integrated structure designs for photovoltaic arrays integ... · san[t-81-7191 / vi integrated...
TRANSCRIPT
..... Vh"lUUU
CONTRACTOR REPORT
DE83013143SAN[t-81-7191
/vi
Integrated Structure Designsfor Photovoltaic Arrays
H. A. Franklin, Project ManagerBechtel Group, Inc.Research and Engineering OperationSan Francisco, California
Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185and Livermore, California 94550 for the United States Department of Energyunder Contract DE-AC04~76DPOQ789
Printed April 1983
RlPROOUCtO 6YNATIONAL TECHNICALINFORMATION SERVICE
us OEPARTMENT OF COMMERCE.. SPRINGfIElD, VA 22161
/9.2./17
.11 / r
Issued by Sandia National Laboratories. operated for the United StatesDepartment of Energy by Sandia Corporation.NOTICE: This report was prepared as an account of work sponsored by anagency of the United States Government. Neither the United States Government nor any agency thereof, Dor any of their employees, nor any of theircontractors, subcontractors, or their employees, makes any warranty, expressor implied, or asllUIDes any legal liability or responsibility for the accuracy,completeness.. or usefulness of any information, apparatus, product, or process disclosea, or represents that its use would not infringe privately ownedrights. Reference herein to any specific commercial product, process, orservice by trade DaIn6, trademark, manufacturer, or otherwise, does notnecessarily constitute or imply its endorsement, recommendation, or favoringby the United States Government, any agency thereof or any of theircontractors or subcontractors. The views and opinions expressed herein donot necessarily state or reflect those of the United States Government, anyagency thereof or any of their contractors or subcontractors.
Printed in the United States of AmericaAvailable fromNational Technical Information ServiceU.S. Department of Commerce5285 Port Royal RoadSpringfield, VA 22161
NTIS price codesPrinted copy: AllMicrofiche copy: A01
SAND81-7191Unlimited ReleasePrinted April 1983
DistributionCategory UC - 63a
Integrated Structure Designsfor Photovoltaic Arrays
H. A. Franklin, Project ManagerBechtel Group, Inc.
Research and Engineering OperationSan Francisco, California
Work performed for Sandia National LaboratoriesUnder Contract 62-9877
AbstractThis report describes the third phase of a multi-year program to investigate the designof low-cost support structures for solar photovoltaic arrays used for central powerproduction. Earlier phases concerned conceptual designs, extensive surveys of solarsystem manufacturers, preliminary cost estimates, development of structural designcriteria for arrays, and wind tunnel tests to provide needed wind design parameters.The present final phase focuses on two low-cost candidate support concepts. Detaileddesign integration was investigated after some comparative costing was done for aselection of possible concepts. Design integration involved detailing various foundationand superstructure systems to reduce structural redundancy and improve systemefficiency for fabrication and installation. Additional wind-tunnel tests of these fixed,flat-panel arrays extended earlier experimental results. The structural responses andaerodynamic stability of candidate designs were checked by wind dynamic analyses.This study led eventually to the fabrication and installation of a non-operativedemonstration array in Albuquerque, NM, which showed the simplicity and effectiveness of the selected design for future large-scale applications.
CONTENTSSection
Page1 INTEGRATED ARRAY DESIGNS - TASK I 1-1
1.1 Basis for Structural Concepts 1-21.1.1 Basic Plant Layout Principles 1-41.1.2 Materials, Codes, and Standards 1-71.1.3 Review of Available Concepts 1-10
1.2 Conceptual Array Designs 1-131.2.1 Panel Parameter Study 1-131.2.2 Concept 1 - Torque Tube W/Caisson 1-211.2.3 Concept 2 - Torque Tube W/Steel Pipe Pile 1-211.2.4 Concept 3 - Torque Tube W/H-Pile 1-251.2.5 Concept 4 - Torque Tube W/Wood Pile 1-251.2.6 Concept 5 - L-Shaped Corrugated Steel 1-281.2.7 Concept 6 - Half-Pipe Corrugated Steel 1-281.2.8 Selected Candidate Designs 1-33
1.3 Final Des i gn Crite ri a 1-351.3.1 Loading Criteria 1-351.3.2 Load Combinations 1-421.3.3 Foundation Criteria 1-44
1.4 Design Optimizations 1-501.4.1 Panel/Module Integration 1-511.4.2 Single Supports versus Double Supports 1-581.4.3 Span Length Variations 1-591.4.4 Variation of Support Locations 1-60
1.5 Selected Final Designs 1-631.5.1 Integrated Superstructure Design 1-641.5.2 Substructure Designs 1-681.5.3 Connections and Details 1-73
1.6 Array Field Fence Study 1-77
v
Section
2
3
CONTENTS (Cont'd)
DYNAMIC WIND ANALYSIS - TASK II
2.1 General
2.1.1 Basic Approach
2.1.2 Scope of Investigation
2.2 Mathematical Models of Photovoltaic Arrays
2.3 Analysis For Steady-State Wind Loads
2.4 Modal Analysis
2.5 Dynamic Instability Evaluation
2.5.1 Flutter Instability
2.5.2 Vortex Shedding
2.6 Dynamic Response Analysis
2.6.1 Dynamic Wind Load
2.6.2 Dynamic Response Analysis Method
and Results
2.6.3 Discussion of Results
2.6.4 Fatigue Considerations
2.7 Summary and Conclusions
2.8 Recommendations for Future Studies
SYSTEM COST ESTIMATES - TASK III
3.1 Bases for Construction Cost Estimates
3.1.1 Field Costs
3.1.2 Engineering Services
3.1.3 Allowance for Uncertainty
3.1.4 Qualifications
3.1.5 Exclusions
3.2 Preliminary Cost Estimate Analysis
vi
2-1
2-1
2-1
2-3
2-3
2-4
2-72-11
2-11
2-13
2-14
2-14
2-18
2-21
2-262-26
2-28
3-1
3-2
3-43-6
3-7
3-7
3-8
3-8
CONTENTS (Cont'd)Section
3 (Cont'd)
3-9
3-223-22
4-1
4-2
4-34-4
4-4
4-54-8
3-14
3-143-15
3-18
Field Costs
Qualifications, Exclusions, andAssumptions
EvaluationCost Estimate Analysis
Field Costs
3.2.3
Fi nal
3.3.1
3.3.2
3.2.1
3.2.2
3.3
4.1 Prototype Array Design Criteria
4.2 Design of Demonstration Array4.3 Dummy Solar Modules
4.4 Construction Specifications
4.5 Construction of Demonstration Array
4.6 Review of Construction
Qualifications, Exclusions, and
Assumptions
3.3.3 Summary
PROTOTYPE HARDWARE - TASK IV4
5 WIND TUNNEL TESTS - TASK V
5.1 Scope5.2 Summary and Conclusions
5-1
5-1
5-2
REFERENCES R-1
vii
Section
APPENDIX A -
APPENDIX B -
APPENDIX C -
APPENDIX D -
APPENDIX E -
CONTENTS (Cont'd)
Drawings for Large Array Field and Constructionof Demonstration Array
Mode Shape Plots of Photovoltaic StructuralSystems (Concrete Single-Post and TimberDouble-Post Designs)Derivation of Fluctuating Wind Pressure Power
Spectral-Density Functions in Reference 2-1
Technical Specifications for the Fabricationand Construction of Demonstration Dummy PhotovoltaicArraysCSU Wi nd Tunnel Study-Phase II
vi i i
Fi gu re
1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
1-9
1-10
1-11
1-12
1-13
1-14
1-15
1-16
1-17
1-18
1-19
1-20
1-21
1-22
ILLUSTRATIONS
Definitions of Terminology
Array Spacing and Roadway Allocations
A 1 MW Array Fi e1 d Layout
JPL Array Support Structure
Single Supported Panel - Design A
Single Supported Panel - Design B
Single Supported Panel - Design C - Torque Tube
End-Supported Panel - Concept A
End-Supported Panel - Concept B
Foundation Concepts for Torque Tube System
Design Curves for Concrete Caissons
Connection of Torque Tube Pedestal to Concrete Caisson
Design Curves for Steel Pipe Supports
Conceptual Connection of Torque Tube to Steel Tube Pile
Conceptual Connection of Torque Tube to H-Pi1e
Conceptual Connection of Torque Tube to Wood Pile
L-Shaped Corrugated Steel Support System for PV Arrays
Corrugated Steel Half-Pipe Array Support
Resultant Wind Forces On Arrays
Assumed Soil Reactions due to Lateral Loads fora Short Rigid Pile in Granular Material
Framed Module Support-Concept A
Framed Module Support-Concept B
ix
1-3
1-6
1~8
1-11
1-14
1-15
1-16
1-18
1-19
1-20
1-22
1-23
1-24
1-26
1-27
1-29
1-30
1-32
1-39
1-47
1-52
1-53
ILLUSTRATIONS (Cont'd)
Fi gu re
1-23 Framed Module Support Concept C 1-54
1-24(a) Conceptual 8' x 20' Panel for Frameless Modules 1-56
1-24(b) Sections For Conceptual Panel 1-57
1-25 Total Installed Direct Costs Versus Span Lengths 1-61
1-26 Typical Row of 8' x 36' Arrays - Double-Supported 1-62Cantilever System
1-27 Typi cal Row of Prototype Arrays 1-65
1-28 Attachment of Photovoltai c Modul es to Panel Tees 1-66
1-29 Attachment of Photovoltaic Modules to Panel Tees 1-67
1-30 Embedment Depths of Caissons in PrototypeArray Field 1-69
1-31 Extended Concrete Caisson for Prototype 1-71Array Field
1-32 Connection of Torque Tube to Extended Concrete Caisson 1-74
1-33 Connection of Torque Tube to Timber Pole 1-75
1-34 Typi cal Connect ion Between Adjacent Torque Tubes 1-76
2-1 Finite Element Model of Single-Post 8' x 20' Array 2-5
2-2 Finite Element Model of Double-Post 8' x 20' Array 2-6
2-3 Steady-State Wind Pressure Loading for Extreme S-WWi nd 2-8
2-4 Simplified Steady-State S-W Wind Pressure Loading 2-9
2-5 Wind Pressure Power-Spectral-Density Input 2-17
3-1 Cost Estimate Flow Diagram 3-3
3-2 Installation Scenario for a PrototypeArray Field 3-19
4-1 Demonstration Reinforced Concrete Caisson 4-10
4-2 Demonstration Panel Fabrication 4-11
4-3 Lifting An 8' x 36' Panel Assembly 4-12
4-4 Completed 2 - Span Demonstration Array 4-13
5-1 Typi cal Corner of an Array Fi el d 5-4
x
Table
1-1
1-2
2-1
2-2
2-3
2-4
2-5
3-1
3-2
3-3
3-4
3-5
3-6
3-7
5-1
5-2
TABLES
Preliminary Support Concepts for 8' x 20' Arrays
Required Embedments of the Timber Pole and ConcreteCaisson
Structural Response Results - Steady-State Wind Loads,S-W Wind On a Single-Post Design
Natural Frequencies of PV Array Structural Systems
Recommended Damping Values
Dynamic Response Results - Fluctuating Wind Loads,S-W Wind on Single-Post design - 1% Damping
Dynamic Response Results - Fluctuating Wind loads,S-W Wind on Single-Post Design - 2% Damping
Comparative Estimate Summary for 1 MW Array Field
Comparative Estimate Summary for 50 MW Array Field
Comparison of Direct Costs For 1 MW Field
Comparison of Direct Costs For 50 MW Field
Total Cost Comparisons of the Double-SupportedArray Structure for Various Spans
Unit Cost Comparisons of the Double-SupportedArray Structure for Various Spans
Unit Costs of the Array Structure Having theDouble-Supported Cantilever System
Maximum Values of Force Coefficients
Recommended Force Coefficients for Design Criteria
xi
1-34
1-72
2-10
2-12
2-20
2-22
2-23
3-10
3-11
3-12
3-13
3-16
3-17
3-21
5-3
5-3
Section 1
INTEGRATED ARRAY DESIGNS-TASK I
This report describes Phase III of work done for Sandia National Laboratories
by Bechtel Group Inc. in the development of low-cost support structures for
photovoltaic array fields for central power production. In the first phase
reported (Reference 1-1), in July 1979, many concepts were discussed in
order to identify low-cost candidates. The second phase reported (Reference
1-2) in May 1980, developed the wind engineering for these flat panel arrays
and gave wind-tunnel test results. Phase III covered detailed design inte
9ration of candidate support concepts, and utilized wind dynamic analysis.
A second phase of wind-tunnel testing was completed and a demonstration
array was constructed in Albuquerque during 1982. This Phase III work was
divided into five technical tasks which provide the main headings throughout
the report.
The objectives of Task I were to:
o Develop integrated conceptual structural designs for two promisingphotovoltaic array support systems
o Develop improved structural design criteria.
The approach taken to begin the study was to review several likely support
concepts on a common design basis, and then compare their costs. Some
alternate designs were explored for possible cost reductions. Finally two
candidate concepts were selected for detailed design integration and final
cost estimates. Design criteria assumed an Albuquerque site as the basis
for concept designs. Dynamic effects due to fluctuating winds were esti
mated from dynamic analyses and new information was obtained from wind-tunnel
tests which was used in the final designs of the support structures.
1-1
Module support concepts were developed which utilized structural integration
and thus achieved reduced costs of the panel structure. Defi nitions and
terminology used in this report are given in Figure 1-1.
1.1 BASIS FOR STRUCTURAL CONCEPTS
Structural concepts for array supports are strongly influenced by panel sizes
and by applied loads. As the array panels increase in size, their average
height above ground also increases, for a fixed angle of tilt. These two
factors both result in a greater total load on the panel from an applied pres
sure. This trend suggests that increasing the size of such structures would
require an increase in strength and this would likely increase unit costs.
On the other hand, the smaller the arrays, the greater the number of repeata
ble components and connections that must be dealt with for a given collector
area. Such an increase in numbers can significantly influence the cost of
fabrication and installation of field arrays. For the purpose of this con
ceptua1 work, and recogn i zi ng trends in photovolta i c technology, modul es
were specified to be 4 ft square.
Previous Bechtel studies had found that 8' x 24' panel sizes for photovol
taic arrays demonstrated minimum cost for a practical configuration (Reference
1-1). In that study, the panels were assumed to be end supported. However,
singly supported panels seemed to offer several advantages over end-supported
ones at the beginning of this study. For example, singly supported arrays
have the advantage of not needi ng e1evat i on ali gnment with adjacent arrays
and thus are more amenable to automated field installation. This offers
potential for cost reductions. For uniform loads, the maximum deflection
of a si ngly-supported array is 60% of the maxi mum deflect i on of an end-sup
ported array. Usi ng si ngly-supported instead of end-supported arrays, one
1-2
PANEL
~,,,~~)
PANELFRAMEWORK
j
ARRAY
/i,,,I ', ,'_.'
I
/ iI
II Ilib __ 1Ir------ ---11II II
MODULE
TORQUE TUBE
SUPPORTSTRUCTURE
SOLAR CELLS
ARRAY SUBFIELD - A group of solar pha!avaltaicarrays assoc'lated by the collection of branch circuitsthat achieves the rated dc power level of the powerconditioning unit.
AR RAY FIELD - The aggregate of all array subfieldsthat generate power within the photovoltaic centralpower station.
PHOTOVOLTAIC CENTRAL POWER STATION The array field together with auxiliary systems(power conditioning, wiring, switchyard, protection.control) and facilities required to convert terrestrialsunlight into ac electrical energy suitable for connection to an electric power grid.
SUPPORT STRUCTURE - The panel (excludingmodules) including caisson or wood pole foundationsthat project above ground level.
ARRAY - A mechanically integrated assembly ofpanels together with support structure (includingfoundations) and other components, as required, toform a free-standing, field-installed unit that producesdc power. The support structure and the modulesmake up the array.
SOLAR CELL - The basic photovoltaic device whichgenerates electricity when exposed to sunlight.
PANEL - A collection of one or more modulesfastened together, preassembled and wired, forming afield·installable unit. The panel frame includes thetorque tube.
MODU LE - The smallest, complete, environmentallyprotected assembly of solar cells and other components (including electrical connectors) designed togenerate de power when under unconcentrated terrestrial sunlight.
ARRAY SUBFIELD
ROADS
If ~- ---- - ..II II
I iI I
I '/I,
II II II II Iu, ___
- --- -L""",,,PHOTOVOLTAIC CENTRAL POWER STATION
Figure 1-1 Definitions of Terminology
1-3
foundation is eliminated in any particular row of arrays. For these reasons,
the single support was chosen for the conceptual array designs to begin this
work.
For ease of handling and installation, 20-ft long panels were chosen for
pre1i mi na ry cost compari sons instead of 24-ft ones. It was expected that
20-ft panel lengths would still be within the realm of low cost. A previous
Bechte 1 study (Reference 1-1) had also proposed and shown the benefit of
using hypothetical 4' x 8' modules. Sheets of plywood are examples of this
unit size which are easily handled by construction equipment and work crews.
Therefore, panel slope height (or width) was taken to be 8 ft. Since 4' x 8'
modules are still mostly in the development stage, and since industry has al
ready developed the 4' x 4' module, the 4' x 4' modules were specified here
for design integration.
1.1.1 Basic Plant Layout Principles
The structural concepts are closely related to the overall layout of a field
of arrays. The degree of repeatabil ity, the sizes, and the requi red spaci ng
of the arrays also require some preliminary definitions.
The power output for an array field is mainly a function of:
o Solar insolation
o Area of the array system
o Effi ci ency of the array system
The solar insolation is defined as the rate of delivery of all direct solar
energy per unit of horizontal surface. The efficiency of the system defines
what percentage of the total energy input from the sun is utilized as output
for the array system. The plant array output divided by the product of the
1-4
solar insolation and the system efficiency will give the required solar
module area. This study assumed a solar insolation of 1 kW/m 2 with a
system efficiency of 15%.
A 15% efficiency is considered high in some photovoltaic systems, the practical
average being in the range of 10-13%. It was assumed that 15% efficiency
would represent a well-designed array field with minimal energy losses.
However, the array cost per unit module area is not significantly related
to the assumption of system efficiency.
In this study, two plant sizes were considered for the conceptual designs:
(1) a 1 MWarray field which would represent a minimum-size central power
producing field and (2) a 50 MW array field which would represent a typical
large central power system. A 50 MW plant size would, most likely, consist
of several building-block systems. That is, the large central power system
would probably consist of identical blocks of baseline array fields. For
example, the 50 MW system could consist of either six 8 MW baseline array
fields or five 10 MW array fields. In this study, the 10 MW array field
was assumed as the baseline in the economic evaluation of the final designs.
Figure 1-2 shows typical spacing between rows for both array fields.
This spacing makes allowances for roads where the simple, low arrays
could be erected using standard mobile cranes. The mobile crane assumed
here can reach over three arrays on each side to erect the panels, meaning that
one roadway between six arrays is needed.
The panels are assumed tilted at 35°, the design latitude of a site located
at Albuquerque, New Mexico. Access between the arrays is assumed to be 10 ft,
1-5
TH
ISA
SS
UM
ES
INS
TA
LL
AT
ION
OF
LOW
PA
NE
LA
RR
AY
S
US
ING
CO
NV
EN
TIO
NA
LE
QU
IPM
EN
T
LIM
ITO
FR
EA
CH
14
'R
OA
DW
AY
)/
35
°
.~
I
8'S
LOP
EP
AN
ELS
>-' I '"
1.5
'
Figu
re1-
2A
rray
Spa
cing
and
Ro
adw
ayA
lloca
tio
ns
which would be adequate for small maintenance vehicles.
Figure 1-3 shows what a typical plant layout of a 1 MW array field would look
like. Using the assumed solar insolation of 1 kWjm2 and a 15% system effi
ciency, 450 units having 8' x 20' panels would be required.
1.1.2 Materials, Codes, and Standards
Specifications and guidelines for use with steel, concrete, and timber are
described in References 1-3 through 1-6. Large portions of such material
specifications are typically taken and incorporated directly into the major
regional Building Codes of this country. For example, the Uniform Building
Code (UBC) (Reference 1-7) of the Western States has chapters devoted to all
the above materials which are almost identical to the particular industry
standards.
In order to ensure proper performance of structures, especially with regard
to public safety, these regional Building Codes also define loads and loading
combinations. Note that these Codes do not govern specialized structures
which require unique considerations, such as bridges, transmission towers,
or photovoltaic arrays.
Each of the common construction materials, i.e. concrete, steel and timber,
have wide variations in strength. These strengths are largely functions of
mix design and placement (concrete), the alloying and rolling process
(steel), or the species and grade (timber). In general, the higher the
strength, the greater the material cost and the less quantity of material
needed. For the purpose of this study, the following material strengths
are used:
1-7
I·M
WA
RR
AY
FIE
LDR
EQ
UIR
ES
450
8'x
20'A
RR
AY
S(A
SS
UM
ES
20'S
PA
NA
RR
AY
S)
\-2
5@
20
'=5
00
'_I
II
II
II
tI
II
II
II
II
II
II
II
II
II-
--,·
r
II
II
II
II
II
II
III~~
---r
=T
1I
I'
JJ
LI'
II
II
II
II
II
II
1_I=
r',
=r=T
I=r
=JI-
.J
II
II
II
I--
-,--
II
II
II
II-
--.L
-T--
-.L1
-I
.LI--::-.::r-=-~
[I
II
r:r:
::::
LI
II
II
II
II
"II=
:LI
I::.::
..LII
I_
I
II
II
II
II
II
~LAIRG~VE~IC~E~CC~SS~
II
II
II
II~
II
II
II
I.I
II
II
II
I1.
=.1.
..~L
IL.:
....
LI
L=
r...
.::J
......
I 00
II
II
II
LI
II
II
II
II-
--.L
=r
L--
-.-T
.LI-
L--
-.L
...J
CU
II
II
TI
II
II.
1I
II
LI
II
II
II
II
Ir.. .....
(I
,Iff
Ir
II
II
II
II
tI
II
II
rI
II
IN
c:::
:LI
1_
1T
'I
II
II
II
II
II.
II
tI
II
II
I
L..
L1
II
II
II
II
LI
IL
I"J
II
II
II
II
I
-LA
RG
EV
EH
ICLE
AC
CE
SS
-~
II
II
II
II
II
jI
1I
II
II
It
II
II
II
!I
II
II
II
II
II
II
II
I!
II
II
II
II
II
-I
'"'"
11
11
11
'"""'~
II
II
IC
l.I
II
II
II
I=$
II
I!
II
f-
!!
II
I~
!I
Ia
II
I.-
[I=
r=I
II
Ir=
L::
:I::
J.=
:LI
IL
r::::
.:.L
II
II
II
II
I.C
o <ri
c...
L.-
---r
--.-
T.
L..L
II
II
II
II
1:1
.I
II
II
II
II
i'I
figu
re1·
3T
ypic
al1-
MW
Arr
ayF
ield
Lay
out
Concrete f'c = 3,000 psi
Reinforcing Steel fy 60,000 psi (Gr. 60)
Structural Steel fy = 36,000 psi ASTM A-36(50,000 psi ASTM A500, Grade C forstructural tubi ng)
Where fy = yield stress of steel
f'c = 28-day cyli nder strength for concrete
These strengths are commonly used in design practice, reflect materials with
general availability and, when used, would lead to lowest total cost for
lightly loaded structures.
To provide a basis for evaluating different array support concepts, the fol-
lowing specifications were assumed for preliminary array designs and design
parameter studies. These include:
0 4' x 4' photovoltaic modules
0 30 psf resultant wind load
0 8' x 20' panel size
0 Steel, concrete, or timber construction
0 Working stress design for parameter studies
0 UBC design method for pile foundations
0 1.5 foundation factor of safety due to uplift
Only wind and dead load combinations were considered in the preliminary array
designs. Applicable codes, specifications, and guidelines described earlier
were used with working stress design for steel and timber construction and
ultimate strength design was used for concrete construction.
1-9
1.1.3 Review of Available Concepts
Most of the central power system concepts revi ewed duri ng the Task I work were
deri ved from a previ ous Bechtel report (Reference 1-1) and from work by JPL
(Reference 1-8). All of these conceptual schemes revealed some possibilities
for providing low costs in large photovoltaic power systems and included
detailed considerations of support structures and foundations.
Basic photovoltaic system classifications are fixed: seasonally adjustable,
or fully tracking. Most tracking systems use motor-driven arrangements supported
on a pedestal or on a carousel. However, tracking systems are not included in
the present work.
Conceptual work for residential photovoltaic applications is shown in References
1-9 and 1-10. Residential photovoltaic systems are characterized by a relatively
Small array size (10 kW or less for an average home) and by attachment to an
existing structural system. Furthermore, structural systems related to con
temporary residential construction are governed by regional conventions in
servi ces provi ded by archi tects and the buil di ng trades. These buil di ng systems
or structures are not easily subjected to changes to suit some extraneous equip
ment; rather, the eqUipment tends to suit the structures.
Therefore residential concepts had little to offer to the problem of evolving
effective structures for a central power system and were not pursued further
in this work.
JPL Concept. One array support concept which showed potential for low cost
was the JPL structure (Reference 1-8). Shown in Figure 1-4, the structure
1-10
>-' I >-'
>-'
4'x
4'
UN
FR
AM
ED
MO
DU
LE
S
20
'
Wi?'ii
/;1B
UR
IED
PO
RT
ION
OF
FO
UN
DA
TIO
N
EM
BE
DD
ED
PL
YW
OO
DB
AS
EP
LA
TE
Figu
re1-
4JP
LA
rray
Su
pp
ort
Str
uctu
re
CO
LD
-FO
RM
ED
ST
EE
LS
UP
PO
RT
FR
AM
E
~/ 7
4"
x4
"W
OO
DP
OS
TS
CG
RO
UN
DSURF~CE
/-----
---/'/
8'
consists of an 8' x 20' frame supported at two ends by a wood foundation.
The frame is made of galvanized steel and array supports are 4" x 4" douglas
fir members chemically impregnated for a 20-year life against deterioration
using a Wolmanizing process. The legs of the support structure are attached
to a 4" x 6" wood beam which is connected to a 1/2-in. plywood base using gal
vanized steel brackets. Foundation support is provided by digging a trench,
inserting the plywood base of the support structure in the trench, and
backfilling the trench with tamped earth. This foundation has been reported
(Reference 1-8) load tested to a pull of 7,000 lb. With the exception of
the two ends, each foundation supports two panel frames.
Some fabrication problems were indicated. Joints of the galvanized steel
panel frame are wire welded and painted where heat from the welding process
had burnt the galvanizing. Welding galvanized steel produces poisonous
gases which requires either adequate ventilation, preferably working outside,
or requi res workers to wear masks. For most regions of the country, the
8' x 20' panel frame is too large to be dipped in a galvanizing tank.
Installation time for the JPL concept is said to be reduced by having the
array supports preassembled in the factory and by having the modules al ready
attached to the panel frame before shipment to the job site. Damage to actual
photovoltaic modules can be avoided by using a dummy panel frame as a jig to
align the foundations. It is estimated by JPL to take approximately 3 to 4
hours to dig a trench, refill, and tamp it to about 85% compaction.
Bechtel Structures. In a previous study, Bechtel conceptualized and identi
fied several array support structures which showed potential for low cost
(Reference 1-1). Some of these concepts included:
1-12
o Single beam (torque tube) - concrete caisson
o Single beam - driven wood pole foundation
o Longitudinal frame - earth screw anchor
o Longitudinal frame wood pile in preaugered hole.
These structures, including the JPL concept, were all examined as candidate
concepts. The resulting preliminary concepts are discussed in the next
sect ions.
1.2
1.2.1
CONCEPTUAL ARRAY DESIGNS
Panel Parameter Study
An early step in arriving at an integrated low-cost array design was to per
form structural parameter studies. Since the module support members contribute
significantly to the total costs, various panel structures were reviewed and
analyzed for structural efficiency and simplicity. Thus, for a single assumed
panel load, the members were sized and weights derived. In this approach the
fabrication costs were assumed to be proportional to panel weights, hence lighter
panels imply lower costs. In this parameter study, it was also assumed that the
4' x 4' modules have adequate stiffness to be supported only along two opposing
edges.
Figures 1-5 through 1-7 show various panel concepts for a panel attached in
its center on a single support. For example, for the panel design shown in
Figure 1-5, loads normal to any panel are transferred indirectly from the
modules, to the longitudinal header angles, to the central tee sections, and
finally to the short longitudinal tube into the center support. However,
the lightest panel weight was found to be the torque-tube concept shown in
1-13
......
I ......
.<>
21
/2x2
x3/1
6L
HE
AD
ER
S-.-J
6x3
1/2
x5/1
6L
5@4'
'=20
'
WT
4x5
21/
2x2x
3116
L
TR
AN
SV
ER
SE
AN
GLE
OR
TE
E,\
'_
HE
AD
ER
AN
GL
E
,..-
ST
UB
TU
BE
WS
EC
TIO
NP
AN
EL
WT
=79
4LB
(EX
CL
UD
ING
MO
DU
LE
S)
Fig
ure
1-5
Sin
gle
Sup
port
edP
anel
-D
esig
nA
C4
x5.4
c..
"..
/
~
~7x
1xV
.TU
BE
C8x
11.5
,..> , >--' <.J'
HE
AD
ER
CH
AN
NE
LS
'---
'
..
CE
NIE
R,U
BE
W_
SE
C11
0 NP
AN
EL
WI~9
94L
!IE
)(C
LV
DIN
GM
OD
ULE
S)
figu
re1.
6Si
ngle
SUpp
oned
Pane
l-
Des
ign
B
al ~...J ...J
~ 8" :;;
~ ~...J 0W ::J2 d« xQ. w
2aIUw<I)
?:
wwI-
1-16
Figure 1-7 where loads are di rectly transferred from the transverse tee
sections, into the longitudinal tube, then to the single support.
For end-supported pane1 concepts, shown in Figures 1-8 and 1-9, a window
pane-type panel was first conceptualized and designed, but this yielded the
1argest total weight. The configuration of that panel is similar to the
JPL concept where unframed modules are also installed into a window pane-type
structure.
The most efficient end-supported pane1 again was the torque-tube concept
shown in Figure 1-9 where loads are taken out more directly into the founda
tion. The resulting analysis demonstrated that identical optimum designs re
sult using the torque-tube concept which are independent of being supported at
the center or at the ends.
Since the integrated pane1 and torque tube provided the least weight, the
simplest structure, and offered the best potential for low cost, then founda
tion schemes for single pedestal supports were investigated using the torque
tube panel concept as a superstructure. These foundation schemes included:
o Concrete caisson
o Steel pipe pile
o Steel H-pile
o Timber pile
Shown in Figure 1-10 is the torque tube panel concept having these foundation
schemes. This panel concept and foundation schemes were cost estimated and
are reported later.
1-17
1::>N
"~@I
""
1-13
...J10-Mx
'"x~N lI:l
...JLO
'"..rIIfs:...JLlJZ<t0.
iilUJ...J::Joo:;:
"2o::J...J<.lXUJ
lI:l
i§()
ICii16...
"C
i~"CC
LlJ
1-19
20
'
6x6
x3
/16
TU
BE S
UP
PO
RT
FO
RT
OR
QU
ET
UB
E•
I
~WT4x5
21>
x2
x3
/16
LA
TE
ND
S
NO
TE
:P
RE
LIM
INA
RY
DE
SIG
NS
US
ING
CO
DE
S
r-
Tr-
rr I1
GR
OU
ND
11
4'·
4"
W6
x1
51
SU
RF
AC
E~
II
II
II
1,
~:f:»
II
I1
1I
1I
I1
3'-
0"
·W:·~!
11
15
'-0
"1
2'-
0"
11
~1
5"
t·~~~.
~~~·
11
HP
8x
36
--"I
1__
12
"~
..."~I+-
16
"1
1'0
"_
....
.0.-.
II
.~~
1I
1I
II
II
1I
1\
I1
..~,
~:ej
:;(1
)Cg
A~~:
g~E
U(2
)S
TE
EL
PIP
EP
ILE
I((4
)W
OO
DP
OLE
~
II(3
)H
-PIL
E~
-
..... I N o
Figu
re1-
10F
ound
atio
nC
once
pts
for
Tor
que
Tub
eS
yste
m
1.2.2 Concept 1 - Torgue Tube with Caisson
The des i gn of cai sson-type foundations i nvo1ves many paramete rs and is very
site-dependent. Two main factors which influence this design are the uplift
forces and the requi red 1atera1 soi 1 res i stance. Therefore, to achi eve a
low-cost design, the design curves shown in Figure 1-11 were drawn. These
curves give the depth of foundation embedment, the factor of safety due to
uplift, and the foundation weight as functions of the caisson diameter. The
requirements for this design were to have an acceptable caisson diameter for
construction which would yield minimal weight and still maintain at least
an uplift factor of safety of 2. For example, as shown in Figure I-II, a 1.5
ft diameter caisson would require an embedment depth of about 10.3 ft for
lateral soil resistance, have an uplift factor of safety of 1.7, and require
concrete weight of 2800 lb.
Shown in Figure 1-12 is the resulting concrete caisson design supporting
the torque-tube panel design. Minimal steel is required in the caisson to
resist the bending and uplift loads and still meet Code steel requirements
for temperature and shrinkage.
In this preliminary design, a W6 x 15 steel section was used as a pedestal
support for the torque tube and woul d be attached to the cai sson by anchor
bolts. The pedestal was assumed to be wel ded to the torque tube prior to
field installation.
1.2.3 Concept 2 - Torgue Tube with Steel Pipe Pile
Another support concept for the torque-tube superstructure which showed prom
ise for low cost was the driven steel pipe pile. Using a similar approach to
that for the concrete caisson, design curves, as shown in Figure 1-13, were
1-21
2.5 10.0
12.0
11.0
10.0
I- 9.0..." 8.0:I: iI-
7.0... ..UJ ..0 ~
2: 6.0 ~!~0
~~tJ)
!!? 5.0 ~~~~« ~lJj !u~
4.0 -o~~~~
3.0 ~
,:;....2.0 ~......1.0
3.0
2.0I.....J...:::l
1.5 a:o...eliu:
1.0
0.5
0.8
6.0
4.0
2.0
I:I:oUJ~
2:o~«()
o 1.0 2.0 3.0 4.0CAISSON DIAMETER - b (FT)
F.S.= FACTOR OF SAFETY
NOTE: CURVES DERIVED FROM USC CODE FORMULA
Figure 1-11 Design Curves for Concrete Caissons
1- 22
o 0
~ TORQUE TUBE PEDESTAL
W 6 x 15
#4 BAR EACH CORNER,4 BARS TOTAL
NOTE: PRELIMINARY DESIGNUSING CODE
11'·0"
ANCHORBOLTS
GROUND /-;= 12" x 10" x \1,," BASE PLATE
SURFA:lT::-..-_---1r~S~~~~~~;::;p~:...I___-------IiiI
r II IJ-,lit- - - --1t-1~ LL-=- __c:!J~
r-------i~-----Ir - - - - - ---1'.......-If-- NO.3 TIES@8"O.Cr------j,-----ir-------Jr-- - - --i++--~--------jI Ir-----,~------i
,-----1L-----lI I1-----1I I
16"
Figure 1-12 Connection of Torque Tube Pedestal to Concrete Caisson
1-23
16.0
70014.0
I- 12.0 60011.
"tlboVERSUS d ..ci
J: FOR LATERAL500I- 10.0 I-0- SOIL RESISTANCEUJ 3:
0UJ -'
UJ-' UJ0- I-
8.0 400 en
6.0
4.0
b 0 VERSUS
STEEL WT
(t=1/4")
300
200
10 20 30 40
OUTSIDE PIPE DIAMETER bo(in)
NOTE: CURVES DERIVED FROM USC CODE FORMULA
Figure 1-13 Design Curves for Steel Pipe Supports
1-24
drawn for the steel pipe pile. Here again, the depth of embedment and the
total steel weight are given as functions of the outside pipe diameter. It
is assumed that the pipe has a constant thickness of 1/4 in. The pile depth
needed to meet the lateral soil resistance criteria will also give an uplift
factor of safety of 1.5. It was found that a 12-in. diameter pipe pile
embedded 12 ft into the ground was sufficient for design.
Figure 1-14 shows a connection detail of the torque tube to the steel tube
pipe. Here a pipe stub is shop welded to the torque tube. Rapid field
installation is achieved by inserting the smaller pipe stub into the pipe
and coupling it together by a steel band of the "Marman" type.
1.2.4 Concept 3 - Torgue Tube with H-Pile
Another foundation alternat i ve for the torque tube superstructure was the
H-pile. With the lightest section available, the HP 8 x 36, a 15-ft embedment
was required to meet the lateral soil resistance.
Shown in Figure 1-15 is a conceptual attachment of the torque tube to the
H-pile. In this attachment scheme, a tee stub is shop-welded to the torque
tube and field installation achieved by field-bolting the tee stub to the
H-pil e.
1.2.5 Concept 4 - Torgue Tube with Wood Pole
From an earlier Bechtel study (Reference 1-1), the driven wood pole demon
strated the most potential for a low-cost foundation. Therefore, using
the preliminary criteria, a driven wood pole foundation was designed for the
torque tube superstructure. A 15-i n. di ameter pole embedded 13 ft was
1-25
W...J
W2i:
ll. W
>-C)ll.
~zll.
z:::; '&<tll.::;;::JceO
IF<t'"
~
w \co::J co I I Ir- ::J I \w r- I I I I::J (/)
0 wce ll. I I I I \0 2i:r- I I I I \ .S1
c::
I I I I \ II
II I I \ ..:II I I \
0;
~I I
\.9
-t=II
\ ::lr-
I \ '"::lI \
.,.~
0
I \r-...0
I \ c:
\.g
I al
I \ c:c:
:::t::: \ 8\
c;;
I ::l...I \ 0-
tl
I \c:0
(,)
I \_L
.r
\......
-r- \ '"~I
::l
'"I \ Lr
------- - -' -____J
r-----
ce0ll.ll.::J(/)
...JWz<tll.
1-26
..... I "".....
HP
8x
36
PA
NE
LS
UP
PO
RT
,
t"'
/T
OR
QU
ET
UB
E
10(
WT
6x2
0
5/8
"D
IA.
BO
LTS
A-3
25W
/NU
TS
&L
OC
KW
AS
HE
RS
-1
1/1
6"
x1
9/1
6"
SL
OT
TE
D
HO
LE
INT
EE
-1
1/1
6"
DIA
.H
OL
EIN
PilE
FL
AN
GE
Fig
ure
1-15
Con
cept
ual
Co
nn
ect
ion
of
To
rqu
eT
ube
toH
-Pile
required to meet both the lateral soil resistance and the uplift criteria.
Figure 1-16 shows a conceptual attachment of the torque tube to the wood
pole. Here curved plates are welded to the torque tube to which a split
ri ng is attached. The torque tube is set atop the wood pole duri ng fi e1d
installation and lag screws are inserted through the split ring and embedded
into the wood.
1.2.6 Concept 5 - L-Shaped Corrugated Steel
One attractive feature of the JPL foundation system was the utilization of
the soil surcharge instead of foundation weight (i.e., concrete weight) to
resist uplift forces. Following the same idea, a support foundation system
was conceptualized which used soil surcharge to resist uplift and lateral
forces.
Shown in Figure 1-17, this support system consists of a bent sheet of corru
gated metal which is lowered into a shallow excavated trench and backfilled
with soil. Tubular struts spaced 10 ft apart were used to provide sta
bility to the bent corrugated sheet. Modules can be attached directly to
the corrugated metal sheet using screws or small bolts. Trench depth of 2
ft is required to provide the necessary resistance to lateral and uplift
forces.
1.2.7 Concept 6 - Half-Pipe Corrugated Steel
After developing the L-shaped corrugated support structure, ARMCO steel
representatives were contacted and were asked their opinion of its feasibility
and potential for low cost. They initially thought fabrication (i.e., bending
1-28
7" MIN
... 3" x %" SPLIT RING
~ 'Ji," LAG SCREWS 6" LONG
~ TORQUE TUBE
I I~ 9" x %" CURVED PLATES
I II I
-,--rr-r...-.---r-ll- - - -.:L1lId"- - - - -II--r-rn---..-----
! I ~ CUT-OFF AFTERr+.!::!=;J'--l.--'--,lj--- _ _ _ I DRIVING,
L+;:;:-r-.,.-,J~=::::;:z;zzcl:::f======#~"":~1 15"" TREATED~OUGLASFIR PILE
Figure 1-16 Conceptual Connection of Torque Tube to Wood Pile
1-29
2'M
INT
RE
NC
H
GR
AD
E
,----
ST
RU
T1%
"S
Q.
TU
BE
,S
PA
CE
D,
10
'D.C
.
BA
CK
FIL
L,1
"x
3"
BE
NT
CO
RR
UG
AT
ED
ST
EE
LS
HE
ET
~5
"x
3l\
"x
5/1
6"
AN
GL
E~
~'--T
WO
4'x4
'PV
MO
DU
LE
S
~~.f
.!J
~'o.
,,()
·'J!
~O~9
,o~O~
"'o'fJ
q'b
.:..
;~-··~d.·c"
60
:'0
'9>
:00
C?o'
':Q~
~C>"
":'
0';
,~.
'.?
:·.)
";ti
··~:
O··.
.·0
"0';
'0".Q
·o'
I~~
~Q:o,,·a.
0',
.-0
0•0.•
Qto
:·o
··a"f:
)"'-
....)
v.
0"
'QO
~~,
"'0
',0·O,O~c
.D',{J
o,~·O
.:.0
:.:
.o·.
o~OO
·&"O
Qd.:
.O'Q
:u',o
9',~
oC?
D~(
j:,:
:·:o
:O·:
.o:"
'6.
'9':
'o~:
;!.c;
:':•.
~j~
••(i
~~·.:r...
60~Oo~~
~9o::(;>.:'~·:·.6·
.~o·~.~:.~·.~
.~',0
.:~:
·6.0
.o~~
:'·:
~·:9
·:~·
·:·?
::.o
::o:
·~·:
O~.:-o'
.~~...
'.'0
''''
"'~
..~
·Dr-
.c.
"'0
"D
./""
.'."
"..
,,·.
·.·.
0-.
·.·0
'0··
····
····
'···
':Go:
'-'0
'(J.o
:0'0
'°0'.V·O~·O.···
:·.<?:oc......-e:O::·-~·~··;....:,.
',a'O
,·c
i··~
·OO·
O·c;'
••;.,
'.:•••
:O·"
:O.!:
f:.C
?o,:o
.','
,0-
oQ.
,0'0
.,',
'o
.....L
/O.:
>t:
:;).
o;t
:l.
,',
·.oe.'-..J·~·o····O·
()~;Q-"":
b...
.;·.
O'..··
c·
....
.•0'
'0''
DO~·.O
o.0
00
:0
'.~
...
"~:
.D
' o' o
O·p
·:-·
::.'
~,0
,.
"0O
.o'
.....:0
.::0
C?'
(:i':
'..0
.-.)
~.~
~'''
0'0
O:O'
()~:
'o'O
"O~·oo~-..
Q;·~O':·'
'0°
0o
{):
···:~~·O
e·,:c
:::.'.'
0o'~'
··R:..
..;,
':'0
';O
P:.o
.o:
~·:4·
'..'
00
\:/..
'...
0•
0I).
"•'0
'O
·"C
J'.
.v
·'.'
,..
,.Q
"~
-'=:-"
6'.p
••-
.···0
..0.0
.·.° 0
'';'
0'&
.'b
9C
o,
00
"()
o?·
0"'
,',0
';0
0.'
<>
::•..
··..{
)·~;
·o:
.:';0
.'·~
~,,!
.o.
.;(t
·O·o
•'-
00
.t:>
.,":
:'.'0
'.0:.,?~
0.0
D.:·~
'..~.
0';"
~:o'~
.00.'0
'c~
·~·.~i/O::~::b:o·:;
·.·.
0·.
(:'.
·.·
00
0"
0·
<:)
.·a
Q:O
"O
",,
0~...
t>(,)
't>~oD
0.
0.'
"'0
."..
...
·~'.b'~.ei~<::>.·~
,.-.:.2
..:::..:
:....:..
•..
~.C
D"
.>..'~.6."
·:·c
J..
.~..~
.........
:...
.I
,.... I W a
8"
x7
"x
%"
PL
AT
E
6'·0
"
Figu
re1-
17L
-Sha
ped
Cor
ruga
ted
Stee
lS
uppo
rtS
yste
mfo
rPV
Arr
ays
of the sheet) of the structure could be done only by press braking which
would make it prohibitive for low cost.
However, ARMCO representatives suggested an alternate support concept as
shown in Figure 1-18. This system, using a corrugated half-pipe, showed
attractions of simplicity, ease of shipping and installation, and utilization
of soil as backfill to resist wind uplift. A regular corrugated steel pipe
(as used for culverts, for example) is cut longitudinally in half and struc
tural tee sections are fixed across the free edges to stiffen them and provide
supports for solar modules. The half-pipe is tilted at the latitude angle,
set into a shallow trench, then backfilled with the excavated soil to provide
uplift resistance.
Through di scussions with ARMCO Company representatives, it was clear that
the major cost of this system came from the pipe section itself. Thus, using
smaller, thinner pipes would lead to reduced costs. However, this concept
had difficulty arranging ground clearance for the PV modules at the lower
edge. This was solved by the following methods:
(a) Provide large enough diameter on the pipe that the lower edge
could project above the ground when adequate backfill was placed.
(b) Support the corrugated pipe above ground level on railroad ties
- which coul d be embedded in the ground - and addi ng soil back
fill for weight.
(c) Provide ground clearance for modules by arranging them higher
on the sloping tee-supports and cantilevering them above the
top edge.
1-31
MT
4"
3.25
MO
DU
LESU
PPO
RT
S
2W
'x2
W'x
3116
"L
ED
GE
ST
IFF
NE
R
w--
2""
2""
3111
"L
FRA
ME
\\
.--
4'.
0"
SQ
UA
RE
MO
DU
LES
\\
\\.
---
SOIL
BA
CK
FIL
L
.... !..,8'
·8"
IfC
OR
RU
GA
TE
DN
ST
EE
LPI
PE./
Fi~r
e1·1
8C
orru
gate
dSt
eel~alf.Pipe
Arr
ayS
uppo
rt
These possibilities were explored along with changes in pipe diameter and
metal thickness. It was determined that a 7-ft diameter was about the largest
that had a chance to remain competitive with other designs. Optimizing studies
led to a 16-gauge half-pipe, 6.5 ft in diameter laid in a shallow, 16-in.
deep trench, wi th soil dumped into the pi pe for wei ght. The sola r modul es
were arranged to have I-ft clearance at the lower end and cantilever beyond
the upper edge. No benefits were gained from putting this pipe on rail road
ties, and, in fact, there was some difficulty in achieving adequate lateral
resistance from the railroad tie foundations.
1.2.8 Selected Candidate Designs
Cost estimates are presented in Section 3 for the six Bechtel engineering de
sign concepts. The basis, qualifications, and exclusions of these estimates
are presented in Section 3. Table 1-1 describes the six design concepts
arranged in order of increasing costs.
Table 1-1 shows the torque tube with the driven wood pile foundation as the
most promi si ng low-cost support structure. The second most promi sing concept
is the torque tube with a concrete caisson foundation. The torque tube
with a steel pipe foundation was cost-competitive with the concrete caisson
design and might have been chosen as the second low-cost alternative. How
ever, the end of the pi pe pil e may eas ily deform if the soi 1 condi t ions are
such that small rocks or boulders are encountered during the driving process.
Pipe pile foundations were recently used in the heliostat installation of
the Solar Test Facility in Albuquerque, New Mexico. During their installation,
some piles encountered small boulders and had to be pulled, the deformed tips
had to be cut off, and the piles then redriven at a slightly different location.
1-33
>-' I W """
TA
BL
E1
-1
PR
EL
IMIN
AR
YS
UP
PO
RT
CO
NC
EP
TS
FOR
8'
x2
0'
AR
RA
YS
(AR
RA
NG
ED
INO
RD
ER
OF
INC
RE
AS
ING
CO
ST
S)
CO
NC
EI'T
DE
SC
RIP
TIO
N
fTO
RO
UE
TU
IE1
.1.3
11
.TO
RO
UE
TUBE
WIT
HW
T4
.51
100U
LESU
PPO
RTS
:15H
DIA
.
1W
/wO
OO
PIL
EW
OO
DPO
LE11
FT.
LON
GP
LAC
EO
INA
13-F
T.-O
EE
PA
UG
ER
EO
HO
LE.
FOU
ND
ATI
ON
1-TO
RO
UE
TU
IE6
.6.3
11
6TO
RO
UE
TUBE
WIT
HW
T4
.51
100U
LESU
PPO
RTS
;2
W/C
AISS
ON
W6
.15
SU
PP
OR
TP
EO
ES
TAL
4FT
.HIG
H;1
6"01
A.
CO
NC
RET
EC
AISS
ON
FO
UN
DA
TlO
Il11
FT.O
EE
P.
4-T
OR
OU
ET
UIE
6.6
.31
16
TOR
DU
ETU
BEW
ITH
WT
4.
5110
0ULE
SUPP
OR
TS;1
2"01
A.
3W
/ST
UL
PIPE
STE
EL
PIPE
16fT
.LO
NG
,14"
THIC
KD
RIV
EN
12FT
.D
EEP.
FOU
ND
ATI
ON
/,./
TO
RO
UE
TU
IE••
6.31
16TO
RO
UE
TUBE
WIT
HW
T4
.51
100U
LESU
PPO
RTS
:HP
1.3
64
W/H
-PIL
ES
TEE
LP
ILE
19FT
.LO
NG
DR
IVE
N15
FT.
DEEP
_FO
UN
DA
TIO
N
4."","
6'-
6"
DIA
.C
OR
RU
GA
TED
STE
EL
PIPE
CU
TIN
HA
LFP
RO
VID
ES
SUPP
OR
T5
CO
RR
UG
ATE
D'
FOR
liT
4.3
.25
MO
DU
LESU
PPO
RTS
:HA
LF-P
IPE
STR
UC
TUR
EP
LAC
ED
IN.'
STE
EL
EX
CA
VA
TED
TRE
NC
HA
ND
BA
CK
FIll
ED
.,.--..
,"..
.-S
TRU
CTU
RE
~Lo
SHAP
EOM
OD
ULE
SA
TTA
CH
DlR
EC
lLY
TO1
".3
"B
EN
TC
OR
RU
GA
TED
STE
EL
CO
RR
UG
ATE
DS
HE
ET:
STR
UC
TUR
EPL
ACED
INA
2-FT
.-DE
EP
TRE
NC
HA
ND
BA
CK
Fill
ED
6S
TEE
LTO
RES
IST
UP
LIFT
.S
TRU
CTU
RE
-.~
This was seen to be a disadvantage for the pipe pile and it was passed over
in favor of the caisson concept.
1.3 FINAL DESIGN CRITERIA
The structural design criteria described here form the basis for the final
designs to be discussed in the subsequent sections. The criteria used for
the conceptual array designs of Section 1.2 were consequently re-evaluated
as a part of design optimization. This is a necessary step since structural
support cost is largely determined by the assumptions made in the design
criteria. These assumptions are in regard to materials, loads, loading
conditions, and foundation stabil ity. The material strengths used in the
final designs for wood, steel, and concrete are identical to those previously
defined in Section 1.1.2 for the preliminary designs.
1.3.1 Loading Criteria
Array support structures should generally be designed to sustain dead, live,
wind, seismic, and snow loads. The general set of loadings that were consider
ed in this work for the design of array support structures are listed below:
0 Dead Load (D)
0 Live Load (L)
0 Wi nd Load (W)
0 Snow Load (S)
0 Earthquake Forces (E)
Each of these loads is assumed to act in certain probable combinations and
this is discussed in Section 1.3.3. For environmental loadings, the
assumed site is Albuquerque, New Mexico. Furthermore, this work considers
1-35
the following environmental conditions:
o Operating Environmental Conditions - The support structuresare designed to maintain normal plant operation without powerproduction or interrupted service.
o Extreme Environmental Conditions - The support structures aredesigned for service during plant operation at reduced power,interrupted power production, and damage repair operations.
The characteristics of each design load are discussed below and recommenda-
tions for design values are made.
Oead Load (0). This load is easily determined and consists of the weight
of the photovoltaic modules, mounted equipment, platforms, and supporting
structures. Assumed dead load of the photovoltaic modules was 3.5 psf.
Other design values were calculated using material unit weights furnished
from handbooks or from suppliers. These unit weights are:
Concrete 145 pcf
Gl ass 160 pcf
Stee1 490 pcf
Aluminum 173 pcf
Timber 35 pcf
Live Load (L). This loading is assumed to be due to maintenance and washing
activity on the photovoltaic systems. For design purposes, the cleaning
load is assumed to be a uniform 3 psf loading applied normal to the surface
of any 4' x 8' section of panel. A 20-lb live load per lineal ft was
assumed to be the support reaction along anyone panel edge due to a panel
washing machine. This load was assumed to act only on one panel edge at a
time in addition to the 3 psf loading.
It is assumed that personnel would not be allowed into the modules. However,
1-36
a 200 1b concentrated load is app1i ed to any pa rt on the module support
beams,' representing possible access by personnel to edge members. This load
is not cumulative to maintenance and cleaning loads.
Wind Load (W). Wind load is the significant load for the design of array
support structures, and is characterized by great uncertainty in the determi-
nation of actual magnitudes of velocities and forces. Wind velocities are
norma11y defi ned at 30 ft above ground. Thi sis a standa rd datum for wi nd
specifications and is the height in open country at which minor obstructions
are unimportant. Wind speed criteria can be developed from studies of the
local climatological data and available site-related records for the particu
lar design locations. Where local wind records are not available, wind load
maps and tables are available (Reference 1-11) which define the wind speeds
for a particular mean recurrence interval.
For example, one may design an array for an operational wind speed correspond
ing to a 25-yr recurrence interval and use a 100-yr recurrence interval
for the extreme wind speed. Gusts are not included in these wind speeds for
a particular recurrence interval and are treated separately (Reference 1-11).
Wind speeds may be converted to velocity pressures using the formula:
2 (1-1)Q30 ; 0.00256 V 30
where Q30 ; basic wind pressure (psf) at 30 ft height
V30 ; design wind speed (mph) at 30 ft height
Effective velocity pressures at elevations other than sea-level may be deter
mined by using an atmospheric density correction appropriate to the site
elevation.
1-37
However, the velocity pressure for the array structures which are close to
the ground surface is somewhat lower. Assumptions may be made as to the
velocity or pressure vari ation from the ground surface to the 30 ft datum,
but these may be very inaccurate. For these reasons, wind tunnel tests were
conducted by Colorado State University at Fort Collins, and practical wind
desi gn criteri a were developed for photovoltai c array structures (References
1-2 and 1-12).
From the Colorado State University study, net force coefficients were given
for vari ous array panel arrangements. These force coeffi ci ents defi ne the
wind forces acting on the surface of the panel caused by the wind specified
at the 30 ft datum height.
Speci fically, for the support structure:
panel normal force FN ~ Kq CNA (1-2)
array pitchi ng moment Mz ~ Kq Cmz ADxy (1-3)
and array yawing moment Mz ~ Kq Cms AD xy (1-4)
where q ~ basic wind pressure at 30 ft height
CN ~ normal force coefficient
Cmz ~ pitching moment coefficient
Cms = yawing moment coefficient
A ~ area of panel
Dxy ~ the half-chord di stance ac ros s the panel
K ~ dynamic turbulence factor
The di rect ions of the des i gn load and force components are gi ven in Fi gure
1-19.
The dynamic effects of the wind on the structure must also be considered.
1-38
WIND
SOUTH NORTH
NORMAL FORCE F = K' q' CN . A
PITCHING MOMENT MZ = K . q' CMZ ' A' D XY
YAWING MOMENT MS
= K· q' CMS ' A . DXY
WHERE A = PANEL AREA
DXY
= PANEL CHORD DISTANCE
CN,CMZ' ."''''D eMS DERIVED FROM WIND TUNNEL TESTS
K IS DY~IAMIC TURBULENCE FACTOR
Figure 1-19 Resultant Wind Forces on Arrays
1-39
For example, the total wind force WT on the structure may be thought of as
being composed of two portions, a steady-state part Ws and a fluctuating
part Wf.
(1-5)
The steady-state wind is a quasi-static pressure which the structure re
sists over a long duration. The fluctuating portion of the wind applies
pressure to the structure over a much shorter duration. If the frequency
content of the fl uctuat i ng wi nd Wf is such that it dynami ca 11y excites one
of the fundamental modes of the structure, then Wf may become larger than
Ws '
For the final design, the fluctuating portion Wf was assumed to be propor
tional to the steady-state portion Ws or by the factor K-l. That is,
Wf = (K-1) Ws (1-6)
Therefore, the total wi nd load Wt may be expressed in terms of the steady
state portion Ws by the factor K.
Here, K is defined as the dynamic turbulence factor. Methods of arriving
at the K factor are discussed in Section 2.
The following assumptions were made for the wind loads in the final design of
the arrays in a 10 MW plant:
o 90 mi/h wind speed, any direction (extreme condition)
o Arrays designed as interior arrays
o Force coefficients (for interior array)
eN = 0.30
1-40
Cms = 0.28
Cmz = 0.04
o Oynamic turbulence factor K 2
A wind speed for the operating condition was not used in the final design.
Instead, the extreme wind condition was defined which would normally dominate
the design. The force coefficients used in this design are further discussed
in Section 5. The determination of the value of the dynamic turbulence factor
K is discussed in Section 2.
Snow Load (S). The determination of snow loads is dependent on:
o Location of the specific site
o Velocities of wind at the site
o Wetness of the snow
o Array surface condition and tilt angle.
This kind of load can be considered to be random and the general approach is
to speci fy an upper bound. The current ANSI Standard A58.1 (Reference 1-11)
suggests that snow load in the Albuquerque region be based on local climate
and topography. Snow as deep as 15 in. has been recorded in the Albuquer
que region. Therefore, for design purposes, a snow load of 10 psf was assumed.
Earthquake Forces (E). These forces for conventional structures come from
the combination of structure mass and the accelerations of the ground supports.
Consequently, a lightweight structure will experience relatively small forces
on its supports for a given peak acceleration. A heavier structure will
experience higher loads. The location and magnitudes of earthquakes are
predicted on a probability basis. Critical structures and systems are de
signed for seismic forces by using advanced dynamic analysis, but simpler
1-41
structures, such as those used in photovoltaic systems, use a static equiva
lent force method. This method is described in the Uniform Building Code
(Reference 1-7). The design equation (UBC eq 12-1) for base shear V is:
V = ZIKCSW
The fo 11 owi ng values are assigned to the above variables:
Z = 3/8 for Zone 2 (Albuquerque)
I = 1.00 for occupancy importance
K = 1.00 for structural type
CS = 0.14 for resonance type
W = weight of the structure
(1-8)
Using these values in the equation for base shear leads to a value of
V = 0.05W (1-9)
for lateral forces. Again, these values are derived for a location near
Albuquerque, New Mexico and show that seismic loads are not significant.
1.3.2 Load Combinations
Another important aspect of the Task I work was the study of load factors and
load combinations for the design criteria of the support systems. Ultimate
strength methods have been developed in the past decade for the economic de
sign of critical structural systems, based upon establishing factored load
combinations and factored member resistances. This is done by considering
the nature of the loads, thei r probabil ity of s imu 1taneous occurrence, and
the degree of confidence in specifying the loads. Likewise, member resist
ances depend on material properties, the ability to control those properties,
1-42
and tile properties and geometrical variations of the structural members
components.
The most recent work in load factor design is summarized in a National Bureau
of Standards Publication (Reference 1-13). Assumptions used in that publica
tion do not correspond to those used here for the design of solar array
structures. Helpful suggestions in attempting to apply load factors in
this work were given by Theodore V. Galambos, Professor in Civil Engineering
at Washington University, St. Louis, Mo. However, load factor design
applied consistently and uniformly with structures composed of various
materials such as steel, timber, and concrete is generally extremely difficult
and still under investigation and development.
Rather than developing consistent and uniform load factors for each different
material used in the solar array structure, modifications of the present Code
factors were eventually used for this design.
Structural Steel and Timber Structures. Steel and timber members were de
signed using allowable stress methods but using higher stress levels than
those normally speci fied by the codes (Reference 1-3 for steel and Reference
1-6 for timber). For the operating condition, a 10% increase in Code allowable
stresses was permitted for all loadi ng conditions. For the extreme case, a 47%
increase in Code allowable stresses was permitted (1.10 x 33%) for all loading
conditions. Given below are the recommended load combinations for steel and
timber:
Operating Basis Condition:
a) D + L •••••••••••• Stress Limit = 1.10 Fs
b) D + Wo •••.•••••••• Stress Limit = 1.10 Fs
1-43
c) D + S•.•••••.•••.. Stress Limit = 1.10 Fs
Extreme Environment Condition:
d)
e)
o + E....•••••... Stress Limit; 1.47 Fs(includes 33% increase)
D + We .•••••••••• Stress Limit = 1.47 Fs(includes 33% increase)
Where Fs is the allowable working stress defined by the pertinent code and Wis
the corresponding environmental wind load.
Concrete structures. The existing load factors for concrete members speci-
fied by ACI 318-77 (Reference 1-5) were all reduced accordingly to corres-
pond to the stress increases allowed for the steel and timber structures. The
same load combinations were used in the concrete design as those considered
in the steel and timber designs. Given below are the resulting recommended
load combinations for concrete:
Operating Basis Condition:
a) U = 1.3D + 1.5L
b) U = 1.3D + 1.55
c) U = 1.3D + 1.5Wo
d) U = 0.80 + 1.5Wo
Extreme Environmental Conditions:
e) U = 0.95D + 1.2We
f) U = 0.8D + 1. 2We
g) U = 0.95D + 1.3E
f) U = 0.8D + 1.3E
1.3.3 Foundation Criteria
For design of the 10 MW array fields, soil properties are assumed which are
1-44
rep resented by test ho 1e #1 as gi ven in Refe rence 1-14. These soil prope r-
ties are also the ones used for design of the demonstration arrays in Albu-
querque, New Mexico. The soil is a clayey-sand having the following properties
and allowable values:
~ Angle of Internal Friction 35°
w Moist unit wei ght of sand 116 psf
Kp Coefficient of Passive Earth Pressure 3.69
Ka Coefficient of Act i ve Ea rth Pressure 0.27
Qp Allowable end soil beari ngpressure for piers 3000 psf
For any combination that includesshort term loads like wind, snow,or 1i ve load 4000 psf
QL Allowable lateral soil pressure passive earth pressure
Criteria for the foundation design are given in terms of operating loads
and extreme loads.
Vertical Soil Pressure. The following values were used for vertical soil
pressures:
a) Operating loads shall produce net bearing pressures less than
allowable values. For combination D + L + Woo the net bearing
pressure shall be less than 1.33 x allowable values.
b) Extreme loads shall produce net bearing pressures less than 2
x allowable values.
Lateral Soil Pressure. The fol lowing values were used for lateral soil
pressures:
a) Operating loads shall produce lateral soil pressures less
1-45
than allowable values.
lateral earth pressure
values.
For combination D + L + Wo , the
shall be less than 1.33 x allowable
b) Extreme loads shall produce passive soil pressures
less than 2 x allowable values.
Uplift. To prevent foundation pullout due to uplift forces, the ratio of
downward resisting force to applied uplift force must be greater than 1.5
for all load combinations.
Torsion. To prevent foundation failure by torsion, the ratio of resisting
torque to app1i ed tors i ona1 forces mu st be greater than 1. 0 for all load
combinations.
Criteria for lateral soil pressure were developed by modification of a method
by Broms (References 1-15 and 1-16) for predicting the failure mechanism of
short (rigid) poles in cohesionless soils under lateral loads. The Broms
method resu Its in 1i ghter foundations than those developed usi ng the UBC de
sign formula. In this method, Broms assumes the ultimate resistance of later
ally loaded short piles is governed by the lateral resistance of the surround
ing cohesionless soil. It is also assumed that the ultimate lateral resist
ance of a cohesionless soil is equal to three times the Rankine passive earth
pressure. Calculated ultimate lateral resistances have been compared with
available test data and found to be in satisfactory agreement (References
1-15 and 1-17).
Shown in Fi gure 1- 20 are the assumed soi I react ions and method used in the
final designs of the wood pole and concrete caisson foundations. This method
for designing short rigid piles differs slightly from the Broms method in
1-46
M
('P
/r-7H / 10( FAILURE MODEI
/ // /
...'I' ~
/ // /
// /
/ /D / /
/ // /
/ /I... TOE REACTION
~ I· ·13lS DB Kp/FS =MAX SOIL REACTION
WHERE FS = FACTOR OF SAFETY AGAINST FAILURE
= 1.50 FOR EXTREME LOADS
= 2.25 FOR D + L + Wo LOAD COMBINATION
= 3.00 FOR ALL OTHER OPERATING LOADS
NOTE: THIS IS A NON-CODE DESIGN PROCEDURE USING MODIFIEDBROMS METHOD (REF. 1·15)
Figure 1-20 Assumed Soil Reactions Due To Lateral Loads For A Short Rigid Pile In Granular Material
1-47
two ways. First, a concentrated moment at the top of the pile has been added.
Secondly, a safety factor has been included in Brom's assumed soil reactions
to account for the different 10adings and environmenta1 conditions. Thus,
the derived equation for a 1aterally loaded pile becomes:
P(H + 0) + M 3= 0.5 Y BO KpFs
where Fs = Factor of safety against fa il ure
Fs = 1.50 for extreme condition
Fs = 2.25 for o + L + Wo load combination
Fs = 3.00 for al1 other operating 10ads
(I-g)
Using the above factors of safety is equivalent to deve10ping 2 x passive
earth pressure for extreme loads, 1.33 x passive earth pressure for the D +
L + Wo load combination, and 1.00 x passive earth pressure for a11 other
operating loads. The 1.50 factor of safety for extreme loads was used due
to the uncertainties in the soil properties and wind forces on the structure.
For the uplift calculations, frictional stresses S on the foundation surface
are assumed proportional to the horizontal earth pressure 0 H and the fric-
tional stresses S are given by
0 H = Ko Y Z (1-10)
S = F 0 H = F Ko 'i Z (1-11 )
where 0 H = hori zonta1 earth pressure
Y = density of soil
Z = depth below ground
Ko = coefficient of earth pressure at rest
S = frictional stress on foundation perimeter
1-48
F = coefficient of friction of sand
Ko depends upon the relative density of the sand with values which range
from 0.40 for dense sand to 0.50 for loose sand (Reference 1-18). For the
final designs, Ko was taken as 0.5.
F depends upon the type of surface contacting the sand. For sand against
timber, F is about 0.35 and for sand against concrete (rough masonry), F is
approximately 0.6 (see Reference 1-19). By integrating the frictional stress
S over the depth and perimeter of the foundation, the uplift resistance Ur
of the foundation due to soil friction is found.
where
Ur = 1/2 Ko F Y IT B 02
B = diameter of foundation
o = depth of foundation
(1-12)
The uplift resistance Ur assumes a granular soil and neglects any contribu-
tion due to soil cohesion. The total resistance of the foundation to pullout
would be the sum of the soil resistance Ur and the dead load on the founda-
tion.
Some research has been performed on the torsional resistance of foundations
(References 1-20 and 1-21) but work in this area has received little attention
in the past. Randolf (Reference 1-21) suggests that, for rigid or stiff piles,
failure in torsion will occur suddenly as the applied torque T approaches
the limiting value of
T = 0.5 1\ B2 0 T f (1-13 )
where T f = average available shaft adhesion (ignoring the contributionfrom the base of the shaft)
1-49
By usi ng an analogous procedure, the appl i ed torque Tis found by static
equilibrium to be equal to the resisting torque caused by the frictional
stresses S. Integrating along the length of the foundation:
oT = TI Bf S (B/2) dz
ousing Eq 1-11
(1-14)
0
T = 0.5 11 B2 F Ko Yf Zdz ,or (1-15)
0
T = 0.25 11 B2 F Ko Y 02 (1-16)
When Eq 1-16 is compared to Rando1f's formula, Eq 1-13, one finds the average
value of the frictional stress on the foundation is equivalent to:
T f = 0.5 f Ko Y D
1.4 DESIGN OPTIMIZATIONS
(1-17)
The preliminary designs were evaluated to obtain two promising support con-
cepts, the wood pole and the concrete caisson. Further work was performed
to achieve not only a low-cost array but also a structurally optimized sup-
port concept.
Four areas were examined to achieve minimum cost and an optimized structure:
0 Panel/module integration
0 Single versus double supports
0 Span length variation
0 Variation of support locations
These topi cs will now be discussed in greater detail.
1-50
1.4.1 Panel/Module Integration
Existing photovoltaic module types may be characterized as framed (with a
metal edge frame) or unframed (with an elastomer gasket around the edge).
This aspect of the module was not defined early in the study and there
fore both possibilities were considered in order to achieve lowest costs.
Telephone discussions were conducted with module manufacturers to explore
design possibilities for cost reductions by eliminating potential structural
redundancies. The design proposals for discussion were (a) that modules
be unframed and therefore be fully supported along all edges by the support
steel, or (b) that the modules be framed and therefore cantilever out from
an abbreviated support structure. The latter arrangement is not conventional
and was pursued as a cost-reduction possibility.
Shown in Figures 1-21 through 1-23 are various conceptual attachment schemes
of framed modules onto a torque tube having a single support. Figure 1-21
shows how a framed module might cantilever out from abbreviated tube members.
Each module is secured at four locations by 3/8 in. diameter bolts which have
rubber washers to spread the concentrated loads.
Fi gure 1-22 shows how the same modules may be conti nuously secured along
the support strips instead of being bolted at four discrete locations.
Figure 1-23 shows how the modules could be secured on abbreviated tee and
angle sections as preViously shown in Figure 1-10. Here, slots are cut into
the projecting webs of these sections where small tee wedges are inserted
to secure the modules. Friction forces between the lower edges of the framed
modules and the tee or angle sections help secure the modules.
1-51
FR
AM
ED
MO
DU
LE
SA
TT
AC
HT
OP
AN
EL
BY
BO
LT
ING
TO
AB
BR
EV
IAT
ED
TU
BE
ME
MB
ER
S4'
x.'
CA
NT
ILE
VE
RE
DF
RA
ME
DM
OD
UL
E
~3/
'"01
....
I!MJL
T
~R
UB
BER
WA
SHER
-...~-..,.
~
50
4'·
20
'
CO
NC
RE
TE
CA
ISSO
N
----
----
-._
-
I I I I , I I , , I i_
, , , I I I I I I I I I I I :I
I
:I-
:I
,I
II
I
:'
I\-
fI
'-~.,/
r..... I <.
TlN
Figu
re1-
21F
ram
edM
odul
eS
uppo
rtC
once
ptA
FR
AM
ED
MO
DU
LE
ED
GE
SU
PP
OR
TE
DB
YC
HA
NN
EL
SM
IDT
UB
ES
CH
MIN
EL
SA
BO
VE
~ , lJ'
W
.~
..,
'"
.~"'f--
_
-S
QU
AR
ET
UB
EB
ELO
W
l
Figu
re1-
22Fr
amed
Mo
du
leS
up
po
rtC
on
cep
tB
FR
AM
ED
MO
DU
LE
SS
EC
UR
ED
TO
PA
NE
LB
YT
EE
WE
DG
ES
TE
EW
ED
GE
INS
ER
TE
DIN
TO
SLO
TS
OF
TE
ES
AN
DA
NG
LE
S
......
j <n
.."
,.
II
II
II
I,
II
Ii
I
iI
II
II
II
II
II
II
II
,,
:J
,-
~~.
AN
GL
E
•?··
";;::-,"
'+C,
,~~
.
Figu
re1-
23Fr
amed
Mod
ule
Su
pp
ort
Con
cept
C
These suggested attachment schemes were presented to module manufacturers for
their comments. Contacts with two manufacturers revealed our suggested
schemes were too advanced for thei r complete approval. They wanted to adopt
the current practice of supporting the modules fully on all four edges.
Since frameless photovoltaic modules are within the realm of current tech
nology, the array panel was mOdified accordingly to accept that type of
module. This would be one necessary step towards fully integrating the
module with the support structure.
Shown in Figure 1-24 is a modified 8' x 20' panel where frameless modules
are slid into the channel sections for support on three edges, with the
fourth channel attached later by bolting a separate piece on that edge.
The channels are cold-rolled sections and are available as stock items from
several manufacturers (e.g., Superstrut and Uni strut). The module has an
extruded section around its edge which provides the necessary retention of
the module to the support structure. It was i ndi cated that thi s type of
module/ structure interface could result in minimal labor for installation of
the modules to the support structure.
The 8' x 20' panel itself (f.o.b. factory) was estimated to have direct
costs of $30.60/m2 based on a 10 MW field installation. With the total in
stalled cost goal of Sandia set at below $40/m2 , this type of panel for frame
less photovoltaic modules would make the total cost of the array installation
too expensive under present guidelines.*
Therefore, with costs unacceptable for the panel concept supporting frameless
modules, the focus of panel optimization was re-directed towards the original
abbreviated tee and angle supports for framed mOdules. After optimizing the
*Second quarter 1981 $ are used in this Report.
1-55
2'%"
20' 8}:;"- •2)t,"_ 20'-3%"
4'-15/8" 4'·0" .1 4'·0" 4'-0" 4'-1 5/8"
i-.i I -A
,
Co-M0 ,0;-M ,
- - -- -- - - -- - - - -,
Co-M0-M 8 B,
-,
V 'p AN
8' x 20' PANEL PLAN
Figure 1-24(a) Conceptual 8' x 20' Panel for Frameless Modules
1-56
3'-10%"
FRAMELESSPHOTOVOLTAIC
/ MjDULES
f ~~·cE·
---'-'--- 1.SECTION A
, '
3'-10%"
UNISTRUT-TYPE CHANNELS
RETAINER GASKET
1..'0((---6 x 6 x 3/16 TUBE
LIFTING LUG
4'-0" TYP. FOR 31NSIDE PANELS 4'-1 5/8" TYP. FOR 2 END PANELS
RETAINER GASKET
/------1--
SECTION B
Figure 1-24(b) Sections For Conceptual Panel
1-57
structure, i.e, span length, this type of support concept for framed modules
was estimated at having direct panel costs (foo.b. jobsite) ranging from
$18.2Om2 to $22.00/m2 (see Section 3.2). These costs were found more acceptable
and this panel support for framed modules was adopted in the final designs.
Details of the panel are further discussed in Section 1.5.
Even though structural redundancy between the module and panel was practically
eliminated using frameless modules, the total cost of the support structure was
not minimized. By using frameless modules, the panel requi res an increased
st i ffness to compensate for the loss in st i ffness of the modu 1es when thei r
individual frames are removed. By increasing its stiffness, the panel weight
was increased which, in turn, increased its total cost.
To evaluate the total system tradeoff, the design details and costs of particular
framed and unframed modules must be included. Since this was beyond this
scope of work, no conclusions could be reached concerning whether framed or un
framed modules achieve mimimum total system costs. One can only conclude that
support structure costs are lower using framed modules instead of frameless
ones.
1.4.2 Single Support versus Double Supports
Array support concepts both from the earlier Bechtel study (Reference I-I) and
for the conceptual designs discussed in Section 1.2 used single pedestal sup
ports. The reasoning behind using the single support has already been discuss
ed in Section 1.1.
At that time in the design process, information was not available concerning
these effects:
1-58
o Torsion forces on the foundation supports
o Natural frequency and mode shapes of the structure
Upon recei pt of result s from t he wi nd-tunnel tests of Colorado State Un i vers i ty
at Fort Collins (Reference 1-12), foundation analyses revealed that torsion
loads, rather than uplift or lateral forces, governed the depth of the single
support foundation. In one case, the depth of a single-support concrete
caisson had increased as much as 3 ft over a double-supported design which
had required consideration of only uplift and lateral loads.
Dynamic modal analyses were performed for the single-support and double
support designs to determine natural frequency and mode shape information
important to the dynamic analysis. The dynamic analysis revealed the single
support design the most vulnerable to dynamic wind effects. For this reason
a dynamic turbulence factor K of 3 was chosen for the single-support design
and a factor of 2 for the double-support design. Details of the dynamic
analyses are described in Section 2.
In any particular row, doubly supported arrays requi re an extra foundation
at one end compared to a row of single-supported arrays. However, in a long
row of arrays, the extra depth requi red for single-support foundations to
resist torsion failure costs much more than the extra foundation required at
the end of double-supported arrays. In addition to deeper foundations, sin
gle-support designs are expected to have a 50 % increase in the dynamic
turbulence factor K. For these reasons, the doubly supported arrays were
chosen in the final design over the singly supported arrays.
1.4.3 Span Length Variation
The third area examined for possible optimization was to vary the span
1-59
length of the arrays. Consequently, the span length of the baseline 8' x 20'
array was changed from 20 ft to 32 ft, 36 ft, and 40 ft. Span 1engths
chosen for the optimizing process were multiples of the 4-ft module size.
Design criteria used for those designs were identical to the final design
criteria described in Section 1.3. Since larger torsion forces would be
associated with longer span lengths, only end-supported arrays were considered.
These designs were cost estimated using the assumptions given in Section
3.3. Plotted in Figure 1-25 are the total installed direct costs for the
arrays having the various span lengths for the extended concrete caisson and
timber pole support concepts.
This plot shows the timber pole costing slightly higher than the extended
concrete caisson. Costs were highest for the 20-ft span array and lowest
at the 36-ft span and then increasing for the 40-ft span.
Therefore, the 8' x 36' array was chosen as havi ng the lowest
di rect cos ts of $34.00/m2 for the extended concrete ca i sson concept.
down of these costs are given in Table 3-6.
installed
A break-
1.4.4 Variation of Support Locations
The final stage of optimization was to take the optimum 8' x 36' end-supported
arrays and move the end supports so as to reduce bending moments and deflections
in the torque tube. With reduced bending moments, a lighter section could be
used for the torque tube and, likewise, reduce the superstructure cost.
Thus, a double-supported cantilever system was chosen in the final design for
the arrays as shown in Figure 1-26. Here, foundations are still spaced 36 ft
1-60
TO
TA
LIN
ST
AL
LE
DD
IRE
CT
CO
STS
IN$
/M2
IBA
SE
DO
N10
MW
FIE
LD
SIZ
EA
ND
MID
·198
1$)
,T
IMB
ER
PO
LES
UP
PO
RT
40
38.4
, '" -a:
35
w .... w ;;; w a: <>: :::> o ~30
37.3
1)
36
j
EX
TE
ND
ED
CO
NC
RE
TE
I/
CA
ISS
ON
34.0
OP
TIM
UM
SP
AN
FOR
TH
ES
ED
ES
IGN
S
oI
IfI
II
Io
20
SP
AN
LEN
GT
H,
FE
ET
3236
40
figu
re1-
25T
otal
Inst
alle
dD
irec
tC
osts
Ver
sus
Span
Len
gths
SU
CC
EE
DIN
GP
AN
ELS
CO
NN
EC
TT
OP
RE
VIO
US
ON
EA
ND
NE
XT
CA
ISS
ON
\ \
----
.-
3R
D
36
'IT
YP
)
SL
IDIN
GC
ON
NE
CT
ION
(TY
P)
(SE
EF
IGU
RE
1-34
)
8'
X3
6'
PA
NE
LIT
YP
)\
1ST
hi
\2N
D
24'
6'
\ \ \ \ I t \
-----l-J
I'III
Ii'Iii
--[--
~
, '"N
DIR
EC
TIO
NO
FE
RE
CT
ION
AL
ON
GT
HE
RO
W
Figu
re1-
26T
ypic
alR
owo
f8
'x3
6'A
rray
sU
sing
aD
oubl
e-S
uppo
rted
Can
tile
ver
Sys
tem
apart with the exception of the fi rst end array where two foundations are
spaced 24 ft apart. These two end foundati ons provi de a stable system to
receive the first panel. Then the end of any succeeding panel in a row is
placed on the previously installed one. This facilitates the erection pro-
cess.
In the revised support arrangement, the torque tube of each panel cantilevers
6 ft from each support with neighboring tubes attached by a sliding connec
tion. This connection is such that no bending moments are transferred between
the tubes. Details of this connection are given in Section 1.5.3.
This cantilever support system arranges that the maximum bending of the torque
tubes under uniform loading is approximately two-thirds of that found in an
end-supported system. Likewise, the maximum deflection of the tubes under
un i form 1oadi ng is about one-half of that found in the end-supported case.
With deflect ions and bendi ng moments reduced by the cantil ever system, a
1i ghter torque tube coul d be chosen over that used in the end-supported system.
This further reduced the installed direct costs of an interior 8' x 36' array
having the extended concrete caisson from $34.00/m2 to $32.60/m2• A breakdown
of these costs is further explained in Section 3.3 and in Tables 3-6 and 3-7.
Therefore, the optimized double-supported cantilever system for the arrays
was adopted as the final design of the arrays. Further details of this
design are given in tne next section.
1.5 SELECTED FINAL DESIGNS
After the optimizing process, the selected design was the 8' x 36' array
having a cantilever support system and having the extended concrete caisson
1-63
foundation since this had the lowest cost. The timber pole concept is given
here as an alternative which was found to be cost competitive. Figure 1-27
shows a typical row of interior 8' x 36' arrays using the double-supported
cantilever system. An interior array is specified here since they have the
lightest forces and thus the Ieast wei ght and lowest costs (see Sect i on 5
for differences between perimeter and interior arrays). This figure is
referenced periodically to show details of the following:
o Superstructure
o Substructure
o Connections
1. 5.1 Integrated Superstructure Design
The superstructure consists of a completed 8' x 36' panel (including the in
stalled PV modules) which can be transported to the jobsite as an independent
unit for installation onto the foundation supports. A typical panel shown in
Figures 1-28 and 1-29 consists of a longitudinal torque tube (6' x 4" x 3/16"
tube) having abbreviated 4-ft long tee sections (MT 4 x 3.25) for cross
members which di rectly support the PV modules. The tees are fillet weI ded
to the upper surface of the tube.
The PV modules are conceptualized here as 3' 11-3/4" square framed modules.
These attach to the tees by bo1ting each modu1e frame at two locations along
the web of the tees as shown in Figure 1-29. The modules then cantilever
about 2 ft beyond the end of the tee support.
These modules are not unique; however, any variations to this module can be
readily adapted for attachment to the tee sections. For convenience, the
framed modules assumed here were identical to the dummy modules postulated
1-64
SO
UT
H 1~
,"o;o
vo,",
,,;oo
o,,"
---..
8"[~
36
'-0
"
_J,
i'
tB3
6'-
0"
(TY
P)
PLA
N
>-' I en on
DE
TA
IL2
(SE
EF
IGU
RE
1-34
)E
XT
ER
IOR
AR
RA
YW
ITH
35
0~
TW
OC
AIS
SO
NS
HIN
TE
RIO
RA
RR
AY
S~
ITYP)~
DE
TA
IL1
(SE
EF
IGU
RE
1-32
)
I'L:
T I1-----~.
:11-
5'-0
-tt=
;l'tt
'\I
I<
0I,
II
I::
"I,
..':1
1(T
YP
)l
tCX
I1·
1co
,I
.JI
I:,
:L
,
0(•
6~2
4'-
0"
k-3
6'-
8"
(TY
P)
)0
Tqjl
',
,I
ITm
L:J
II
II
II
I,IT
FL;JJ
jI
II
Ir.
3'-3~UTH
:1:
/~E
LE
VA
TIO
N1
6"
\lC
ON
CR
ET
EP
OS
T
••S
EE
FIG
UR
E5-
1FO
RL
OC
AT
ION
OF
INT
ER
IOR
AR
RA
YS
WIT
HIN
AR
RA
YF
IEL
D
Figu
re1-
27Ty
pica
lR
owo
fP
roto
type
Arr
ays
-----------'=I~
I~~l
~ >- S ~ N
3'·1
1jI
,"S
QU
AR
EF
RA
ME
OP
HO
TO
VO
LT
AIC
MO
DU
LES
3"
(TY
P.)
c
I'~--~--+---
IS"
(TY
P.)
3'·1
1jI
,"(T
YP
.)
......
I C1>
C1>
24
"(R
EF
.)6
"x
4"
x%
"L
SU
PP
OR
TC
ON
NE
CT
ION
TO
CA
ISS
ON
6"
x4
"x
3/1
6"
N.T
.S.
SE
CT
ION
A-A
(SE
EF
IGU
RE
1·27
)(P
AN
EL
SH
OW
NH
OR
IZO
NT
AL
)
35°
Figu
re1
·28
Att
ach
men
to
fP
ho
tovo
ltaic
Mod
ules
toP
anel
Tees
~ I '"....,
TY
PIC
AL
MT
4x
3.2
5M
OD
UL
ES
UP
PO
RT
AR
MS 4
'-0
"(T
YP
.I
3'-1
1'Y
.,"S
QU
AR
EF
RA
ME
DP
HO
TO
VO
L TA
lCM
OD
ULE
SH
IMA
SR
EQ
'D.
%"0
BO
LT
v:
1/II
'\
,/I
?if
\if
/t/
6"
x4
"x
3/1
6"
TU
BE
SE
CT
ION
B-B
(SE
EF
IGU
RE
1-27
)
Figu
re1
-29
Att
ach
men
to
fP
ho
tovo
ltai
cM
odul
esto
Pan
elTe
es
for the demonstration arrays. For certain locations in a large array field,
changes in the torque tube size were necessary to reflect variations in the
wind forces over the arrays. As mentioned before, a 6" x 4" x 3/16" tube
section was used for a typical interior row. However, a 6" x 6" x 1/4" tube
was found necessa ry for each end array of any i nteri or row to refl ect 1a rger
wind forces at those positions.
Likewise, the 6" x 6" x 1/4" tube can be used throughout the extreme north and
south row of arrays where larger wind forces are expected. The MT 4 x 3.25
tee section can be used throughout the array field for support of the photovol
ta ic modul es.
1.5.2 Substructure Designs
The extended concrete caisson foundation was the preferred design which
achieved the lowest cost. The caisson is 16 in. in diameter and extends
slightly over 3 ft above the ground surface. In a large array installation,
removable forms may be used to cast this 3-ft extension.
The slightly-more-expensive timber pole design may be used as an alternate but
the availability of sufficiently large quantities would require extra consid
eration. For this concept, a hole is augered and the pole inserted and back
filled with suitable material, e.g., pea gravel. The timber pole shown in
this work is 12 in. in diameter after optimization of the design.
As was the case for the torque tubes, variations of wind loads throughout a
large array fi el d requi re di fferent embedment depths of the foundations.
Shown in Fi gure 1-30 are four different types of cai ssons whi ch have been
standardized for a large array field installation.
For example, Type 1 caissons are typical foundations designed for an interior
1-68
WE
ST
EN
DO
FF
IEL
DN
OR
TH I
EA
ST
EN
DO
FF
IEL
D
~
, '" <.0
\N
OR
TH
PE
RIM
ET
ER
RO
W,
AL
LT
YP
E4
CA
ISS
ON
S
~r0
0I
0ci
f0
I0
I0
I0
I0
I
1ST
2W
ES
TE
ND
,T
YP
E3
CA
ISS
ON
S1S
T2
EA
ST
EN
D,T
YP
E3
CA
ISS
ON
S
~~
,~
,1
00
1€
_o
l01
~I01
01
[TY
PIC
AL
INT
ER
IOR
RO
WT
YP
E1CAI
SSON~E
STOF
INT
ER
IOR
RO
WS
I00
I0
l=til
0I
0I
0I
0I
0I
!L/
LS
OU
TH
PE
RIM
ET
ER
RO
W,
AL
LT
YP
E2
CA
ISS
ON
S
CA
ISS
ON
DE
PT
HS
TY
PE
S
15
'-0
"
26'
-6
"
38
'-6
"
41
1'-
0"
NO
TE
:C
AIS
SO
ND
EP
TH
SD
ER
IVE
DF
RO
MN
ON
-CO
DE
DE
SIG
NP
RO
CE
DU
RE
US
ING
MO
DIF
IED
VE
RS
ION
OF
BR
OM
S(R
EF
ER
EN
CE
1-15
)
Fig
ure
1-30
Em
bedm
ent
Dep
ths
Of
Cai
sson
sIn
Pro
toty
pe
Arr
ayF
ield
row of arrays. These caissons experience the li ghtest loads and have the
least embedment of 5 ft. The first two caissons at each end of each interior
row of arrays experience higher wind loads due to being on the perimeter of
the array field. These are labeled Type 3 caissons having an embedment of
8-1/2 ft. The extreme south perimeter row of arrays experience higher
wi nd forces than a typi ca1 i nteri or row and these foundat ions, Type 2,
have an embedment of 6-1/2 ft. The deepest caissons, Type 4, are on the
extreme north perimeter of the array field and have an embedment of 11 ft.
Lateral loads governed the design of the Type 1 and Type 2 caissons while
up 1i ft forces governed the des i gn of the othe rs. Type 1 cai ssons represent
the majority of foundations in a large array field and was therefore the one
used in the final cost estimates of Section 3.
Shown in Figure 1-31 is a typical Type 1 concrete caisson. Two anchor bolts
are butt welded to the 1ongi tudi na1 rei nforcement and extend above the
top of the concrete for connection to the panel. These bolts may be held
in proper alignment by a simple jig while the concrete hardens in the forms.
Details of this connection are given in the next section.
If the 12-in. timber pole is used as an alternate design, different embedment
depths are required in the same locations in the array field than the concrete
caissons. Shown in Table 1-2 are the concrete caisson depths labeled accord
ing to type with equivalent depths of the timber pole. Uplift governed the
depth of all the timber poles since they do not have as much pullout resistance
as the caissons. Again, the Type 1 wood pole was used in the final cost
estimates given in Section 3.
1-70
#3TIES
10"
~
"-~I- '"r
I I
~I
"'~ I I
"I I en
~"-
'?00
0'7
"'-'--.... -f "c--... ,- '----- 0-~ '\r-- ::::'r-(_::::: ~
~1\- I--- r--., '\
: I'--. "4.u;
"' r.:'" z- ;,.
%" £l ANCHOR BOLTS
Co I'00 ;:, I I
to1 I I'I I
BUTT WElD REBAR& ANCHOR (TYP.)
# 7 LONGITUDINAL BARS
(DEPTH SHOWN FOR END OF
INTERIOR ARRAYS, FIGURE 1·30)
Figure 1-31 Extended Concrete Caisson Foundation for Prototype Array Field
1- 71
Table 1-2
REQUIRED EMBEDMENTS OF THE TIMBERPOLE AND CONCRETE CAISSON
Required Embedment in Feet
Type* Concrete Caisson Timber Pole
1 5.0 9.0
2 6.5 11.5
3 8.5 15.5
4 11.0 18.0
*Refer to Figure 1-30 for location in array field
1-72
1.5.3 Connections and Details
Figure 1-32 shows the connection between the torque tube and the extended
concrete caisson for a typical interior array. A 6" x 4" x 1/2" angle is
shown cut and fillet welded to the 6" x 4" x 3/16" tube. This angle provides
a seat for the torque tube and is attached to the top of the cai sson by two
3/4-in. anchor bolts. Before the concrete has set, the top of the caisson
may be screeded to form a smooth level surface. This would eliminate the
need of costly grouting between the seat angle and the caisson. Shims may
be used for any necessary adjustments. For the larger (6" x 6" x 1/4") torque
tube, the connection uses the same seat angle and is welded in similar manner.
Figure 1-33 shows how the seat angle support may be adjusted to the timber
pole. For this connection two opposite sides of the pole are assumed pre
shaved at the mill to provide flat surfaces for the placing of two metal
plates. These plates are attached by two bolts which extend through pre
drilled holes in the pole. These plates have 3/4-in. threaded rods welded to
them to which the seat angle is bolted. This connection may be used through
out the array field using timber poles.
Figure 1-34 shows a typical connection between torque tubes. Here a 6" x 4"
x 3/8" angle is fillet welded to one tube and provides a seat for the instal
lation of the adjacent tube. The adjacent tube has two predrilled holes
through which the tube is attached to the seat angle by two 5/8-in. bolts.
Slotted holes in the seat angle allow for any longitudinal adjustments re
qui red duri ng the i nsta 11 at i on of the torque tube. The connecti on between
the 6" x 6" x 1/4" tube is similar to this one.
1-73
SYM. ABT.
/MT 4 x 3.25
6x4x3/16TUBE.
I
tCONC. POST
2"
9/160 HOLES
7"
16X4XlHJ,- \
,
~-~
4" rftr , --L ,. ..,.,. , ~....:
\- -it--
1-:--3"
10"
DETAIL 1(SEE FIGURE 1-27)
Figure 1-32 Connection of Torque Tube to Extended Concrete Caisson
1- 74
7/8
"~
HO
LE4
"x
6"x
%"
L
-t;f~,
:t!L
.---
--.-
-
7"
x5
"x
3/1
6"
TU
BE~
Co
.~
III':
:
~
IM
T4
x3.
25
----
>' ou..
LU a:
...
'i-iI7
-
%"~EADEI
RO
Dn
~~
~~--+i
(.
u. w a: N
>-' , ...., '-"
u. w a: >' N
"A~;;''
'[It:=,:
=~-:-
:0:'1
"1rJ
~15/
16"H
OLE
ST
HR
UP
OS
T
~.:k-----
.-=-=
-L
.rr-
----
-----
u:
,I
---..j
_.....--
-----
----
'1
2"
IIIT
IMB
ER
PO
LE
Co ;" ;;.
N
15
/16
"D
IA.
HO
LES
Fig
ure
1-3
3C
on
nec
tio
no
fT
orq
ue
Tu
be
toT
imb
erP
ole
5/6
"0B
OL
T
E
3"
xII
"P
LAT
E
r--
8x
4X
3111
•
I--.I
--.
::.....
.. 6"
~-
--=
-'
.
% '%
%"x
LIF
TIN
GLU
G&
TE
EN
OT
SH
OW
NFO
RC
LA
RIT
Y
6"6
"
DE
TA
IL2
(SE
EF
IGU
RE
1-27
)
Figu
re1-
34T
ypic
alC
onne
ctio
nB
etw
een
Adj
acen
tTor
que
Tub
es
II 1,
14 '3 11
__
.LI
_
~
00
.,..
3%"
1%"
,\ I
---
::J
::fWIJ
.U-
--
W.1
.lJ.
~Q
.
Jr:~
~r
1'--
:-"7
-
IT
T~
I
I•
\J
+"
-,-,
..:l
-+<
••••
~
"'-00 in
z6
x4
x3
/8L
......
I ..... en
More detail of the final design of the array superstructure, substructure,
and connections may be found in drawings given in Appendix A.
1.6 ARRAY FIELD FENCE STUDY
From the Phase II wind tunnel study (Reference 1-12), it was determined the
use of fences drast i ca lly reduces the wi nd loads on the peri meter arrays but
remains ineffective for the interior arrays (See Section 5). Therefore, a
study was performed to see if the use of fences would be economically benefi
cial in a large array field.
Since fences affect the forces only on the perimeter rows of an array field,
estimates were made of the increase in steel weight/linear foot (measured
along the perimeter array) and the increase in foundation depth (also measured
along the perimeter array) if fences were not used. This required doing com
parative array designs with and without fences. Using the panel design de
scribed in Section 1.4.1 for frameless modules, it was thus estimated that
the torque tube weight would increase about 2.73 lb/linear ft and the frame
member wei ghts woul d increase about 0.264 1b/l i near ft if wi nd fences were
not used. Likewise, the caissons would requi re an increase in depth of
about 0.05 ft/ linear ft along the perimeter row.
With the above increases in material quantities, using the bases and assump
tions of the cost estimates of Section 3, it was estimated that perimeter
array costs would increase about $4.00/linear ft compared to not using wind
fences.
Checking with vendors who supply fencing, the most common type of wind bar
rier would be an 8-ft chain-link fence with thin wood slats inserted between
the links. This type of fence would cost approximately $8.00/linear ft
1-77
more than a common security fence which would otherwise typically be required
around an array field. Even though a wind barrier fence would, at most,
save $4.00/ li near ft in the array structure costs by lowering the wind
forces, the overall effect of using fences for wind protection would raise
the balance of costs about $4.00/1inear ft. Therefore, it was concluded at
this point that the use of fences for wind protection of array fields would
not provide a cost savings and their use was not investigated further.
1-78
Section 2
DYNAMIC WIND ANALYSIS - TASK II
2.1 GENERAL
This section reports the accomplishments and results of the Task II investi
gation on wind dynamic analysis of photovoltaic (PV) array structures. The
dynamic analysis was required for predicting structural response to the fluc-
tuating component of wind load. Prediction of the response to the steady-state
component was obtained by a static analysis that assumed quasi-static loads
loads. The objectives of Task II were to:
o Provide necessary data input on dynamic (modal) properties of PVarray structures for dynamic design evaluation
o Demonstrate an analysis methodology which could be used to determi ne the resonance and fati gue effects of array structures underwind excitations
o Evaluate dynamic results and estimate dynamic load factors (orturbulence factors) for comparison with those used in the designof PV array structures
o Draw conclusions and recommendations; identify important parameters and unresolved issues
2.1.1 Basic Approach
The dynamic analysis procedure adopted for this Task II were:
1. Establish dynamic wind load characteristics in terms of intensityand frequency contp~t
2. Develop mathematical (finite element) models of the array structura 1 systems
3. Perform modal analysis to extract natural frequencies and modesof vibration
2-1
4. Investigate possible instability problems due to interaction ofst ructura1 response wi th wi nd loads
5. Perform dynamic structural response analysis using load andmodal information from Steps 1 and 2.
The fluctuating wind load characteristics in Step 1 were based on the measured,
non-steady wind pressure data on similar PV array models reported by Boeing/
CSU (Reference 2-1). Since these wind pressure data were in the form of
power spectral densi ty (PSO) functions, a stationary, random-frequency re-
sponse analysis using a PSD analysis routine called OYNRE3, was selected in
Step 5. DYNRE3 is a routine in the STAROYNE program (Reference 2-2). The
required structural admittance functions (transfer functions) for this anal
ysis were established from structural modal information previously obtained
from the STAROYNE eigenvalue analysis routine. The analysis results (forces
and displacements) are in terms of root mean square (rms) values.
Pri or to the pub1i cat i on of Refe rence 2-1, two diffe rent approaches were
tried to predict the structural response to the fluctuating wind load.
The first approach is the methodology proposed by Cervallos-Candau and
Hall (Reference 2-3), which uses the measured fluctuating component of the
wind velocity reported by Oavenport (Reference 2-4) as the dynamic load input.
A deterministic response spectrum analysis, similar to that commonly used in
earthquake engineering, is employed to predict the peak structural response.
Another approach follows the wind-load calculation procedure specified by
ANSI A58.1-1972 (Reference 1-11). The so called "Gust Response Factor" of the
wind load is expressed as a function of structural frequency so that the
dynamic amplification associated with a gi ven structural frequency can be
readily determined. Since both methodologies are intended for predicting
2-2
response of tall structures subjected to atmospheric gust loads, their appli
cat i on to the PV array st ructu re subjected to ground-l eve1 wi nd tu rbu 1ence
is not clearly valid. Therefore no further studies were made in this di
rection.
The i nstabil ity problems cons i de red inStep 4 were those that mi ght be
associ ated with sta11 fl utter (Reference 2-5) and possi b1e resonant vi bra
tions caused by alternate vortex shedding in the wake of flat-plate PV
structures (Reference 2-6). The risk of these instability problems was
assessed using simple calculations based on the previously-determined struc
tural modal information.
2.1.2 Scope of Investigation
The complete dynamic response analysis was demonstrated for the single-post
array design which was found most susceptible to wind dynamic effects. The
dynamic analysis of both the single and double-post design concepts was
limited to modal analysis and dynamic instability evaluation.
The wind condition analyzed for the single-post design corresponded to a
wind speed of 80 mi/h in the S-W direction incident upon the south perimeter
row of a large array field. This wind condition provided one of the prelimi
nary loadings which governed the final design of the array structures.
2.2 MATHEMATICAL MODELS OF THE PHOTOVOLTAIC ARRAY
Two basic mathematical (finite element) model configurations were developed
for the dynamic analyses. They corresponded to early versions of the single
and double-post des i gn confi gurat ions. Both confi gurati ons used the 16-i n.
diameter extended concrete caisson and an 8' x 20' panel for frameless modules
2-3
as described in Section 1.4.1 (see Figure 1-24). Differences in various
si ngl e- and doubl e-post desi gn concepts were accommodated by assi gni ng member
properties to these two basic model configurations. The definitions of
the two basic configurations are described in Figures 2-1 and 2-2, respec
tively.
Both models were coded according to the specifications of the STARDYNE pro
gram. The array superstructures (panels) of the two models were identical
and both idealized the baseline 8' x 20' arrays. Since identical superstruc
tures were used, element properties were the same for the two design configura
tions. Ten 4' x 4' modules of the 8'x 20' array were each modeled by four
plate elements connected to the adjoining module supports modeled by beam
elements. The torque tube and the supporting posts were also modeled by
beam elements. The posts were assumed to be fully fixed at the ground surface.
Thus, the effects of soil foundation flexibility and the contribution of soil
damping to energy dissipation were not investigated in this work.
To account for the reduced stiffness of the concrete posts due to cracking,
the ACI-suggested formula was used for estimating the cracked-section flex
ural moment of inertia. (Section 9.5.2.2, ACI 318-77, Reference 1-5). The
torsional moment of inertia was taken as twice the computed flexural value.
The mass of the array structures was 1umped at the node poi nts of the model.
The element mass was computed from mass (weight) density information and
evenl y di stri buted to the nei ghbori ng node poi nts. The node poi nt mass was
assumed effective in all three translational degrees of freedom.
2.3 ANALYSIS FOR STEADY-STATE WIND LOADS
Since the steady-state component of wind load remains constant with respect
2-4
20'
I"6
1116
2126
3136
4146
5~T
YP
ICA
L4
'x4
'G
LAS
SM
OD
UL
ER
EP
RE
SE
NT
ED
YP
LA
TE
EL
EM
EN
TS
N J U1
MA
INT
OR
QU
ET
UB
EU
SIN
GB
EA
ME
LE
ME
NT
~V
,7
121
722
2732
3742
4I
-"""I
813
1823
28
3338
4348
53t-
-J.
...-......
I9
1419
2429
3439
4449
54
l
;10
1520
2530
3540
4550
55--
l'.-
VIE
WA
-A
PA
NE
L
TY
PIC
AL
MO
DU
LE
SU
PP
OR
TS
US
ING
BE
AM
ELE
ME
NT
S
A\>
-
2.29
'
4.00
'
I"3.2
8')
28
PA
NE
L
~_
PE
DE
ST
AL
1.71L.
--
(BE
AM
S)
>7>77J~77
7!
\>-(
Fig
ure
2-1
Fin
ite
Ele
men
tM
odel
of
Sin
gle-
Pos
t8
'x2
0'A
rra
y
N,-
... 1",
... POST
\'" \, ..:J r -,
\~ •. \r':",d
~
- .s~...Y
- ",
,"'"
'I '~ ..
"\~.l<,, :",," - "j~ .,~.,,\:, - ",r~ . -..rl;...",-"'" '-
, ". ,{"
,
, ., ~ -,.
J
\ ',,\.'
, .
- I' " .., ~,
,.I.
"II. ' .~, • , ~. ' .. , .,'
" ','" u.,.i , , :I~.....
.. , ,\" -
I, ~
. Iv. .
POST---1
Figure 2-2Isometric Model Configuration of Double-Post 8' x 20' Array
2-6
to time, structural response could be predicted based on a static analysis
which considered the steady-state wind loads as quasi-static loads. This
static analysis was performed for the single-post design to provide a base
line for estimating the dynamic amplification, or the turbulence factor, in
Section 2.6.
The appli ed loads were derived from the Phase II wind tunnel test results
for the south peri meter array subjected to the S-W wi nd (Sect ion 5). The
loads were transformed to linearly distributed normal pressures applied on
the su rface of the array panel as shown in Fi gure 2-3. The actual 1oadi ng
used in the STARDYNE static analysis was a further-simplified version consist
ing of uniform pressure blocks on each PV module as shown in Figure 2-4.
The static analysis results at critical points of the array structural members
are summarized in Table 2-1. These values also were in reasonable agreement
with hand calculations for the S-W wind which were used to design the single
post array structures.
2.4 MODAL ANALYSIS
Information regarding modal frequency and shape for various PV array design
concepts was obtained using a standard eigenvalue analysis routine (the House
hold Q/R method) in the STARDYNE program. This information was essential in
put to the dynamic instability evaluation (Section 2.5) and dynamic response
analysis (Section 2.6).
As a direct result of modal analysis, some design modifications were immediate
ly introduced for the designs. These modifications were primarily strength
ening the structural stiffness. For instance, for the single-post design, a
2-7
N
. - ~(-, .
ARRAY PANEL
/
- . -
0.1433 psi
. ".' -
--------.. ).', .k' -'r : 'I 0- ,: - 'J' k' -( - ,/' _" ... )~...., I
, II, ,D. - ,I,[
.-\ ( .'f>
{' " I '..~. -. {, r. .-
-•• J
0.0567 psi (6 Pmean )
~~===:zJ0.01067 psi
,,(,11, .\, " ,"
, ,
.....
0.0277 psi
NOTE: THIS ASSUMES 80 MPH WIND AT 10-METER HEIGHT USING PHASE II WIND
TUNNEL TEST RESULTS
Figure 2-3 Steady-State Wind Pressure Loading for Extreme SoW Wind
2-8
.. )(\
Figure 2.4 Idealized Steadv·State s-W Wind Pressure Loading
2-9
Table 2-1
STRUCTURAL RESPONSE RESULTS - STEADY-STATE WIND LOADS,S-W WIND ON SINGLE-POST DESIGN
Structural Response STARDYNE Static Value CalculatedMember paramet~l Analysis Results For Des i gn
(maxium
Torque Tube Yawi ng Moment 4,894 ft-Ib 5,222 ft-Ib
Di spl. 0.41 in. 0.41 in.
Module Support Pitching Moment 427 ft-Ib 555 ft-Ib(Cross)
Module Support Yawi ng Moment 40 ft-Ib 51 ft-l b(LongitUdinal)
Post Pi tch i ng Moment 2,544 ft-Ib 3,293 ft-lb
Yawi ng Moment 2,421 ft-Ib 2,512 ft-Ib
Tors ion 1,693 ft-Ib 1,759 ft-l b
2-10
steel H-section pedestal was found to have a very low fundamental frequency
in the torsional response mode (0.2 cps), and consequently a more effec
tive torsional section, e.g., 16-in. diameter concrete post section re
placed that design.
The fi rst four natural frequencies of the final single- and double-post
designs are listed in Table 2-2 and the associated first three mode shapes
are plotted in Figures B-1 to B-6 of Appendix B.
As expected, the natural frequencies of the single-post design were lower
than those of the double-post design. The fi rst mode of the single-post
design had a frequency of 6.6 cps. This mode involved bending of the post
and caused the top array panel to rock spanwise (see Figure B-1). The funda
mental mode of the double-post design (10 cps) showed the panel bending
symmetrically between the end supports, accompanied by some pitching displace
ment near the mid-panel section (See Figure B-4).
2.5 DYNAMIC INSTABILITY EVALUATION
2.5.1 Flutter Instability
The flat-plate PV array designs in this study had a fixed angle of 35° rela
tive to the ground. With this angle of attack, the only possible flutter
problem would be stall flutter with a separated wind flow, since classical
flutter occurs only at a small angle of attack with an attached flow. Stall
flutter is less critical than classical flutter, since the lift coefficient
variation of the layer angle is less significant than that of the smaller
angle. (See Figure A-25 of the theoretical analysis data reported in Reference
2-7). Therefore, freedom from stall flutter can be demonstrated if it can
be shown that this instability will not occur in a classical flutter situation.
2-11
Table 2-2
NATURAL FREQUENCIES OF PV ARRAYSTRUCTURAL SYSTEMS
Single-Post Design(Concrete Post)
Configuration III
Double-Post Design(Wood Post)
Configuration II
Mode No. Frequency (cps)Mode ShapeRemarks * Frequency (cps)
Mode ShapeRemarks*
1
2
3
4
6.60
8.76
8.89
10.34
Panel rocki ng,spanwi se
Symmetricalbendi ng of panelwith smallpitch i ng mot ion
Panel in torsion
Primarily panelpi tch i ng
10.04
10.71
12.85
13.48
Symmetrical bendi ngof panel betweensupports with pitching motion towardhigher end
Same as above
Same as above exceptpitching motion towardlower end with twisti ng of panel
Higher twisting mode
*First three mode shape plots are given in Appendix B
2-12
Classical flutter nearly always occurs at a reduced frequency, Fr less than
0.5, with 0.2 being a typical value (Reference 2-8), The reduced frequency
is defined as:
Fr = 2 'IT f n (C/2)Vw
(2-1)
where f n is the fundamental structural frequency, C is the chord length,
and Vw is the incident wind speed.
Using the Fr value of 0.2, and substituting the minimum array structural
frequency for a single-post design of 6.6 cps and a chord length of 8 ft, the
critical wind speed is:
(VW)cr = 2 'IT (6.6)(8/2)/0.2 = 829 ft/sec
or 564 mi/h
(2-2 )
Since this critical wind speed was an order of magnitude higher than the
maximum design wind speed of 90 mph anticipated at a typical design site,
freedom from flutter was assured.
2.5.2 Vortex Shedding
Periodic vortex sheddin9 can occur in the wake of bluff bodies such as cylin-
ders and flat plates, and cause periodic lift forces to develop normal to
the wind direction. The critical period of structural response occurs when
the lift force frequency approaches the basic structural frequency and a resonant
condition develops. This is "lock-in" or "synchronization" (Reference 2-6),
and the wind speed at which this occurs is the critical wind speed. The critical
wind speed is determined by:
Vcr = ¥'
2-13
(2-3)
where S is the Strouhal number (for flat plates this is about 0.13 in the
range of Reynolds number considered here), C' is the projected chord length
normal to the wind or 8 ft x sine 35° = 4.59 ft, and f n is the structural
frequency.
Based on a minimum structural frequency of 6.6 cps used previously, the
critical wind speed was:
Vcr = (6.6)(4.59)/0.13 = 211 ft/sec or
or 143 mi/h
(2-4 )
This means that the wind speed has to reach 143 mph before synchronization
can develop. Since the maximum wind speed anticipated at the field is 90
mph, it is highly improbable that large, vortex-induced structural vibrations
could occur.
2.6 DYNAMIC RESPONSE ANALYSIS
2.6.1 Dynamic Wind Load
Wind pressure exerted on the structure can be separated into a mean or static
(steady-state) pa rt and a fl uctuat i ng or dynami c part. The mean wi nd load
was extensively studied in the wind-tunnel tests conducted at the Meteoro
logical Laboratory of Colorado State University (CSU) and is fully discussed
in Section 5. Although a 1/7 power atmospheric boundary layer with its corres
ponding turbulence intensity distribution was simulated in the wind-tunnel
testing of the array structures, fluctuating wind loads were not measured.
The fluctuating wind loads used for this dynamic response analysis were de
rived from the non-steady wind pressure measurements on similar flat plate
PV models used in a recent CSU wind tunnel tests conducted for Boeing and
2-14
JPL (Appendix C of Reference 2-1). Those tests were also conducted in the
wind tunnel of CSU using the same simulated 1/7 power law boundary layer and
the reduced (l/24 scale) models. The panel size, pitching angle, array
spacing, and ground clearance were similar to Bechtel's array arrangement
except Boei ng' s reduced model s of the 8' x 20' array were supported on
four corner posts. These posts, however, would not affect the wind pressure
measurements over the panel. The wind di rection was either 00 (north wind)
or 180 0 (south wind). With- and without-fence conditions were also studied.
Pressure data was obtained for the peripheral (first) array and an inner
(5th) array. The detailed test matrix is in given Appendix C of Reference
2-1.
Eight pressure taps were placed in pairs on the front and back surfaces of
the panel along the center mid-chord. The fluctuating pressure measurements
were simultaneously recorded and presented in the form of normal ized Power
Spectral Densities (PSD). The derivation of these PSDs and the data reduction
procedure are described in Appendix C, taken from Reference 2-1.
Establishing the dynamic wind loads to analyze the single-post array, the
following was assumed:
1) The dynamic wind load condition is an extreme S-W wind incident
upon the south perimeter row of a field of si ngle-post arrays
unprotected by fences. The reference wind speed is 80 mi/h.
2) The frequency content of the fluctuating wind pressure is the
Same everywhere in the panel and can be represented by a single
PSD.
This single PSD was computed from the normalized back and front PSD pressure
2-15
data from Reference 2-1 for the north wi nd without fences (this was the
closest available case):
Ie, ¢ = ¢ ~. C 2 + ¢~. C 2 + 2 ¢ ~. C . C .11 prmsi JJ prmsj lJ prmsl prmsJ (2-5)
x xwhere ¢ i i and ¢ jj were the normal ized auto-PSD of the front and back
xpressures respectively, and ¢ ij was the cross-PSD of the two pressures, where-
as the Cprms's were the respective pressure coefficients.
xThe summation in the above equation involved only real parts of ¢ s, since
the imaginary parts of both auto- and cross-spectra are essentially zero
(Reference 2-1).
The log-log plot of Ie, ¢ computed from Eq. 2-5 is shown in Figure 2-5. The
xsimple, straight-line relationship shown is because ¢ functions on the
right side are similarly constructed to approximate the original PSD test
data (Figures 4 and 10 of Reference 2-1).
The accuracy of the linear (log-log) approximation of Ie,¢ vs. Nwas checked
by integrating the curve to yield the square of the corresponding rms pressure
coeffi ci ent:
00f Ie, ¢ (N) dn
o
(2-6)
The calculated Ie, Cprms (0.12) agreed with the average of Ie, Cprms values
measured along the chord (Reference 2-1).
3) The dynamic rms pressure is a constant fraction (25%) of
the mean wind load, or
Ie, Cprms / Ie, Cpmean ; 0.25
2-16
(2-7)
14.7 cps (FULL SCALE)WHICH IS EQUIVALENTTO 129 cps (1/24 SCALEI
1. 2 4 6 810 20 400.1om1O-s ......jL-r-r-,-,.,.,n-rr---r-r-rrrrm---r-r-,-TT1rm----"....,..,
W0:::>~w0::Q.
fwZ
~
oUl .Q.
1=zwUu.u.wou
N - FREQUENCY (cpsl
Figure 2-5 Wind Pressure Power Spectral Density Input
2-17
The distribution of ~Cpmean on the panel was shown in Figure
2-3. The 0.25 ratio was based on the wind tunnel test data of
~Cpmean and ~Cprms obtained for a similar flat-plate
model under a comparable corner wind condition (Figure 4-11a of
Reference 2-9).
2.6.2 Dynamic Response Analysis Method and Results
The fluctuating wind load defined in Section 2.6.1 was input to a stationary
random-frequency response analysis implemented by the DYNRE3 routine of the
STARDYNE program (Reference 2-2). The generalized modal response was computed,
then the local response was obtained from the sum of contributions (in
PSD's) from each mode. Fifteen modes were used for the response analysis of
the single-post array structural system. In each mode, the response in PSD
is related to the forcing function (load) in PSD by the square of the struc
tural admittance (transfer) function (detailed description can be found in
Section 25-2 of Reference 2-10).
Since this case involved a separated flow, the aerodynamic damping and stiff
ness matrices (derived from the potential flow theory adopted in the classical
flutter analysis) were not considered (Reference 2-8).
In the preliminary investigations, a 10% structural damping was selected for
dynamic evaluation of steel single-post concepts. Later it became apparent
that a higher damping value could be used for the first design with concrete
or wood posts. Using concrete or wood posts, the designs of the foundations
were governed by the lateral resistance of the soil and not by the stress
levels in the posts. Therefore, it was decided that a 2% structural damp
ing would be a reasonable lower-bound value, based on the latest recommen-
2-18
dation by Newmark and Hall (Reference 2-11). The recommendations were
specified according to the range of stress level and type of structure shown
in Table 2-3. For the expected stress level below yield in the posts, and
for a welded steel structure. then Table 2-3 gives 2% damping as lower-bound
val ue.
The dynamic analysis results. in the form of structural member internal loads
and other response parameters, are summarized in Tables 2-4 and 2-5 for 1%
and 2% structural damping. The (rms) response values obtained from STAROYNEj
OYNRE3 analysis were based on 6C prms being 25% of 6Cpmean (Eq. 2-7). Assum
ing a Gaussian distribution. the rms response value was related to the extreme
response value according to the selected probability level of nonexceedance.
as follows:
Probability of Nonexceedance Ext reme Value
68.27% (1 Cl - standard deviation) 1 x rms value
95.45% (2 Cl ) 2 x rms val ue
99.73% (3 Cl ) 3 x rms value
The dynamic amplification factor (OA) is defined as the ratio between the
dynamic response and the equivalent static response or response obtained
from applying dynamic loads in a quasi-static manner. Since 6 Cprms is
25% of 6Cpmean. the equivalent static response is simply 25% of the
steady-state response.
The turbulence factor (K) obtained from the response analysis is defined
as:
K = Steady-state response + extreme dynamic response (2-8)steady-state response
2-19
StressLevel
Working stress,tono more thanabout 1/2 yi el dpoint
At or just belowyield point
Table 2-3
RECOMMENDED DAMPING VALUES{Newmark and Hall (Reference 2-11))
Type and Conditionof St ructu re
a. Vital piping
b. Welded steel, prestressedconcrete, well-reinforcedconcrete (only slight cracking)
c. Reinforced concrete withconsiderable cracking
d. Bolted and/or riveted steel,wood structures with nailedor bolted joints
a. Vital piping
b. Welded steel, prestressedconcrete (without completeloss in prestress)
c. Prestressed concrete with noprestress 1eft
d. Reinforced concrete
e. Bolted and/or riveted steel,wood structures, with boltedjoi nts
f. Wood structures with nailedjoints
2-20
PercentageCritical Damping
1 to 2
2 to 3
3 to 5
5 to 7
2 to 3
5 to 7
7 to 10
7 to 10
10 to 15
15 to 20
or K = 1 + no x rms dynamic responsesteady-state response
(2-9)
Different turbulence factors may be assigned to various portions of the struc-
ture. Therefore, the array structures were recommended to be designed on the
basis of extreme, steady-state wind loads multipled by an overall design K fac-
tor of 3 for the single-post design and 2 for the double-post design (Reference
Section 1.3.1). This factor can be validated by using the turbulence factors
estimated above for the various portions of the structure.
2.6.3 Discussion of Results
The local dynamic response of array structures depends on the characteristics
of important contributory modes and how effectively the dynamic wind load
excites these modes. The local response is related to the modal response by
the mode shapes. The effectiveness depends on modal frequency, the spatial
distribution, and correlation of wind loads and their relationship to the
mode shapes.
The input wind spectrum shown in Figure 2-5 indicated that most of the wind
energy was contained in the low-frequency end; thus the lower the structural
frequency, the higher the dynamic response.
With the above background, a trend was observed in the dynamic response re-
sults in Tables 2-4 and 2-5. The largest dynamic amplification was in the
bending and torsion loads in the post. This is because several lower modes
(first, third, and fourth) involved bending and torsion of the post causing
the rigid-body rocking and pitching of the array panel (Appendix B).
The maximum dynamic amplification (DA) is 2.84 for 2% damping, comparing to
3.94 for 1% damping (Tables 2-4 and 2-5). The general dynamic response of
2-21
"I " "
Tab
le2-
4
DYNA
MIC
RESP
ONSE
RESU
LTS
-FL
UCTU
ATIN
GW
IND
LOAD
S,S-
WW
IND
ONSI
NG
LE-P
OST
DESI
GN-
1%DA
MPI
NG
IS
tru
ctu
ral
Res
pons
eRM
SE
xtre
me
Ext
rem
eS
tead
y-S
tate
Dyn
amic
Tur
bule
nce
Fac
tor
KM
embe
rP
aram
eter
Uni
tsV
alue
sV
alue
sV
alue
sV
alue
sA
mpl
ific
atio
n68
.27%
95.4
5%99
.73%
*(m
axiu
m)
1+1+
1+(1
)(2
)(3
)(4
)4
x(1
)/(4
)(1
)/(4
)(2
)/(4
)(3
)/(4
)
Tor
que
Tube
Yaw
ing
Mom
ent
ft-l
b2,
968
5,93
68,
904
4,89
42.
431.
612.
212.
82
IM
odul
eSu
ppor
tP
itch
ing
Mom
ent
ft-l
b19
138
257
342
71.
801.
451.
892.
34(C
ross
)
Mod
ule
Sup
port
Yaw
ing
Mom
ent
ft-l
b18
3654
401.
801.
451.
902.
35(L
ongi
tudi
nal
)I
IP
ost
IPit
chin
gM
omen
tft
-lb
2,45
14,
902
7,35
32,
544
3.85
1.96
2.93
3.89
Yawi
ngM
omen
tft
-lb
2,35
54,
710
7,06
52,
421
3.89
1.97
2.94
3.92
IT
orsi
onIft
-lb
1,66
93,
338
5,00
71,
693
3.94
1.99
2.97
3.96
IIA
xial,
X3Il
b52
91,
058
1,58
71,
122
1.89
1.47
1.94
2.4
Shea
rX2
1b50
91,
018
1,52
778
42.
601.
652.
302.
95
*P
rob
abil
ity
leve
lof
none
xcee
danc
e
N I N W
Tab
le2-
5
DYNA
MIC
RESP
ONSE
RESU
LTS
-FL
UCTU
ATIN
GW
IND
LOAD
S,S-
WW
IND
ONSI
NGLE
POST
DESI
GN-
2%DA
MPI
NG
Str
uct
ura
lR
espo
nse
RMS
Ext
rem
eE
xtre
me
Ste
ady-
Sta
teD
ynam
icT
urbu
lenc
eF
acto
rK
Mem
ber
Para
met
erU
nits
Val
ues
Val
ues
Val
ues
Val
ues
Am
plif
icat
ion
68.2
7%95
.45%
99.7
3%*
(max
ium
)1+
1+1+
OJ
(2)
(3)
(4)
4x
(l)/
(4)
01
/(4
1(2
)/(4
1(3
)/(4
)
Tor
que
Tube
Yaw
in9
Mom
ent
ft-l
b2,
198
4,39
66,
594
4,89
41.
801.
451.
902.
35I I
IM
odul
eSu
ppor
tP
itch
ing
Mom
ent
ft-l
b15
030
045
042
71.
401.
351.
702.
05(C
ross
)
Mod
ule
Supp
ort
Yaw
ing
Mom
ent
ft-l
b14
2842
401.
401.
351.
702.
05(L
ongi
tudi
nal)
I
Pos
tP
itch
ing
Mom
ent
ft-l
b1,
724
3,44
85,
172
2,54
42.
711.
682.
363.
03
Yaw
ing
Mom
ent
ft-l
b1,
701
3,40
25,
103
2,42
12.
811.
702.
413.
11
Tori
son
ft-I
b1,
203
2,40
63,
609
1,69
32.
841.
712.
423.
13
Axi
alX3
lb43
987
81,
317
1,12
21.
571.
391.
782.
17
She
arX2
lb36
973
81,
107
784
1.88
1.47
1.94
2.88
*P
rob
abil
ity
leve
lof
none
xcee
danc
e
2% dampi ng is about 1/ .[2of the 1% dampi ng response or by a factor equal
to the inverse of the square root of the damping increase ratio. DA results
can be checked using the simplified equation for estimating the response of
a lightly-damped single-degree-of-freedom (SDOF) system, e.g., Eq. 10.3-8
of Reference 2-12:
_ en Sen) 1f ) 1/24E:
(2-10)
where Xis the rms response value, n is the structural frequency of the SDOF
system, S(n) is the value of the input load PSD at n : n, and £ is the
dampi ng rat io.
The expression for (DA) can be obtained by dividing S(n) by the integral
of S (n):
(2-11)
1/2
)DA "( (n :(n) '%IE
J S(m) dmo
If S is in terms of pressure coefficients, the integral is equal to:
J00
2S(m) dm = /', Cprms
(2-12)
o
Assuming a SDOF representation of the single-post design with the fundamen-
tal frequency, 6.6 cps (yawing mode, Table 2-2), and the load PSD specified by
Figure 2-5, the calculated values from Eq. 2-11 compared with those from the
more rigorous STARDYNE/DYNRE3 analysis (Tables 2-4 and 2-5) are shown below:
1% damping
2% dampi ng
Eg. 2-11
3.3
2.3
STARDYNE
3.89 (post bending, yawing)
3.94 (max. value, post torsion)
2.81 (post bending, yawing)
2.84 (max. value, post torsion)
2-24
The comparison indicates that OA predicted by a multi-degree-of-freedom (MOOF)
PSO analysis could in some cases be about 20% higher than those using Eq. 2-11
using a single OOF representation.
Equation 2-10 also indicates that the rms dynamic response is proportional to
the square root of damping. Therefore, increasing the dampi ng ratio from 1%
to 2% reduces the dynamic response by 1/,/2. This is confirmed by comparing
dynamic response data in Tables 2-4 and 2-5.
The turbulence factor K results in Table 2-5 were computed using the selec
ted 2% structural damping. They range from 2.05 to 3.13 for a 99.73% (30 )
probability of nonexceedance, and from 1.70 to 2.42 for a 95.45% (20) probabi
lity of nonexceedance. The design K factor recommended for the single-post
structure design was 3 (Section 1.3.1) for the 99.73% (30 ) confidence level.
However, in the final design of the preferred double-post design (Section 1.4),
the design turbulence factor K was reduced from 3 to 2. In examining Table 2-5
of the single-post design, K factors for elements of the superstructure, i.e.,
torque tube and modul e supports, were all below 2 for a 95.45% (20 ) probabi
lityof nonexceedence. Even though the single-post design had a maximum K
factor of 2.42 for the support post for the same confidence level, a double
post design would not be as susceptible to amplifications.
Since the design of the posts is not governed by stresses but by the ultimate
resistances of the surrounding soil, a lower design K factor could be used
for the double-post design. With optimization and low cost in mind, it was
felt the 95.45% (20) confidence level would be adequate for the final design
of the double-support concept and a K factor of 2 was recommended and used
for design.
2-25
2.6.4 Fatigue Considerations
Wind-induced dynamic alternating stresses could cause a fatigue failure at a
stress level below that normally expected during monotonic loading conditions.
Fatigue analysis was performed on the post joint, since all the wind loads
on the array panel have to be transmitted through the joint to the foundation.
Other structural members and welded connections were also checked.
The dynamic wind stresses were computed for fatigue analysis, based on the
previous dynamic response analysis results for the 80 mi/h extreme S-W wind,
scaled down to an operating basis wind speed of 60 mi/h.
The results showed that the dynamic stress level (3 values) in the post is
insignificant (not more than 1/3 of concrete cracking strength), and the
stress produced elsewhere in the array panel steel (A36) members is less
than the fatigue allowable corresponding to 2 x 106 stress cycles. The
fatigue allowables used were the suggested values given in Table 1, Section
2.9 of Reference 2-13.
A preliminary fatigue analysis of the panel/post joint (details in Figure
1-32) indicated that the most critically stressed area was the butt weld
connecting the torque tube to the flanges of the mounting bracket (which is
attached to the concrete post by four anchor bolts). The anchor bolts are
sUfficiently preloaded to prevent fatigue damage.
2.7 SUMMARY AND CONCLUSIONS
The results of the dynamic wind analyses completed in Task II are:
1. The finite element modal analysis of the final PV array design
2-26
showed that the single-post design had the lowest frequency of
6.6 cps, compared to 10 cps of the double-post design.
2. A stationary, random, dynamic response analysis of the single
post design indicated a large wind-response effect even though
the structural frequency is at the higher-frequency end of the
wind spectrum, where the wi nd energy is relatively small. The
important parameters affecting the results were the assumed damp
ing value, the acceptable level of probability of nonexceedance,
the structural dynamic properties, and the input wind spectra.
3. A simple dynamic instability evaluation indicated that the chance
of flutter and resonance with vortex shedding forces was very
remote for the range of wind speed anticipated at a typical site
of a large array field.
4. The dynamic wind stresses produced in the single-post array struc
tures were below the fatigue allowable for the structural members.
Fatigue damage of the critical array panel and post joint can be
avoided by careful design detailing.
5. The turbulence factors K computed from the dynamic analysis of
single-post design [corresponding to 99.73%{3a) probability
of nonexceedance] were either below or slightly above the design
value of 3 used in the design. This was acceptable, since the
turbulence factor of 3 occurred at the post-and-ground interface
where the combined wind and static stresses were considerably
below the allowable.
6. For the double-post design, the turbulence factors computed from
2-27
the dynamic analysis [corresponding to 95.45% (20) probability of
nonexceedences] were all below 2 for the panel and below 3 for the
posts. Since the combined wind and static stresses for the post
were considerably below the allowable, a design turbulence fac
tor of 2 was selected for double-post design.
2.8 RECOMMENDATIONS FOR FUTURE STUDIES
Several unresolved technical issues were identified during the study, and
should be further investigated:
1. The input wind spectrum for the dynamic response analysis was
based on the wind-tunnel test results from Reference 2-1. The
simulation of near-ground turbulence in terms of turbulence
intensity and frequency content needs to be substantiated by
field measurements. There are also insufficient measurements
on spanwise distribution and correlation of fluctuating loads on
the array panel.
2 There is a need to determine realistic damping values for dynamic
analyses of different PV array structures.
3. The effect of soil foundation flexibility and the contribution
of soil damping to structural response should be investigated.
4. Dynamic analysis of the double-post design with wood and concrete
pedestals should be performed. The dynamic stresses produced in
the post and superstructure of the double-post design will not
be as high as that in the single-post design, because the panel
is more effectively supported and the structural frequencies
are much higher than for the single-post design.
2-28
Section 3
SYSTEM COST ESTIMATES - TASK III
Preliminary and final cost estimates and analyses were performed to evaluate
the cost impact on the integrated structural designs for photovoltaic arrays.
The results of these conceptual estimates, their basis, qualifications, and
exclusions are presented in this section.
The photovoltaic array structure provides support for photovoltaic panels
for power production from solar energy. A solar central power array field
consists of hundreds or even thousands of simple ground-mounted array
structures. The three main components of each array structure are:
o Superstructure frame carrying the modules
o Structural connections
o Pedestal and foundation
During the preliminary cost estimates and analysis stage of this study, various
array design concepts were considered from the cost point-of-view. The objec
tive of these preliminary estimates was to identify and explain major cost
elements of di fferent concepts for 1 MW and 50 MW array i nsta11 at ions. The
emphasis on this portion of the work was placed on relative cost comparisons
among these alternatives.
From the results of this preliminary cost study, two of the most promising con
cepts were selected for further detailed testing and optimization. Consequently
final cost estimates and analyses were evaluated for the projected construction
of large array fi e1ds. The object i ve of the second port i or 'lf thi s study
was to project the likely actual construction cost requiren. ,Its to build a
10 MW solar power system. Designs of end-supported arrays having four different
3-1
span lengths were also estimated. Expected costs were also estimated for
the construction of a brief demonstration array structure at an Albuquerque,
New Mexico test site.
3.1 BASES FOR CONSTRUCTION COST ESTIMATES
The technical scope and design bases for the photovoltaic array structures
have been di scussed in Section 1. The cost estimates here are based on the
conceptual and final designs and engineering information prepared for the
study in the form of engi neeri ng drawi ngs, outl i ne speci fi cati ons, sketches
and scope narrat i ves. Est i mat i ng methods cons i stent with the conceptual
nature of the design information were employed and supported by informal
vendor contacts, and written quotations, as well as by extrapolation from
current Bechtel information.
These estimates include modest site preparation, requisite earthwork, and
location marking. These estimates also include all necessary materials and
installation costs to provided structural support for the modules. Modules
and module mounting hardware are excluded.
These estimates reflect an engineer/constructor's direct-hire operation
employing local manual labor and specialist subcontractors. All costs are
stated at second quarter, 1981, (mid-1981 $) price and wage levels with
no allowance for future escalation.
The schematic breakdown of the total construction cost is shown in Figure
3-1. Three main components of the total construction cost are:
o Field costs
o Engineering services
o Allowance for uncertainty
3-2
w , w
TO
TA
LC
ON
ST
RU
CT
ION
CO
ST
(PR
ICE
)
ff
TO
TA
LF
IEL
DE
NG
INE
ER
ING
AL
LO
WA
NC
E
CO
ST
SE
RV
ICE
SF
OR
UN
CE
RT
AIN
TY
Tf
4D
IRE
CT
FIE
LD
IND
IRE
CT
FIE
LD
CO
ST
CO
ST
•f
+T
1M
AT
ER
IAL
MA
JOR
LA
BO
R
CO
ST
EQ
UIP
ME
NT
INS
TA
LL
AT
ION
SU
BC
ON
TR
AC
T
CO
ST
CO
ST
CO
STS
Figu
re3-
1C
ost
Est
imat
eF
low
Dia
gram
3.1.1 Field Costs
The fi el d costs are composed of the (1) di rect fi el d costs and (2) i ndi rect
field costs.
Direct Field Costs. These costs are comprised of the following:
o Materials
o Major equipment
o Construction labor
o Subcontracts
These costs included in the estimate are based on the following discussion.
Materials - Quantity evaluations were developed from the engineering drawings
and sketch take-offs for such materials as fabricated galvanized steel, corru-
gated steel, lumber, timber piles, etc. Material pricing was based on formal
and informal vendor contracts, and estimates using current unit prices from
Bechtel sources. The general level of material pricing includes:
0 Fabricated galvanized steel $3,600 - $5,000 per ton
0 Fabricated structural steel $1,280 per ton(torque-tube)
0 Fabri cated corrugated steel $1,200 per ton
0 Fabricated carbon steel $2,000 per ton
0 Structural board lumber $540 per thousand board ft
0 Treated wood pole, 12 in. $5.40 per linear ftdi ameter
0 Structural steel tubing $600 - $700 per ton
Major Equipment - Generally, the costs of any major equipment defined by the
conceptual design descriptions are estimated based on oral vendor quotations,
3-4
vendor catalogs, material pricing bulletins, and recent estimating experiences
with similar equipment. In this study, there is no major handling or process
equipment anticipated in the construction of a photovoltaic power plant. There
fore, major equipment costs were not included in this estimate.
Construction Labor - The on-site construction manhours for the installation
of materials were estimated using Bechtel standard unit manhours, adjusted
for recent productivity in the state of New Mexico. Recent productivity
experience for the Albuquerque, New Mexico, resulted in the use of a factor of
1.0 based on Bechtel standards.
An on-site labor rate of $20.00 per manhour was used based on an evaluation
of craft labor agreements for the area. This labor rate represents an assumed
mix of crafts appropriate to the type of construction. It includes base rate,
fringe benefits, payroll taxes, and allowances for casual overtime, and an
estimated travel and subsistence allowance.
Subcontracts - Subcontracts for construction operations normally performed
by speci al i st subcontractors were est imated and pri ced in accordance with
Bechtel experience and information from potential subcontractors. The sub
contractor estimate includes the following operations:
o Shop fabrication for the superstructures/panels
o Concrete caisson foundation installation (includes material costof foundation)
o Steel pipe pile driving
o Steel H-pile driving
o Wood pole driving
The use of specialist subcontractors results in higher productivity and lower
3-5
costs when compared to field operations due to the following conditions:
o Repetitive assembly operations and design specifications
o Better and consistent working conditions
o Use of specialized tools and equipment
o Use of well-trained and experienced craftsmen
Indirect Field Costs. These costs consist of construction costs that cannot
be ascribed to di rect portions of the facilities and thus are accounted
separately, and typically include the following:
o Temporary Construction Facil it i es - Temporary buil di ngs, worki ngareas, roads, parking areas, utility system and general purposescaffoldi ng.
o Miscellaneous Construction Services - General job cleanup, maintenance of const ruct ion equi pment and tool s, materi a1 hand1i ngand surveyi ng.
o Construction Equipment and Supplies - Construction equipment, smalltools, consumable supplies, and purchased utilities.
o Field Office - Field labor for craft supervision, engineering,procurement, scheduling, personnel administration, warehousing,first aid, and the costs of operating the field office.
o Preliminary Check-Out and Acceptance Testing - Testing of materialsand equipment to insure that components and systems are operable.
o Project Insurance - Public liability, property damage, and builder's risk insurances.
3.1.2 Engineering Services
The engineering services include engineering costs, other home office costs
and fees. Engineering includes preliminary engineering, optimization studies,
specifications, detail engineering, vendor-drawing review, site investigation,
and support to vendors. Other home office costs are comprised of procurement,
estimating and scheduling services, quality assurance, acceptance testing,
and construction and project management. Fees are included as a function of
3-6
the total project cost. The sum of these three categories falls historically
into consistent percentages in the range of 5% - 25% of total field costs
depending upon the complexity and cost of the project.
3.1.3 Allowance for Uncertainty
Included in the estimate is an allowance for the uncertainty that exists with-
in the conceptual design in quantity, pricing, or productivity that is under
the control of the engineer/constructor and within the scope of the project
as defined. Implicitly, the allowance will be expended during the design
and construction of the project and it cannot be considered as a source of
funds for overruns or additions to the project scope. Thus, if the concep-
tual arrangement of the plant components contains major uncertainties, or
the design duty of plant components proves to be more severe than anticipated,
or if additional major subsystems are ultimately found to be necessary, then
the scope of the project is deemed to be inadequately defi ned and is not
covered by the allowance.
Allowance for uncertainty is also known as contingency. The contingency
allowance included in the construction costs is normally estimated in the
range of 15 to 25 % of the sum of the total field costs and engineering
services.
3.1.4 Qualifications
The following are the major qualifications in the estimate:
o The estimate was prepared assuming that the scope of service willbe that of a prime contractor responsible to the owner forengineering, procurement and construction.
o Equipment and materials will be procured from U.S. sources, andlead times will be able to support the project schedule withoutcost penalties.
3-7
o Sufficient manual and non-manual personnel to complete the projectwithin the construction schedule are assumed to be available inthe project vicinity.
o Equipment mobilization cost is included in the indirect fieldcost est imate.
o Site prepa rat i on has been compl eted. The soil condi t ions arestable and essentially free of rock.
3.1.5 Exclusions
The following items are excluded from the project scope and are therefore not
included in the estimate:
o Cost of solar modules and associated module hardware
o Wiring and any auxiliaries
o Any special construction such as widening and strengtheningexisting roads
o Owner costs, such as engineering, land acquisition, cost offinancing, licensing, royalties, etc.
o State and local taxes
o Future escalation.
3.2 PRELIMINARY COST ESTIMATE ANALYSIS
Severa1 photovolta i c array support structure des i gn concepts were selected
and considered in the preliminary studies. These designs include the exist-
ing Jet Propulsion Laboratory (JPL) concept and six Bechtel design concepts.
The JPL support concept was included to serve as a comparison basis for the
six Bechtel concepts which were all costed on the same level of conceptual
details. Costs were estimated for a 1 MW array field size (approximately
6690 m2 of array area) and then adjusted to a 50 MW array field size (approx
imately 334,500 m2 of array area). Costs are subdivided into two categories
which include the frame superstructure and foundation installation costs.
3-8
As indicated earlier, the emphasis for these preliminary cost estimates
was placed on relative construction cost comparisons among these seven
alternati ves. Tables 3-1 and 3-2 summarize photovoltaic array fi el d construc
tion costs for the seven design concepts based on 1 MW and 50 MW array field
sizes.
3.2.1 Field Costs
The technical scope and design basis for these seven array-supporting concepts
are described in Section 1.
Direct Field Costs. The direct field cost estimates for these seven concepts
were expressed as unit costs for comparison purposes. Table 3-3 and Table 3-4
show the unit cost breakdown between superstructures and foundation installa
tions of the seven concepts based on 1 MW and 50 MW field size respectively.
The unit cost per array surface area is the total direct field costs derived
from Table 3-1 and Table 3-2 divided by either 6690 m2 or 334,500 m2 for
the 1 MW and 50 MW array installation scenarios.
Indirect Field Costs. Indirect field costs were estimated as a percentage
of direct labor costs. They were estimated with reference to Bechtel's
construction experience and consideration for the repetitive nature of the 1
MW and 50 MW array field installation scenarios. These resulted in assess
ments of 100"~ and 65% of di rect labor costs for the 1 MW and 50 MW plant
sizes respectively.
In the 1 MW plant study, the ratio between direct labor and direct field costs
ranges between 18% and 22%. Direct field costs range from approximately $240,000
to $590,000 and then 100% of the di rect labor cost was selected to represent
indirect costs at this stage.
3-9
Table 3-1
COMPARATIVE ESTIMATE SUMMARY FOR 1 MW ARRAY FIELD(450 - 8' x 20' Arrays)
($ in $l,OOO's, 2nd Quarter 1981)
Items JPL 1 2 3 4 5 6
FramePu rchase Price 247 154 154 154 154 525 326
FoundationInstallation 75 128 132 165 87 63 72
Total FieldDi rect Cost 322 282 286 319 241 588 398
IndirectField Cost 59 49 60 ..Jl 62 63 72
TotalField Cost 381 331 346 387 303 651 470
Engi neeri ngServices 95 82 87 96 75 130 94
All owance forUncertainty 74 62 67 72 57 119 86
TotalConstruct ionCost 550 475 500 555 435 900 650
----- ---- ==== ==== :::;:;;=:;; ==== ----
$/Sq.M:
Di rect Cos t: 48 42 43 48 36 88 60
Total Cost: 82 71 75 83 65 134 97
Note: These costs only compare concepts on a common basis and are not optimizedat this stage.
3-10
Tab Ie 3-2
COMPARATIVE ESTIMATE SUMMARY FOR 50 MW ARRAY FIELD(22,000 - 8' x 20' Arrays)
($ in $l,OOO's, 2nd Quarter 1981)
Items JPL 1 2 3 4 5 6
FramePurchase Price 10,520 6,910 6,910 6,910 6,910 23,620 14,670
FoundationInstallation 3,550 6,100 6,280 7,840 4,130 2,990 3,420
Total FieldDirect Cost 14,070 13,010 13,190 14,750 11,040 26,610 18,090
IndirectField Cost 1,810 1,500 1,860 2,090 1,920 1,940 2,220
TotalField Cost 15,880 14,510 15,050 16,840 12,960 28,550 20,310
EngineeringServi ces 1,190 1,090 1,130 1,260 970 2,140 1,520
Allowance forUncertainty 2,530 2,350 2,420 2,700 2,070 4,610 3,270
TotalConstructionCost 19,600 17,950 18,600 20,800 16,000 35,300 25,100
======= ======= :::=::::=::;== ::;=::=::;== ====:;:; :::;:;;::::::::;= ======
$/Sq.M:
Di rect Cost: 43 40 40 45 34 81 55
Total Cost: 60 55 57 64 49 108 77
Note: These costs only compare concepts on a common basis and are not optimizedat this stage.
3-11
TABLE 3-3
COMPARISON OF DIRECT COSTS FOR lMW FIELD
(BASED ON 450 UNITS - 8' X 20')
($ per Array)
1981 $
FOUN DATION $ FRAME $ TOTAL COST $ S/m2 $/SFJPL
h~/f/GALVANIZED
7 -/i STEEL
:;po FRAMEW/WOODEN 166 550 716 48 4.5
TRUSSFOUNDATION
1
-t TORQUE TUBEW/CAISSDtJ 285 341 626 42 3.9FOUNDATION
2
1- TORQUE TU8E\'IISTEEL PIPE 294 341 635 43 4.0FOUNDATION
3 .,.//TORQUE TU8EW/H·PILE 367 341 708 48 4.4FOONDATION
4
f TORQUE TUBE
" W/WOOD PILE 193 341 534 36 3.3
J.FOUNDATION
5 ,C\ L-5HAPEDCORRUGATED
1,167STEEL 140 1,307 88 8.2STRUCTURE
.~
6 HALF PIPECORRUGATED 160 724 884 59 5.54 STEEL
~.-.--.":\.;.' STRUCTURE
3-12
TABLE 3-4
COMPARISON OF DIRECT COST FOR 50MW FIELD
(BASED ON 22,000 U~J1TS - 8' X 20')
($ per Array)
1981 $
FOUNDATION $ FRAME $ TOTAL COST $ $1m2 $/SFJPL ///;/ GALVANIZED
"7 7"- STEEL
~FRAME
W/WDDOEN 161 478 639 43 4.0TRUSSFOUNOATIDN
1
-t TORQUE TUBEW/CAISSDN 277 314 591 40 3.7FOurlOATION
2
1- TQRQUE TUBEW/STEEL PIPE 285 314 5S9 40 3.7FDUNDATIDN
3 .,//TORQUE TUBEW/H·PILE 356 314 670 45 4.2FDUNOATIDfJ
4
f TORQUE TU3EWIWOOO PILE 188 314 502 34 3.1
f<FOUNDATION
5 Ie:: L·SHAPEDCORRUGATEDSTEEL 136 1,073 1,209 81 7.6STRUCTURE.-
6 HALF PIPECORRUGATED 155 &&7 822 55 5.14 STEEL
~ _.--... :\.,:.' STRUCTURE
3-13
The 50 MW plant study shows the same ratio between di rect labor and di rect
field costs as the 1 MW plant but, due to the larger volume of arrays in
stalled, di rect fi el d costs were rai sed to between $11 milli on and $26
mill i on. Here 65% of di rect labor cost was sel ected to represent i ndi rect
costs.
3.2.2 Qualifications, Exclusions, and Assumptions
The qualifications and exclusions itemized in Sections 3.1.5 and 3.1.6 apply
to the preliminary cost estimate. The following are the major assumptions
for which design data were not available when this estimate was prepared.
a Engi neering services were estimated to be between 20-25% of thetotal field costs for the 1 Mill array field and taken at 7.5%for the 50 MW array field.
a The contingency was estimated at 15% of the sum of the total fieldcosts and engineering services.
o Installation costs of the solar modules were excluded from theestimate.
o The materi al costs for the JPL superstructure were estimated assuming press-brake processing for the 1 MW array field (450 units)and roll-form processing for the 50 MW array field (22,000 units).
o For wage levels, the site is taken to be Albuquerque, New Mexico.
3.2.3 Evaluation
While investigating the costs of various support concepts for the photovoltaic
arrays, those structural components which contributed significantly to major
costs were identified. The relative costs of these structures were then used
for the purpose of i dent i fyi ng promi 5 i ng candi date confi gurat ions and thus
discarding uneconomical designs from further consideration.
As shown in Table 3-3, the JPL concept was evaluated at $48 per square meter
of array area. This served as a comparison on a common basis with the
3-14
other six Bechtel concepts which were estimated on the same level of con
ceptual details. Bechtel concepts, Cases 1 through 4, were estimated ranging
from $36 to $48 per meter square based on direct field costs.
Bechtel concepts, Cases 5 and 6, used corrugated steel pipe for structural
support of the panel s. They both indicated a higher cost per array area
than the other concepts. Thi sis due to the heavy wei ght and hi gh cost of
the corrugated steel. Therefore, these concepts were not considered further.
The degree of accuracy at this conceptual level was such that the estimates
served only to indicate approximate magnitude of costs and probable ranking
between alternates. Alternates which appear within 10% of each other should
be ranked together in cost, since they may incorporate inaccurate assump
tions of opposite values that could reverse their ranking. Similarly, sites
with conditions other than those listed for the baseline design criteria
(e.g., soil type, wind and snow loading, etc.) could influence their ranking.
Two support structures, the concrete caisson and wood pole foundation
concepts, showed the most promise for low cost and were selected for the
detailed designs and cost studies discussed in further detail in Section 3.3.
3.3 FINAL COST ESTIMATE ANALYSIS
As mentioned in Section 1.4.3, double-supported arrays were designed having
various span lengths. The cost breakdown in actual dollars and unit costs
of those designs are given in Tables 3-5 and 3-6. The most economical span
was found to be 36 ft.
Usi ng that optimum span length, the array supports were shifted to achieve
further cost opt imi zat i on and a1so reduce the moments and defl ect ions in
the 36-ft span. Thus, the arrays were changed from being end-supported to
3-15
w , ..... Gl
Tab
le3-
5
TOTA
LCO
STCO
MPA
RISO
NSOF
THE
DOUB
LE-S
UPPO
RTED
ARRA
YST
RUCT
URE
FOR
VARI
OUS
SPAN
S
(Est
imat
esar
eba
sed
on10
MWfi
eld
size
qu
anti
ties
)($
in10
00's
and
use
mid
-198
1$)
$C
ost
for
5000
$C
ost
for
3125
$C
ost
for
2778
$C
ost
for
2500
8'x
20'
8'x
32'
8'x
36'
8'
x40
'A
rra.
ysA
rray
sA
rray
sA
rray
s
Ext
ende
dT
imbe
rE
xten
ded
Tim
ber
Ext
ende
dT
imbe
rE
xten
ded
Tim
ber
Cos
tC
oncr
ete
Pol
esC
oncr
ete
Pol
esC
oncr
ete
Pol
esC
oncr
ete
Pol
esE1
emen
tsC
aiss
onSu
ppor
tC
aiss
onSu
ppor
tC
aiss
onS
uppo
rtC
aiss
onS
uppo
rt
1.Pa
nel
f.o
.b.
job
site
1,35
01,
350
1,51
61,
516
1,50
01,
500
1,65
01,
650
2.Pa
nel
inst
alla
tio
n30
030
026
626
625
025
025
025
0
3.M
odul
ein
stal
lati
on
300
300
281
281
278
278
275
275
4.S
truc
tura
lco
nnec
tion
s22
537
514
023
412
520
811
218
8
5.In
stal
led
Fou
ndat
ion
600
525
422
390
375
347
333
325
Dir
ect
Fie
ldC
ost
2,77
52,
850
2,62
52,
687
2,52
82,
583
2,62
52,
683
Ind
irec
t-
50%
Fie
ldL
abor
300
360
273
315
264
301
263
297
*Tot
a1In
stal
led
cost
=3,
075
3,21
02,
898
3,00
22,
792
2,88
42,
888
2,98
5I
*Inc
1ude
sov
erhe
adan
dp
rofi
tbu
tex
clud
esen
gine
erin
gan
dco
ntin
genc
y
W I ......
'-oJ
Tab
le3-
6
UNIT
COST
COM
PARI
SONS
OFTH
EDO
UBLE
-SUP
PORT
EDAR
RAY
STRU
CTUR
EFO
RVA
RIOU
SSP
ANS
(Est
imat
esar
eba
sed
on10
MWfi
eld
size
qu
anti
ties
)($
in10
00's
and
use
mid
-198
1$)
$/m
2$/
m2
$/m
2$/
m2
8'
x20
'8'
X32
'8'
x36
'8
'x
40'
Arr
ays
Arr
ays
Arr
a.ys
Arr
ays
Ext
ende
dT
imbe
rE
xten
ded
Tim
ber
Ext
ende
dT
imbe
rE
xten
ded
Tim
ber
Cos
tC
oncr
ete
Pole
sC
oncr
ete
Pole
sC
oncr
ete
Pol
esC
oncr
ete
Pol
esE
lem
ents
Cai
sson
Supp
ort
Cai
sson
Supp
ort
Cai
sson
Supp
ort
Cai
sson
Supp
ort
1.Pa
nel
f.o
.b.
job
site
18.2
018
.20
20.4
020
.40
20.2
022
.20
22.2
022
.20
2.Pa
nel
inst
alla
tio
n4.
004.
003.
603.
603.
403.
403.
403.
40
3.M
odul
ein
stal
lati
on
4.00
4.00
3.80
3.80
3.70
3.70
3.70
3.70
4.S
truc
tura
lco
nnec
tion
s3.
005.
101.
903.
101.
701.
501.
502.
50
5.In
stal
led
Foun
datio
n8.
107.
105.
605.
305.
004.
504.
504.
40
Dir
ect
Fie
ldC
ost
37.3
038
.40
35.3
036
.20
34.0
035
.30
35.3
036
.20
Indi
rect
-50
%F
ield
Lab
or4.
104.
803.
704.
203.
603.
603.
604.
00
*Tot
alIn
stal
led
cost
=41
.40
43.2
039
.00
40.4
037
.60
38.9
038
.90
40.2
0
*Inc
lude
sov
erhe
adan
dp
rofi
tbu
tex
clud
esen
gine
erin
gan
dco
ntin
genc
y
being double-supported in a cantilever mode. The final designs reflect this
change to the double-supported cantilever system with their costs being re
ported here.
The technical scope and design basis for the array designs have been described
in Section 1. The cost estimate methodology developed for the two most promis
ing array support structures, the wood pole and the concrete caisson concept,
is presented in the following discussion. The basis for this methodology and
the assumptions made in these estimates are also given.
3.3.1 Field Costs
The field costs are based on the installation scenario as shown in Figure 3-2.
In this scenario, the materials required for the panel fabrication are supplied
by steel mills which are located in the midwest of the United States. The
materials are then transported by rail from steel supplier to a local steel
fabricator (assumed within 50 miles of the jobsite location). Panel super
structures are fabricated in the shop and then shipped to the jobsite by
trucks. Modules provided by the module manufacturers are also received at
the jobsite. Fabricated panels and solar modules are assembled at a site
assembly area before the final installation on to the foundation supports
(timber poles or concrete caissons).
Direct Field Costs. Direct costs were estimated for a 10 MW size array
field for each array concept and are subdivided into the following five
categori es:
o Panel costs (f.o.b. jobsite)
o Panel installation
o Module installation
3-18
PANELC><1 S::~L • C><J1-.....:P~A::;N;:;E;,;L;....;J1~C><J W/MOoULE.. C><1STEEL MATERIAL STEEL SITE FIELD
SUPPLIER FABRICATOR ASSEMBLY AREA INSTALLATION
•Ik><}----~~~----~PV MODULE
MANUFACTURER
Figure 3-2 Installation Scenario For A Prototype Array Field
3-19
o Structural connections
o Installed foundation
Emphasis for these final estimates was placed on projected construction costs
to build a typical 10 MW size array support system. The array support struc
tures considered here represent typical interior array structures.
The panel cost includes the superstructure material cost, the panel fabrica
tion cost, and the transportation cost of material from the steel mill to
the local steel fabricator shop and then to the jobsite (assuming a 50 mi
distance between the local fabrication shop to the jobsite). The panel
installation cost includes the labor cost of handling the panel superstruc
ture, loading and unloading, and panel erecting and installation. The module
installation cost includes the labor cost to install the 4' x 4' modules
received from module suppliers onto the support panels. The costs to trans
port the modules from the module suppliers to the jobsite were excluded in
the estimate.
The cost of the structural connection between the panel and the foundations
was separately estimated in order to identify any design/cost sensitivity.
This cost included the material cost of the connection and the cost to
fabricate it. However, different connection designs did not vary greatly in
costs. The installed foundation costs included the material and labor costs
to install the foundation by a subcontractor.
The final array designs, the 8' x 36' array having the wood pole and the
concrete caisson foundations, are described in Section 1. The direct cost
breakdown in unit costs of those designs are given in Table 3-7. The con
crete caisson design had a total direct cost of $32.60/m 2 while the timber
pole concept showed $33.40/m2•
3-20
Table 3-7
UNIT COSTS OF THE ARRAY STRUCTURE HAVING THE DOUBLE-SUPPORTEDCANTILEVER SYSTEM
(Estimates are based on 10 MW field size quantities and use mid-1981 $)
$/m2 IB' x 36'
Arrays
Extended TimberCost Concrete Poles
Elements Caisson Support
1. Panel Lo.b. jobsite 18.80 18.80
2. Panel installation 3.40 3.40
3. Module installation 3.70 3.70
4. Structural connections 1. 70 2.80
5. Installed Foundation 5.00 4.70
Direct Field Cost 32.60 33.40
Indirect - 50% Field Labor 3.60 4.00
*Total Installed cost = 36.20 37.40
*Includes overhead and profit but excludes engineering andcontingency
3-21
Indirect Field Costs. To make reasonable estimates of indirect field costs
for the const ruct i on project, it was necessary to look into the overall
construction plan and services. Three major cost elements of the indirect
costs include temporary construction facilities, construction equipment and
field office costs. These cost elements were estimated in detail with re-
ference to Bechtel's construction experience. Miscellaneous construction
services were estimated at an allowance of $0.80 per di rect labor manhour.
Preliminary check-out and project insurance were excluded in this estimate
by definition. Because of similarities in the repetitive nature of the
installations and in the total direct field costs, indirect field costs were
estimated at 50 % of the total di rect labor cost. As previously shown in
Table 3-7 for the preferred 8' x 36' span arrays, the timber pole concept
had indirect costs of $4.00/m2 while the concrete caisson concept was
$3.60/m2•
3.3.2 Qualifications, Exclusions, Assumptions
The qualifications and exclusions itemized in Sections 3.1.4 and 3.1.5 apply
to the final cost estimate except as noted below:
o Installation costs of solar modules are included in the estimate
o Engineering services and a contingency are excluded from the estimate
o The estimate for a large array field is based upon the engineeringdesign of an inner-field array structure. It is assumed that thisestimate is adequate for a large field that includes the perimeterarray st ructures.
3.3.3 Summary
The final cost estimates, their basis and methodology, for the installation
of a 10 MW photovoltaic array field have been presented. Based on these esti-
3-22
mates, the 8' x 36' span arrays having either timber pole or concrete caisson
foundations achieved the lowest cost.
As previously shown by Tables 3-5 and 3-6, the total installed costs of con
structing a 10 MW photovaltaic array field were estimated in the range of $2.8
million to $3.2 million based on mid-19Bl dollars. In unit costs, the total
installed costs for the preferred 8' x 36' span array came to $36.20/m2 for
the concrete ca i sson foundat i on concept and $37.40/m2 for the timber po Ie
concept.
3-23
Section 4
Prototype Hardware - Task IV
This Section summarizes the work performed in Task IV to select one of the
low-cost designs described in Section 1 in order to fabricate and install
an array for demonstration of the support system. The most attractive
design for this purpose was determined to be the 8' x 36' panels mounted
on reinforced concrete caissons (see Section 1.5). This design gave one
of the lowest cost estimates for large field applications and represented
very practical design features easily utilized for mass production situations.
Two panels, each 36 ft long, mounted on three concrete caissons, typical
of the end of any row in a field, were authorized for the demonstration
construction. A site was selected for this in the Photovoltaic Test Facility
area at Sandia National Laboratories in Albuquerque, NM and design criteria
were correspondingly formulated.
All of the low-cost designs were developed for specified, hypothetical 4
ft square photovoltaic modules. Since modules of this size were not cur
rently available on the market, and since a functional array was not needed
to demonstrate the structure, the construction work included the design
and fabrication of dummy modules made of tempered glass sheets in metal
frames. This also saved cost for the installation.
The following sections describe the development of the array design criteria,
the design of the array structures, the development of dummy solar modules,
the development of the necessary construction specifications, and finally
several aspects of the overall construction activity.
4-1
4.1 PROTOTYPE ARRAY DESIGN CRITERIA
The earlier design work showed that the major loading experienced by the
arrays comes from regional winds. An elementary approach is to use a single
wi nd speed value, perhaps taken from ANSI Code wi nd speed maps. Thi s
approach, using a single wind speed, was convenient for the large array
field design studies (see Section 1.3.1) but was not sufficient for the
demonstration array. For a specific site it is desirable to determine the
directional nature of the wind regime, to identify prevailing wind speeds
and di rect ions. Accordi ngly the expected wi nd regi me was investigated for
the designated site in Albuquerque.
Wind speeds and directions for Albuquerque were sought from studies of local
climatological data and from local site-related records. Regional data was
found for a 40-yr period from the Albuquerque Airport Weather Office.
Site records were found that related to wi nd energy studi es. After some
discussions with Sandia, estimates were made of the short-duration peak
gusts and for the longer duration maximum winds. These results were adopted
for the design wind criteria and are tabulated as follows:
Maximum Wind Speed (mi /h)
Wi nd Short Duration Long DurationDirection Peak Gusts Winds
N 40 30
E 90 60
S 90 65
W 65 *prevail i ng NW * 10
* Uncertain but not critical
4-2
Note the directionality of winds for this site. The maximum winds are expected
from east and south, and are only significant from the south. The north
sector, most sensitive for uplift calculations, has relatively low winds.
The other improvements needed for the design criteria of Section 1.3.3 were
in the soil condition parameters. A critical review of the site soil
parameters was made. The soil is described as a medium plastic, sandy clay
with internal friction angle 35° and cohesion value 1500 psf. Since a
sandy clay soil is not expected to have an internal friction angle greater
than 20°, a compromise was decided and the foundation design assumed 15°
for this angle. Other foundation criteria remained as before for the
prototype array designs.
4.2 DESIGN OF DEMONSTRATION ARRAY
Si nce the wi nd domi nates the 1oadi ng, and has di rect i ona1ity specifi ed in
the previous section, then the demonstration array is rather similar to the
south row of a large array field. Figure 1-30 shows the caisson variations
expected over a 1arge fi e1d wi th a fi xed wi nd speed speci fi cat i on. The
demonstration array caissons were calculated to require 5-ft depth at the
ends, and 7-1/2 ft depth at the center.
Rebar cages for the caissons were designed to be prefabricated and included
two anchor bolts prewe1ded to the rebar. This allowed the rebar to be
suspended from the formwork by the anchor bolts before concrete was poured
and eliminated the conventional procedure of setting bolts during or after
the pour. Grouting was eliminated from this design in order to again reduce
field costs. Caissons were simply specified to have horizontal top surfaces,
trowelled smooth and flat.
4-3
Panel details are as given earlier in Figures 1-27 to 1-29 and in the
drawings of Appendix A. This desi9n required 7" x 5" x 3{16" torque tubes
as the main structural members. All panel steelwork was degreased, prime
coated, and finish painted with white exterior enamel for display purposes.
Two 1ift i ng 1ugs were ori gi nally specifi ed at the ends of each tube but
they were finally moved 8 ft in from each end in order to reduce the
flexing, and risk of damage to glass, during the lifting and handling
operations of completed panels.
4.3 DUMMY SOLAR MODULES
Large photovoltaic modules, 4-ft square, were conceptualized during the
array field development (see Section 1.5.1). That work was put on a more
practical basis by investigating actual materials that might be used for
the dummy modules of this construction task. The resulting dummy modules
are shown in Figures 1-28 and 1-29, as well as being detailed in Appendix
A. Off-the-shelf materials were successfully used for this demonstration
and consisted of extruded aluminum frame members; special metal corner keys
to lock the square frames together; standard rubber gaskets to protect the
glass edges; and square sheets of 3{16 in. tempered glass. The resulting
dummy assemblies simply provided a reasonable physical representation of
the structural characteristics of future large solar photovoltaic modules.
4.4 CONSTRUCTION SPECIFICATIONS
The approach to generating suitable specifications was first to utilize
the guidelines of the Construction Specifications Institute which prOVide a
standa rdi zed approach to the development of documentation. Thi s was then
altered to suit the specific requirements of this job, to recognize this
4-4
was a small construction activity to be done by a local contractor. It
was important not to burden this work with unnecessary constraints, over
restrictive tolerances, or excessive inspections. The specifications
document that resulted from this approach is given in Appendix D, and this
was used for the construction bid package released in May 1982.
To simplify the procurement process, Bechtel supplied the metal frames,
corner keys, and rubber gaskets for the dummy modules, while the construction
contractor supplied the glass. The specifications were thus written so
that the contractor took care of all other procurement, fabrication, and
installation work.
Concrete was assumed to be supplied by a local ready-mix firm and was
specified for required strength. Aggregate size and slump were sufficient
to give suitable workability in the caissons. Vibration was also specified
because of the need to densify the concrete over the depth of the caissons
and around the rebar cages. No additives were thought necessary for this
job, however accelerated strength gain for concrete in a large array field
might benefit a construction schedule.
The specification was written to maximize shop operations and to avoid
welding or module assembly in the field at the site. This proved to optimize
the construction work.
4.5 CONSTRUCTION OF DEMONSTRATION ARRAY
A bid package for the construction work was assembled by Bechtel and released
to several Al buquerque contractors in May 1982. Responses were recei ved,
evaluated, and construction was started by Cardenas Construction Company in
4-5
early June 1982. Materials procured by Bechtel were delivered to Cardenas
by 1 June 1982 so that module fabrication could begin in early June along
with fabrication of the steel panels.
The three caissons were augered and formed, rebar was placed and concrete
was poured on a single day. The 16-in. diameter holes averaged 6 ft deep
and were augered in an average of 10 minutes each, including movements
of the truck-mounted auger between holes. The sandy clay soi 1 requi red no
1i ner as expected. Formwork for the caisson above ground was custom-bui It
at each hole and would not be satisfactory for large scale installations.
For a large field, the formwork would be prefabricated, removable and re
usable. The rebar cages were simply suspended within each hole from the top
side formwork by using the anchor bolts. Concrete was poured directly from
the delivery truck-mixer and vibrated during the pour. Curing of the
concrete was promoted by 1eavi ng the tube forms in place for 7 days before
stripping. For this mix design, available data showed that 90% of the design
strength would be achieved at 21 days, with full design strength (3000 psi)
attained at 28 days after pouring.
Fabrication of the steel panels was straightforward, being a simple design
involving some cutting, drilling, and welding. Three support assemblies
were fabricated and final spacing on the torque tube was determined from a
field measurement from the caissons. Panels were cleaned and painted as
specified before the dummy modules were attached.
Care was required in welding the tees to the torque tubes because of the
attachment method of the dummy modules. Since the attachment bolts were
horizontal, in the plane of the panel, no adjustment was readily available
4-6
for slight variations in the sizes of modules. Future attachments of
production photovoltaic modules should be by means of bolts perpendicular
to the plane of the panel and using slotted holes. This will greatly expedite
attaching modules to the final panel assembly.
Fabrication of the dummy modules by Cardenas was expedited by using simple
jigs for repetitive cutting of metal sections and rubber gaskets. A two-man
crew manufactured the metal frames and completed the final assembly with
the tempered glass. No problems were experienced with this work.
Each panel was shipped complete with the dummy glass modules except for two
modules, removed symmetrically from each panel, to allow easy access to the
lifting lugs. Each panel was easily handled by the hydraulic crane unit of
a self-unloading tractor-trailer vehicle. This vehicle delivered and
unloaded each panel at the site, released its trailer nearby, then the
tractor/crane portion returned to erect the panels. A different arrangement
could of course be anticipated for a large array field installation.
Each panel stayed horizontal when suspended by the crane. As the support
seats came down over the anchor bolts and touched the concrete, the whole
panel would rotate to the correct slope, while the support seats made full
contact with the flat concrete surfaces.
From beginning the lift to releasing the hooks after erection, the fi rst
panel took eight minutes and the second took seven minutes. It took six
mi nutes between pane Is to change the crane pos i ti on and prepare for the
next 1i ft. Fi tt i ng the four dummy modu Ies to fi ni sh the paneI s took IB
minutes to complete.
4-7
Altogether it took about 40 minutes for the complete erection of two panels
or about 20 mi nutes average per pane1. Indeed thi sal so suggests that thi s
demonstrat ion had an erect i on rate of 108 ft x 8 ft ~ 864 sq ft of array
per hour, using crews in an unfamiliar process and without special training.
The construction work was completed later with a small amount of touch-up
of the paint and slight adjustment of one module.
4.6 REVIEW OF CONSTRUCTION
Construction of the demonstration array cannot easily be extrapolated to
the high production type of activities required for a future large-scale
array field. However, despite the overall relative inefficiency of this
small construction effort for a two-panel array, there were some promising
indicators for future applications.
A major activity for producing the foundations is augering the caisson
holes. For an average depth of 6 ft, this job demonstrated a production
rate, using a single truck-mounted auger, of more than 45 caisson holes for
an eight-hour day. It would be expedient to match this with prefabricated,
removable, reusable cylinder forms that could be deployed at the above
rate over each caisson hole for the future large array fields.
The module installation effort for each panel had been estimated, for a
large array field, as requiring five manhours per 18 modules. Observation of
the Cardenas module assembly work suggested that this could be improved to
at least four'manhours per 18 modules. This anticipates a reduction of direct
cost for this item from $3.70/sq m to $2.96/sq m for a large field.
4-8
Furthennore the panel installation activity had been estimated for a large
field as requiring three manhours per panel (among other costs). It was
clear from observing erection of the demonstration array that one manhour
per panel would be a reasonable expectation for a production operation.
When this change is incorporated into the cost data, this element of direct
cost reduces from $3.40/sq m to about $1.90/sq m.
The above two items suggest at least a 6% reduction may be expected for
the direct costs predicted in this study for a 10 MW array field (see Tables
3-5 and 3-6). To this extent the construction of the demonstration array
provided confi rmation that this basic structural concept, torque tube and
caissons with cantilevered modules, has significant potential for future
low-cost field developments. Figure 4-1 shows one of the demonstration
reinforced concrete caissons used to support the arrays. Figures 4-2 and 4-3
show some of the fabrication and erection activities that led to the com
pleted demonstration two-span array shown in Figure 4-4.
FIGURE 4-1 DEMONSTRATION REINFORCED CONCRETE CAISSON
4-10
4-12
co
MI...
Section 5
WIND TUNNEL TESTS - TASK V
5.1 SCOPE
This section summarizes the results of a Phase II study of a series of wind
tunnel tests on flat-panel photovoltaic array structures performed under
the direction of Bechtel at the Colorado State University at Fort Collins,
Colorado. These tests were to extend the results of a previous Phase
I study (Reference 5-1) of wind loads on photovoltaic arrays.
A seri es of wi nd-tunnel tests had been conducted in the Phase I study to
determine the effect of various design parameters on wind loadings for
different wind directions and wind profiles on a single array and on indi
vidual arrays at different locations in a large array field. The tests
showed that arrays on the perimeter of the field are subjected to very large
1oadi ngs, but that these 1oadi ngs may be drast i ca11y reduced by fences de
signed to act as Wind barriers. However, few measurements were made concern
ing the effect fences have on wind loadings on arrays further in the field.
All the previous array field tests were conducted with the height of the
array H = 2.0 ft above ground.
Subsequent analysis of the wind-tunnel data from the Phase I study by Bechtel
showed that more information was needed for wind forces at the interior arrays
of the field and particularly without the use of wind fences so that economic
comparisons could be properly considered.
Therefore the present wind tunnel tests extended the previous study and exam
ined more closely the mean forces and moments on individual arrays in the
fi el d with H = 1.5 ft, wi th and without a fence. The standard fence used
5-1
in the study had 30 % porosity and an additional corner fence was used to
reduce the corner turbulence identified in Phase I.
It is important to know that the wind tunnel tests used a Ill-power law velocity
profile for the atmosphere which correctly causes turbulence near the ground.
This turbulence envelops the arrays and provides the uncertainty for designers
tryi ng to use convent i ona1 methods for any low structures. The wi nd tunnel
results are the averages of many samples taken over a few seconds. Hence,
gusting and turbulence effects are excluded. Separate analyses are required for
those effects, such as for turbulence as discussed in Section 2 and for gusting
as given in the Codes or from other wind engineering sources.
The detailed results of the present wind tunnel tests are presented in Appen
dix E, adopting the previously defined aerodynamic dimensionless force and
moment coefficients and notation used in Section 1.3. These results are now
summarized and recommendations made for design.
5.2 SUMMARY ANO CONCLUSIONS
The wind tunnel coefficient CN, Cms , and Cmz are defined in Section 1.3.
The maximum experimental coefficients taken from the present wind tunnel tests
are shown in Table 5-1 and recommended values to use for design are given in
Table 5-2. The coefficients are given for northerly and southerly wind loads on
perimeter and internal arrays, with and without fences. The array ground clearance
is 1.5 ft and the fence is 5 ft high. The perimeter arrays as shown in Figure
5-1 are defined as those arrays which are right at the four sides of the field.
The coefficient Cms is a newly introduced wind-force coefficient which was
not measured in the previous Phase I study. From this coefficient the eccen-
5-2
Table 5-1
MAXIMUM VALUES OF FORCE COEFFICIENTS
(Experimental Wind-Tunnel Test Results by Colorado State University)
Condit ion* Peri meter Rows Internal RowsNortherly Southerly Northerly SoutherlyWinds Winds Winds Winds
CN Cms Cmz CN Cms Cmz CN Cms Cmz CN Cmz Cmz
Wi thout fe nee 0.79 0.21 0.05 0.5:! 0.4b 0.09 0.27 0.2e 0.04 0.08 0.07 0.0"
With fence 0.37 0.13 0.02 0.27 0.25 0.01 0.33 0.1 0.01 0.1 0.13 0.01
(signs omitted, see Appendix E)
Table 5-2
RECOMMENDED FORCE COEFFICIENTS FOR DESIGN CRITERIA
Condit ion* Perimeter Rows Internal RowsNortherly Southerly Northerly SoutherlyWinds Winds Winds Winds
CN Cms Cmz CN Cms Cmz CN Cms Cmz CN Cms Cmz
Without fe nee 0.80 0.25 0.05 0.55 0.5 0.1 0.3 0.3 0.05 0.1 0.1 0.03
With fence 0.40 0.15 0.05 0.3 0.25 0.02 0.35 0.1 0.02 0.1 0.15 0.02
(signs omitted, see Appendix E)(NORTH) 0°,360°
*Notes: 1. For North wi ndsFor South winds
0<; = 3000 600~ = 1200 2400
2700--f-
WINDDIRECTION
2. For East-West winds ~ = 60 _ 1200 °180 (SOUTH)0( = 240 - 3000 coefficients
are uncertain but not critical.
3. Design criteria are for H = 1.5'
5-3
10W~~@W~~~~~~
l00Y~ I I I ItI
f7Z//ad +II
l00Yhi +W~ PERIMETER ARRAYS - AT EDGES OF FIELD
__IINTERIOR ARRAYS
Figure 5-1 Typical Corner of an Array Field
5-4
tricity of the normal wind load Fn along the span of the array can be found.
This coefficient also determines the torsion on a foundation. As explained
in Section 1.4, for long spans torsion governed the depth of single-support
foundations. Therefore, double-supported arrays were used to eliminate
torsion problems. This could not be identified in the earlier Phase I work.
Table 5-1 shows that the use of fences drastically reduces the values of
eN for the perimeter arrays but was ineffective for the interior arrays.
One explanation could be the perimeter arrays provide wind protection for
the interior arrays much in the same manner as the fence provides wind protec
tion for the perimeter arrays. Therefore, the interior arrays are not affected
by the use of fences. The question of us i ng a fence cou 1d now be answered
from an economic standpoint considering perimeter arrays alone. As explained
in Section 1.6, a preliminary cost estimate for a large array field showed
the use of a fence having 30 percent porosity was not economically advantageous.
5-5
REFERENCES
1-1 Bechtel National, Inc., Design of Low-Cost Photovoltaic Arrays, Volume 2of 3, prepared for Sandia Laboratories under Contract Number 05-6195,July 1979.
1-2 Bechtel National, Inc., Wind Design of Flat Panel Photovoltaic ArrayStructures, prepared for Sandia Laboratories, SAND 79-7057, 1980.
1-3 American Institute of Steel Construction (AISC), Specification for theDesign, Fabrication, and Erection of Structural Steel for Buildings,November 1, 1978, with Commentary.
1-4 American Iron and Stee1 Institute (AISI), Specification for the Designof Light Gage Cold-Formed steel Structural Members, 1968 Edition.
1-5 American Concrete Institute (ACI), Building Code Requirements forReinforced Concrete, (ACI 318-1977).
1-6 American Institute of Timber Construction (AITC), Timber ConstructionManual, 1974.
1-7 International Conference of Building Officials, Uniform Building Code(UBC), 1979.
1-8 Wilson, Abraham H., Low-Cost Solar Structure Development, prepared byJet Propulsion Laboratory, Pasadena, CA, DOEIJPL-I012-53, June IS, 1981.
1-9 Design Practice Manual, "Active Solar Energy System", DOE National SolarData Program, October 1979, Solar 10802-79-01.
1-10 Residential Photovoltaic Module and Arra Re uirement Stud, June 1979,o J 1
1-11 American National Standard, Building Code Requirements for Minimum DesignLoads in Buildings and Other Structures, ANSI A58.1 - 1972.
1-12 Poreh, M., Peterka, J. A., and Cermak, J. E., Wind-Tunnel Study of WindLoads on Photovoltaic Structures--Phase II, Final Report prepared forBechtel Group, Inc., by Colorado State University, Fort Collins, Colorado,January 1982.
1-13 Ellington, B., Galambos, 1. V., MacGregor, J. G., and Cornell, C. A.,"Development of a Probability Based Load Criterion for American NationalStandard A58," NBS Special Publication 577, 1980.
1-14 Albuquerque Testing Laboratory, Inc., "Foundation Investigation Report",Report No. 97105, December 19, 1978.
1-15 Broms, B. B., "Lateral Resistance of Piles in Cohesionless Soils", Journalof the Soil Mechanics and Foundation Division, SM3, May 1964,
R-l
REFERENCES (Cont'd)
1-16 Woodward, Jr., R. J., Gardner, W. S., and Greer, P. M., Drilled PierFoundations, McGraw-Hill, 1972 , pp. 68-76.
1-17 Merkle, Douglas H., Full Scale Load Tests of Experimental SolarCollector Foundations, prepared for Sandia Laboratories by AppliedResearch Associates, Inc., SAND80-7076, June 1981.
1-18 Terzaghi, K., Peck, R. B., Soil Mechanics in Engineering Practice,2nd Edition, John Wiley &Sons, 1967, pp. 188.
1-19 Jumikis, A. R., Foundation Engineering, International TextbookCompany, 1971, pp. 43.
1-20 Stall, U. W., "Torque Shear Test of Cylindrical Friction Piles",Civil Engineering - ASCE, April 1972, pp 63-65.
1-21 Randolf, M. F., "Piles Subjected to Torsion", Journal of theGeotechnical Division, ASCE, Vol 107, No. GT8, August 1981.
2-1 Miller, R. D., and Zimmerman, D. K., "Wind Loads on Flat PlatePhotovoltaic Array Fields (Non-Steady Winds)", Phase IV FinalReport, Boeing Engineering and Construction Co., Seattle, Wash.,August 1981.
2-2 "STARDYNE, A General-Purpose Structural Ana lys i s Computer Code,"System Development Corporation" 2500 Colorado Ave., Santa Monica,CA 90406.
2-3 Cervalos-Candau, P. J., and Hall, W. J., "The Commonality of Earthquake and Wind Analysis", Dept of Civil Engineering, Report NSFjRA800022, University of Illinois, Urbana, Illinois, January 1980.
2-4 Davenport, A. G., "Rationale for Determining Design Wind Velocities",ASCE, Structural Division, Vol. 86, 1960.
2-5 Fung, Y. C., An Introduction to the Theory of Aeroelasticity,John Wil ey and Sons, New York, 1955.
2-6 Simiu, E., and R. H. Scanlan, Wind Effects on Structures, John Wileyand Sons, 1978.
2-7 Miller, R. D., and Zimmerman, D. K., "Wind Loads on Flat PlatePhotovoltaic Array Fields", Phase II Final Report, Boeing Engineeringand Construction Co., Seattle, Washington, September 1979.
R-2
REFERENCES (Cont'd)
2-8 Communication with Prof. Holt Ashley of Stanford University, PaloAlto, Cal ifornia, in the period from September to November 1981.
2-9 Miller, R. D., and Zimmerman, D. K., "Wind Loads on Flat PlatePhotovoltaic Array Fields", Phase III Final Report, Boeing Engineering and Construction Co., Seattle, Washington, April 1981.
2-10 Clough, R. W., and J. Penzien, Dynamics of Structures, McGraw-Hill Co.,1975.
2-11 Newmark, N. W., and W. J. Hall, "Development of Criteria for SeismicReview of Selected Nuclear Power Plants", U.S. Nuclear RegulatoryCommission, NUREGjCR-0098, 1978.
2-12 Thomson, W. T., Vibration Theory and Applications, Prentice-HallInc., 1965.
2-13 Blodgett, O. W., Design of Welded Structures, Published byJames F. Lincoln Arc Welding Foundation, 1976 Edition
5-1 Poreh, M., Peterka, J. A., Cermak, J.E., "Wind-Tunnel Study ofWind Loads On Photovoltaic Structures", Final Report prepared forBechtel National, Inc. by Colorado State University, Fort Collins,Colorado, September 1979.
R-3
ApPENDIX A
CONSTRUCTION DRAWINGS
FOR LARGE ARRAY FIELD
L
• I 0 I . I . I 0 I u I . <I
I ' u I
~~~h~,
l , i -IM
, ~,I -:.~ ~ I •.' ~ ~i! ~ •
, -i I ~ c ai ~'" ~ c:I,~ !i.1 1 f ,~
i ~ : ~~ ,II ~ ~ ~f ~~
I ;-. I ~,). Ci-,- <" ,
-.-. ,-,
I
~. ,;1 I-§~
r ,Q;' !i[-,-- ~~ ~-- ,FYi ·. ., ~\ Ii "" illlJI\\-- 1-
-
- f]l ~ If-
';~\, ~;,I~".
I t ~ i 1 I ' ·-,
? t II I I'_.-, i~~ ';1-, ~ J ' . '-, - • I' l" '1-, ;Il[~- ,i ,1-
-<-, 'l+ r- X). ,\
1 ~J- "-, 1-,- "-_1[':---1 ~
-'". i;
1::
t~~, ,,
- , ~ 'Ii- ,
t91, ., i-rI,
~~ .--1 --- ,, .Ii __ 1, "J: '",, }'I-, , Id - ,.--, -T'\t i,
:~; ',:,
U ·..~ cr .., I : I, , ·, ,, ';- ~- _: t
,< s[~rl :9
,.I
~ --II: j
I • II"\\~ ,, ~, '.,, ,
:f-,
": r\
-- : J ,!~: iJL_
, , I • i_I ' ,-- ·. !----IC-,-
!
I I I I I I I
- ----
I I I o I u I
I-I
.~ i i _ _
~ -~._._--t--_.-
I
_.",.
4;
. .
.
I I I I I I I
------- ~- - - --~-------
1 o I • 1 I o , o I I <
I-
I-
. "
; II~~i
1
I
I
§~; ib,,~'
,I
rI",
r Illt_, ;:r~
!I.
0!1'~··'I ~ l'~)
F._~:
I--J
1-'
~!
--
\'-
-
I-
l
I--
1-1I--
I I I I 1 I
/p:
!"'~i
/,f"
~
~~
\~
jt;'L
I\"
J\
\I,
Jr.
.....,\
~I.
f
~IJ~'
~c,.
\c,~li
c;c',
I~k~R\~~'r~~~§
,!4
:I'
",
'-,\J~
16"
"U">
AnO
N3'~
I:>I-l
'~!i-O
li"''c
l-(
iTlYP
)I'~,
(1,'p
,~~
_/l-
:...
c+-~
r-y-
,,\,\~
~I
5'
5
ItFOI
.JNDAT
lO~1I4
'RtM
X
~ ~
3!~-,<1
liz"
~i-
3116
Vli
)26
"(T
'r'P
l6;1~
I~..!
0-"
'11~2"
lTYP
},&
r@
~,..<
m,7
1.
CLa
I.
'I
II·.
...>:
T~.
r~~~c..
~::~
~!.;
~o,
,
+b"f'~
I~Ljj
'---
.'.~
DETA
IL3
ImERNALARRA~
-S
LID
ING
SEAT
'3l1
6Vli
li'"=
'4
INIT
RN
AL
ARR
AYCO
NNEC
TIO
N
",\~';
EXlE
RI.(
ALAR
R,l.
YCO
NNEC
TIO
N
.•,
"7'
D-I~5
SH.
2C
f",
'0.....
PR
£LIM
INA
RV
ISSU
E
ISSU
EDfO
RFI
NAL
REPO
RT
ISSU
EDFO~
REVI
EW
~Y',O
"i,
.;/
1."'-
ill'
"I
....·~
....
;:I:
I-I
IY~
!~
BE
CH
TE
LG
RO
UP
,IN
C.
s"'~
FR
.o,N
C
".,
ISC
C
INTE
GRAT
EDST~UCTURE
DES
IGN
FOR
PH
OTO
'/a.T
AIC
ARR
AYS
SA
ND
IACO
NTRA
CTN
o.62
-9
m
LARG
EAR
RAY
FIEL
DI~TAUATlml
fOU
llnA
HO
NS
AN
lJO
HA
ILS
II
-..1
.14
""
f0
IIr-
--t.
:,"'1
-,.~
I---
--<
S=
--__
Ln--
'--- A
-5
~i "
~o"
1I
~-----,
6"
IJli
Wfif
.2I
3J14
"11
12
'"~"
~f-
-I!
I;w
61\1
112
<'
~-
-=~-
'-.
~112~1
'~
~!,l
0~
~.!
~~
!I~;:,
--=::::r
'I
~'31
16V
l1l2
-4----~--I
tXlF
Ri'L
IIlARR~~IL
4 51!P
ING
SE
AT
GA,
""
O£T
A!U
SCA
LI11
2"·
T"
-, "
_.I---
i-~
1'1]--
---I;-
~]/
4'
--
YF..
'"'"~.!
\'W~
~L~I
Htj_
_\'-<
Ie,I:;
:
I"'. "
'"
Jf:h
I!"
.-I
l.
,i=
J'~-
,1-
-----"
--'~U
4'R
.""
i__
,_,,
_5
"I
DET
A.!1
..1
,
1
<•
II
u•
,,
"
,~
---
illi[,II'Ii, ~
iii ~~ iii:, '
:
" ii 't::
-~.
•
,"
-4: " ~,
, , -I, '
- 1-
i
APPENDIX B
MODE SHAPE PLOTS OF PHOTOVOLTAIC STRUCTURAL SYSTEMS
(CONCRETE SINGLE-POST AND TIMBER DOUBLE-POST DESIGNS)
Figure
B-1
B-2
B-3
B-4
B-5
B-6
APPENDIX B
ILLUSTRATIONS
First Mode (6.60 cps) of Single-Post Concrete Design
Second Mode (8.76 cps) os Single-Post Concrete Design
Third Mode )8.89 cps) of Single-Post Concrete Design
First Mode (10.04 cps) of Double-Post Timber Design
Second Mode (10.71 cps) of Double-Post Timber Design
Third Mode (12.85 cps) of Double-Post Timber Design
B-i
Page
B-1
B-2
B-3
B-4
B-5
B-6
~=:':J:~:=w...;=_----.s •• __
.. Xl
Figure B-' First Mode (6.6 cps) Of Single-Post Concrete Design
B-1
Figure 6-2 Second Mode (8.8 cps) Of Single-Post Concrete Design
B-2
.. Xl
---r---'- - -r I
L __ .l---.l---
- - - --r ,- ,
- --r I I
'" - --r- --r I 1 I
- - --\- --r I I I
- - --\-1 I- - --\-I I J_ - - --\-1 I II
,- ---=+ - ---=f - - -. .-
- - --\- I I I
- - --\- 1 1
- - --\I,
1 I- - ---II I
I I I - - -'__ .-J
I- -I - - -'
__-:::J -- -- - - ---I
- - - -Pt----11----
- - - 1-- - , ,1----l- - - - - - - 1 I I
W1---- I I lL - -I ,~ --I I
~ --,1
J- - -,
I 1\I I I, -.l..--- - -
rI
I
Figure B-3 Third Mode /8.9 cps) Of Single-Post Concrete Design
B-3
PANEL
POST POST
.. Xl
Figure 8-4 First Mode (10.0 cps) Of Double Timber-Post Design
B-4
__ --::l:::l:::J:::l:::3:-- __= =---=:_~ __ i: __ ~ __ ~ :=~=_~_==
... xl
Figure 8-5 Second Mode (10.7 cps) Of Double Timber-Post Design
8-5
/
Figure B-6 Third Mode (12.8 cps) Of Double Timber-Post
B-6
APPENDIX C
DERIVATION OF FLUCTUATING WIND PRESSURE POWER SPECTRAL
DENSITY FUNCTIONS IN REFERENCE 2-1, APPENDIX C
APPENDIX C
C.l RELATIONSHIP BETWEEN AERODYNAMIC PRESSURE AND PRESSURE COEFFICIENTS
(a) Surface pressure P and the associated coefficient Cp:
Cp = P/qref (C-l)
where q f is the stagnation aerodynamic pressure at there2reference height (10 meters), and is equal to .00256 V f wherere
Vref is the wind speed of 43 fps at the reference height.
(b) For net pressure {:, P (di fference between the front and back
surface pressures) the relationship is:
Cp = (:, P/qref
B.2 CORRELATION FUNCTIONS
(C-2)
For this analysis, auto- and cross-correlation functions are defined by
R i j
T
fja
[Cpi (t) - C pmeani ] [CpJ" (t+"t) - C " ] dtpmeanJ
(C-3)
where CPk(t) and Cpmeank are instantaneous and time-averaged values
of pressure coefficient with respect to tap number k. Defining a fluc
tuating component of pressure coefficient, C'pk(t). as
(C-4 )
C-l
the auto-correlation function, at T = 0, becomes for one tap
Rii (0) =
=
Cpi (t) Cpi (t)
2C prms i (C-5)
in which an overbar denotes a time averaging.
B.3 POWER SPECTRAL DENSITY FUNCTIONS
Power spectral density functions are defined by
¢ij (N) =
00
4f R.. ( T ) exp (-2 1T NT) d T11
o(C-6)
where j in the exponential function refers to j2 = -1. The integral
property of the auto-spectral function requires
00
f 1> •• (N) dn = R .. (0) = C2 .11 11 prms1
o(C-7)
Because of this property, the power spectral density can be normalized
by
¢ .. (N) =1J
I 1> ij (N) I
Cprmsi • CprmsjWith Units
(C-8)
It is also common practice to normalize the frequency N by
N* = NcY;:ef
where Vref is the reference wind velocity.
C-2
(C-9)
The auto- and cross-spectral analysis was performed digitally by a
Fourier Transform subroutine using standard techniques. The transforms
were performed on 8 time segments, each with 2048 samples in length,
for each pressure record (recorded at 500 samples per second). Trans
forms were combined to form cross-spectra, and appropriate averaging
across data points in frequency and time segments was performed to re
duce normalized standard error of the spectrum. Because of memory
limitations, the auto- and cross-spectra were only calculated to a
maximum frequency of 125 Hz (N*" 1.0). This frequency retained vi rtually
all the energy in the fluctuating pressures. Normalized standard error
of the cros s-s pect ra reached a maximum 11 pe rcent at the lower frequenci es.
C-3
Bechtel Group Inc.April 1982
TECHNICAL SPECIFICATIONS
FOR THE
FABRICATION AND CONSTRUCTION
OF
DEMONSTRATION DUMMY PHDTOVOLTAIC ARRAYS
AT
SANDIA NATIONAL LABORATORIES
ALBUQUERQUE, N.M.
APPENDIX D
Technical Spec. No. 14614/2Revision 1, June 1982
14614/2Rev. 1
1.0 GENERAL
1.1 OESCRIPTION OF PROJECT
A. The work of this contract comprises the general construction ofsteel support framework and concrete caisson foundations, and thefabrication and installation of dummy photovoltaic modules.
The project is on the site of the Sandia National LaboratoriesTotal Energy Facility, 12th Street near F-Street, KirtlandA.F. Base, Albuquerque, New Mexico.
B. The metal frame and rubber gasket materials for the dummy moduleswill be furnished by Bechtel.
1.2 WORK TO BE DONE
A. Auger three (3) holes to construct 16 in. diameter reinforced concrete caissons as follows: 2 holes 5 ft deep, 1 hole 7 ft 6 in.deep.
B. Pre-assemble rebar cages and anchor bolts for concrete caissons,install sonotube forms for pedestal extensi ons above ground,place and compact the concrete, provide top surfaces smooth andlevel using a suitable template.
C. Fabricate structural steel panel frames.
D. Fabricate dummy solar modules, nominally 4 ft square, andassemble into the panel frames.
E. Install the panel assemblies (2) onto the support pedestals.
2.0 SITE CONDITIONS AND EXCAVATION
2.1 SUBSURFACE CONDITIONS
2.1.1 General: a soils investigaiton report has been prepared forthe site of this work and is filed with the Owner.
2.1.2 Availability: The soils investigation report may be inspectedat the office of Sandia National Laboratories Total Energy Facility (Owner),and copies may be obtained at the cost of reproduction and handling uponrequest accompanied by full payment and addressed as specified by the Owner.
2.1.3 Use of Data: This report was obtained only for Bechtel's use indesign and is not a part of the Contract Documents. The report is availablefor bidders' information, but is not a warranty of subsurface conditions.
D-l
14614/2Rev. 1
2.2 EXCAVATING AND GRADING
2.2.1 Work Included: Removal and storage of surface gravels, excavatingfor caisson foundations, and replacing surface gravels during clean-up.
2.2.2 Excavating for Caissons: Clear surface gravel from site of caissonexcavations and stockpile for final restoration of the site ground surface.Excavate for caissons as shown on Drawing No. D-101. Clear loose materialfrom bottom of excavations before setting rebar cages. Take precautions toprevent entry of soil into excavations. All soil from excavating operationsshall be hauled to dump area designated by Owner and disposed in the mannerrequested by Owner.
2.2.3 Restoration: Stockpiled surface gravel shall be replaced on the groundaround the final arrays installation to the approval of the Owner.
3.0 REINFORCED CONCRETE CAISSONS
3.1 Reinforcing Steel: Provide complete, in place, all steel requiredfor cast-in-place concrete caissons as shown on the Drawing No. D-I0l.Materials specified are:
Bars: ASTM A-615-79, Grade 60
Wire: ASTM A-82-79,
Anchor Bolts: ASTM A-307
3.1.1318.shallrust,steel
3.2
Fabrication: In accord with CRSI Manual of Standard Practice and ACIWel di ng of anchor bolts to rebar in accordance with drawings. Barsnot be spliced. Remove dirt, grease, oil, loose mill scale, excessiveand foreign matter that will reduce bond with concrete. Keep reinforcingin position shown on drawings during concrete placement.
CONCRETE
Provide cast-in-place concrete, complete, in place, as indicated on theDrawing No. D-101 specified herein, and needed for complete and properinstallation of caisson foundations.
3.2.1 Concrete Specification: Provide ready-mixed in accord with ASTMC-94-8.
Cement: Type II, ASTM C-150-80
Fine and Coarse Aggregate: ASTM C-33-80. Coarse aggregate shallbe size number 67 (3/4 in. to No.4)
Slump: 5 in. ASTM C-94-80, and ASTM C-143-7
D-2
14614/2Rev. 1
Mix proportioning: suitable to achieve 28 day compressive strengthof moist cured laboratory samples of 3000 psi
Curing Materials: Liquid membrane, ASTM C-309-74 or wet sand.
3.2.2 Concrete Certifications: Provide the following Certificates and TestReports:
1. Manufacturers certification that materials meet specificationrequirements.
2. Ready-mix delivery tickets, ASTM C-94-80.
3.2.3 Placing Concrete: Convey concrete from mixer to final position bymethod which will maintain uniformity of mix, and which will not affect theposition of the reinforcing steel as shown on the drawings.
A. Consolidation Concrete: Use mechanical vibrating equipmentfor consolidation with a minimum vibrator speed of 7,000 rpm.Vibrate concrete minimum amount needed for consolidation.
B. Finishing: Strike and level concrete at top of caissons toel evati on shown on drawi ngs. Hand float fi ni sh for smoothlevel surfaces.
C. Curing: Keep topcuring compound,method approved
of caisson moist by use of liquid membranewet sand, continuous sprinkling, or other
by Bechtel. Conti nue curi ng for 7 days.
4.0 STRUCTURAL STEEL PANELS
Furnish, fabricate, mark for erection identification, pack, crate, or otherwise properly prepare for shipment, and ship to the site all structural steelindicated on the Orawing Nos. 0-101, 0-102, described in these Specifications,or otherwise required for proper completion of the Work.
4.1 MATERIALS
A. Steel Shapes, Bars and Plates:Fy 36 ksi steel: ASTM A36-77a
B. Structural Steel Tubi ng:Fy 46 ksi cold-rolled tubing ASTM A500 Grade B
C. Standard Threaded Fasteners:
1. Standard bolts and nuts: ASTM A307-78, Grade A
2. Plain washers: ANSI B27.2-1965, Type A
0-3
14614/2Rev. 1
o. Filler Metals for Welding: AISC Specification
E. Electrodes for Welding: Comply with AWS Code, usingASTM A233-E70 Series electrodess
F. Chrome oxide primer, prime undercoater and exteriorenamel finish paints.
4.2 FABRICATION
A. Fabricate structural steel in accordance with AISCSpecification Structural Steel for Buildings
B. Shop connections: All welding shall be done in the shop.Welding procedures, welders, welding operations and tackersshall be qualified in accordane with ANSI/AWS 01.1-81"Structura1 We1ding Code-Steel."
C. Field connections:
1. Stee1 contact surfaces shall be clean before bolting.
2. Frame connection operations to the caisson pedestalsshall not commence until after the concrete has reachedits design strength.
4.3 PAINTING
Paint and finish all externa1 surfaces of the steel panel frames. Paintingis not required on concea1ed and inaccesib1e surfaces of the fabrications.
4.3.1 Paint Types: Painting required under this Section is defined asfollows:
(1 ) Prime coat of ch rome oxide primer #15.
(2) Second coat of prime undercoater.
(3) Finish coat of exterior white enamel.
4.3.2 ourabi1it~: Provide paints of durab1e and washable quality. Do notuse pa i nt mater; a s wh i ch will not wi thstand normal washi ng as requi red toremove dust and dirt, pencilmarks, and similar materials without showingdiscoloration, loss of gloss, staining or other damage.
4.3.3 Paint Schedules: Cleaning and painting of steel panel frames is tobe done in the shop and before dummy modules are installed to comp1ete the assembly.
0-4
14614/2Rev. 1
4.4 ERECTION
A. Erect structural steel in accordance with AISC SpecificationStructural Steel for Buildings.
B. Erection in Tolerances: The steel panels shall be erected sothat the devi at i on from specifi ed slope shall not exceed 10
•
5.0 DUMMY PHOTOVOLTAIC MODULES
Furnish, fabricate, mark for erection identification, pack, crate, or otherwiseproperly prepare for shipment, and ship to the site all dummy photovoltaicmodules as indicated on the Drawing No. D-103, described in these Specifications,or otherwise required for proper completion of the Work.
5.1 MATERIALS
A. Aluminum framing members and glass retainer strips:
EASCO ALUMINUM Items #811273 and #811441 (suppliedby Bechtel)
B. Rubber retaining gasket material, EPDM Shore 60 rubber fromPauling Rubber Company, Pauling, New York, (supplied by Bechtel)
C. Tempered glass sheets, 3/16 in.
D. Standard bolts and screws, as specified on the DrawingNo. D-103.
E. Special corner frame fasteners, DECO Products Item#204201 (supplied by Bechtel).
5.2 FABRICATION
A. Fabricate module frames, retaining strips, glass sheetswith rubber edge gaskets as described on the Drawings.
B. Assembled module frames shall have the metal surface finishpreserved, free or mars, dents, or scratches.
D-5
14614/2Rev. 1
C. JOB MOCK-UP
1. First dummy module fabrication shall be assembled for approvalby Bechtel as a standard for acceptance of the fabrication work.
2. One dummy module shall be mounted and affixed to the steel panelframe as shown on the drawi ngs for approval by Bechtel as astandard for acceptance of the installation work.
3. Leave accepted dummy module mock-up in place in the panel aspa rt of the completed work.
5.3 ERECTION
A. Install the dummy modules in the steel support frames accordingto the Drawings.
B. Erection Tolerances: Dummy modules shall be attached to supportframes so that they provide a visibly uniform flat glass surfacein each complete panel assembly.
C. Erection Methods: Sequence of assembling modules into panels,and installing panels onto foundations, shall be discussed withand approved by Bechtel.
D. Field connections: Contact sufaces between dummy modules andsteel support frames shall be clean before bolting.
5.4 CLEANING
Clean dirt and marks from both surfaces of installed modules according tomanufacturers instructions, after arrays are installed complete on thefoundations. Touch-up painting will be required to repair paint damaged byshipping and handling.
5.5 INSPECTION
A. Inspect and correct any deficiencies in dummy modulefabrications installed.
B. Inspect structural steel panel framing prior to mountingfor suitability to receive the modules.
D-6
APPENDIX E
CSU WIND TUNNEL STUDY - PHASE II
WIND-TUNNEL STUDY OFWIND LOADS ON PHOTOVOLTAIC STRUCTURES-
PHASE II
by
M. Poreh,* J. A. Peterka**and J. E. Cermak***
for
Bechtel National, Inc.P.O. Box 3963
San Francisco, California 94119
Fluid Mechanics and Wind Engineering ProgramFluid Dynamics and Diffusion Laboratory
Department of Civil EngineeringColorado State University
Fort Collins, Colorado 80523
CSU Project 5-36060
January 1982
*Visiting Professor (from Department of Civil Engineering,Technion, Haifa, Israel)
**Associate Professor***Professor-in-Charge, Fluid Mechanics and
Wind Engineering Program CER8l-82MP-JAP-JEC2l
E-l
ACKNOWLEDGEMENTS
The authors would like to acknowledge the contributions of
Mr. Morgan Downing to the experimental and data acquisition work. The
help of Mr. Noriaki Hosoya should also be acknowledged.
Further, the authors would like to thank Dr. Andy Franklin and
Mr. Nathan Kotlyar from Bechtel National for their cooperation in
establishing the project and in aiding development of a test plan.
E-2
Chapter
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
LIST OF FIGURES
i
iii
1
LIST OF SYMBOLS
INTRODUCTION
• v
1
2
3
4
EXPERIMENTAL CONFIGURATION, INSTRUMENTATIONAND DATA ACQUISITION • • • • • • • •
ANALYSIS OF THE EXPERIMENTAL RESULTS3.1 The Single Array Tests ••••3.2 The Array Field Tests •••••
3.2.1 The Normal Force Coefficient3.2.2 The Pitching Moment Coefficients3.2.3 The Yawing Moment Coefficients
CONCLUSIONS
REFERENCES
FIGURES
2
666799
10
11
12
APPENDIX A--AERODYNAMIC COEFFICIENTS FORPHOTOVOLTAIC ARRAYS • •
E-3
. . . . . . . 35
Figure
1
2
3
4
5
6
7
B
LIST OF FIGURES
A view of the array field model and fence in theMeteorological Wind Tunnel (configuration NFA4)
Roughness configuration in the MeteorologicalWind Tunnel for generating a 1/7 power law(NEWBYL) near the metric array • . . . • .
Velocity distribution in the test section
Mean velocities and turbulent intensities in thelower section of the boundary layer
A view of the 6-component balance system and the1:24 scale model .
Conceptual low-cost support for aphotovo1taic array . .,. •
Directions of forces and moments for northerlyand southerly winds . . . •
The values of eN and CMZ in the presentand the previous study (1979), H = 2.0 ft
13
14
15
16
17
1B
19
21
9 The values of CN andand H = 1. 5 ft
CMZ for H = 2.0 ft. . .. .... 22
10
11
The values of CMS for HH = 1. 5 ft
Distribution of ICNI x 100(left number, without fence;with fence) ...•
2.0 ft and
in the fieldright number,
23
24
12
13
14
Comparison of values of ICNI x 100 forH = 1.5 ft (left) and H = 2.0 ft (right)without a fence
Comparison of values of ICNI x 100 forfenced fields with H = 1.5 ft (left) andB = 2.0 ft (right) .......•..
Maximum values of ICNI x 100 without afence for all wind directions forH = 1.5 ft (left) and H = 2.0 ft (right)
E-4
26
27
29
Figure
15 Maximum values of ICNj x 100 with a fence(and a corner fence) for all wind directionsfor H = 1.5 ft (left) and H = 2.0 ft (right) 29
16
17
Distribution of ICMZI x 100(left number--without fence,with fence)
Distribution of ICMsl x 100(left number--without fence,with fence) •••••.••
in the fieldright number--
in the fieldright number--
30
32
18 Notation used for array location and fieldtest files . • • . • . • • . . .
E-5
34
LIST OF SYMBOLS
Definition
Array surface area = 192 sq ft in prototype
Array chord length = 8 ft in prototype
Force coefficient in X-direction
Force coefficient in Y-direction
Symbol
A
C
CFX
CFY
CMS Moment coefficient about S-axis
CMX Moment coefficient about X-axis
CMY Moment coefficient about Y-axis
CMZ Moment coefficient about Z-axis
CN Normal force coefficient
DXY = C/2
ES Normalized eccentricity by DXY for MZ
EZ Normalized eccentricity by DXY for MS
FN Normal force
FX Force in X-direction
FY Force in Y-direction
Ground clearance
(yawing moment)
(pitching moment)
Reference dynamic pressure
Center chord of the photovoltaic array
Reference wind velocity
Wind direction
Tilt angle
Density of air
Fence height
Moment about S-Axis
Moment about X-axis
Moment about Y-axis
Moment about Z-axis
H
HF
MS
MX
MY
MZ
QREF
S
UREF
WD
S
p
E-6
1. INTRODUCTION
The present work is a continuation of a previous study on the
magnitude of wind loadings on photovoltaic arrays. (1) A series of
wind-tunnel tests were conducted in that study to determine the effect
of various design parameters on the wind loadings for different wind
directions and wind profiles on a single array and on individual arrays
at different locations in a large array field. The tests showed that
arrays at the upwind edges and corners of the field are subjected to
very large loadings, but that these loadings can be drastically reduced
by fences designed to act as wind barriers. No measurements of the
relative effect of the fence further in the field were made. All the
previous array field tests were conducted with a "standard array
configuration" (see Figure 6) in which the height of the arrays above
ground was H = 2.0 ft.
Subsequent analysis of the wind-tunnel data by Bechtel National,
Incorporated, showed that it might be advantageous to use a smaller
array height, H = 1.5 ft, and that in many situations the use of fences
for reducing the wind loadings on the upwind edges and corners of the
field would not be economical.
It was therefore decided to extend the previous study and to examine
more closely the mean forces and moments on individual arrays in the
field with H = 1.5 ft, with and without a fence. The tests were per-
formed in a 1/7 power-law boundary layer (BLl) at which the highest
loads had been observed. The fence used in the study had 30 percent
porosity and an additional corner fence (see [1], Figure 26).
The results of the present investigation are presented in a ready
to-use form, adopting the previously defined aerodynamic dimensionless
force and moment coefficients and nota.tion.
E-7
2
2. EXPERIMENTAL CONFIGURATION, INSTRUMENTATION AND DATA ACQUISITION
The tests were conducted in the Meteorological Wind Tunnel of the
Fluid Dynamics and Diffusion Laboratory at Colorado State University
in which the earlier study was performed, using the 1:24 scale array
(Figure 1). Some changes have been made in the wind tunnel since 1979
and a different upstream roughness configuration (see Figures 1 and 2)
had to be installed in the tunnel to obtain a 1/7 power law at the test
section. One of the changes was the installment of a larger turntable
which made it possible to rotate most of the array field model as one
unit. This has, however, resulted in a slight dependence of the equiva-
lent roughness in the neighborhood of the model on the wind direction
which could cause a 2 to 3 percent change of the mean local velocities
very close to the floor (1 to 2 in.) which might affect the single
array tests. Figure 3 compares the velocity profiles at the test
section (NEWBYL) with the profile measured in 1979 (OLDBYL). The
dimensionless velocity profiles were normalized by the velocity at
50 in. above the floor. Figure 4 shows the velocity distributions in
the lower portion of the boundary layer. The velocity at H = 10 in.
was used to normalize the profiles in this graph. The old and the new
velocity distributions appear to be very similar but small differences
between them exist. Although these differences can affect the values
of the aerodynamic coefficients by a few percent, they are undoubtedly
small compared to the observed differences in the atmospheric velocity
distributions above apparently similar smooth sites.
The mean force and moment measurements on the instrumented model
were made with a new 6-component balance designed by G. K. Keily and
(2)J. A. Peterka. Figure 5 shows a view of the new balance and the
E-8
mean dynamic pressure QREF
3
metric array connected to it. The vanes which are attached to the bottom
of the balance were submerged in a viscous oil bath to damp vibrations
of the array. The balance was designed to measure relatively small
forces and moments of the order of 1 lb- and 3 lb-in. Its load-voltage
coefficients and its axes of zero moments were determined by a careful
calibration prior to the tests. Frequent checks of its response were
also made during the course of the tests. Since the balance used in the
previous study was a 50-lb balance which worked at the lower part of its
useful range, it is estimated that this could have caused relatively
larger scatter and errors in the moment measurements and that the
present moment measurements are much more reliable.
The use of the new balance made it possible to measure, in
addition to the normal force and pitching moment coefficients CN and
CMZ, the yawing moment coefficient on the structure CMS around the cen-
ter chords of the array S (see Figures 6 and 7). The pitching and yaw-
ing moments were then used to calculate the displacement of the normal
force from the center of the balance, which will be designated by ES
and EZ (see Figure 7). Note that the sign of ES and EZ is deter-
mined by the sign of both the moments and the normal forces. The same
2(= 1/2pUREF ) at the reference height of
30 ft was used in this study so that the dimensionless coefficients
given in the report are:
CN FN= QREF
. A
CMZMZ
= QREF• A • DXY
ES CMZ= -eN
(1)
(2)
(3)
E-9
CMS MS=
QREF - A - DXY
4
(4 )
EZ =CMSCN (5 )
where A is the area of the array (192 sq ft in the prototype) and
DXY = C/2 is half the chord length of the array (C = 8 ft in the
prototype)_ The normal force was calculated using the equation
FN = FX - sin 35° - FY - cos 35°
where FX and FY are the forces in X- and Y-direction as
shown in Figure 7 respectively_
(6 )
It should be stressed that the shape of the array which resembles
to a large extent that of a flat plate ensures the resultant force on
the array is practically equal to the normal force FN. Thus the co-
efficients of FX and FY on the array can be determined by
!CFXI =
ICFYI
ICN I sini>
ICNI cosi>
(7)
(8)
where i> in the present study is 35°.
The moment coefficients acting on the arrays CMX, CMY and CMZ
(see Figure 7) are determined by the position of the resultant force
FN and are given by
and
CMZ
CMS
CN - ES
=CN-EZ= CMYsin 35°
CMX=
cos 35°
(9)
(10)
where MS is the moment around the S axis (see Figure 7).
In the above equations we have used the values of !CFXI and !CFYI
since the sign of FX and FY varies with the wind direction. Note
E-IO
5
that FX is defined as the force in either the south or the north
direction and not in the direction of the wind.
The above calculations are, however, correct only within an error
of approximately 5 percent in the force coefficient calculations and
10 percent in the moment calculations.
E-ll
6
3. ANALYSIS OF THE EXPERIMENTAL RESULTS
3.1 The Single Array Tests
The single array tests were made for comparing the data obtained
in the present and the previous report to examine the effect of the
reduced array height H from H = 2.0 ft to H = 1.5 ft and to measure
the yawing moment coefficient CMS, which was not measured in the pre
vious study. The data for the single array runs is tabulated at the
end of the report in Files S20 and SIS. Figure 8 compares the previous
and the new measurements for H = 2.0 ft. The agreement between the CN
data is satisfactory in view of the slight changes in the velocity
profile. The data confirms the previous observation that the maximum
value of CN is obtained around a wind direction of 45°.
Larger differences exist, however, between the present and the
earlier measurements of the pitching moment coefficient CMZ. As
explained earlier the previous measurements are less reliable than the
present ones.
Figures 9 and 10 compare the aerodynamic coefficients of the single
array for H = 2.0 ft and H = 1.5 ft. The dependence of the coefficients
on the wind direction is very similar. A small reduction in the values
of the coefficients is observed for H = 1.5 ft. As one sees from the
yawing moment coefficient the maximum moment around the S axis occurs
around wind directions of 45° and 135°. Slightly larger moments are
observed for the southerly winds.
3.2 The Array Field Tests
The results of the array field tests are tabulated in the Appendix
at the end of the report using the notation shown in Figure 18. The
data is also presented in Figures 11 through 17 using a schematic
E-12
7
description of the array field. The H = 2.0 ft data are from the
Phase I study.
3.2.1 The Normal Force Coefficient
The measurements of the normal force coefficients showed that all
the coefficients measured at the northeast corner of the field for wind
directions a·, 30·, 45· and 60· are negative, indicating an upward lift.
Similarly, all the coefficients measured at the southwest corner of the
field are positive, except for a few cases in which the absolute value
of CN is very small. We shall therefore refer in the following dis
cussion to the absolute value of the normal force coefficient expressed
in percent ICNI x 100.
Figure 11 shows the values of [CN I x 100 measured in 176 runs.
Each section of these figures shows the values measured at the northeast
corner for one wind direction together with the values measured at the
southwest corner for the opposite wind direction; WD = O· and 180·,
30· and 210·, etc. Two numbers are shown in the space allocated for
each array. The left-hand number shows the value of ICN[ x 100 for
the field without a fence and the right-hand number shows the value of
ICNI x 100 for the fenced field. The 30 percent porosity fence used in
this study was always augmented at the corners by the corner fence used
in the previous study (see [1], Figure 26). Figure 26 also shows the
spacings between the rows and the distances to the fence. The height of
the fence in the prototype was HF = 5 ft.
One sees from Figure 11 that the normal force coefficients recorded
at the upwind edges of the unfenced field were close to those recorded
in the single array tests and that the largest value CN = - 0.79, was
E-13
8
recorded at the corner of the northeast field. Slightly smaller values
were recorded at the east and west edges of the field.
A drastic reduction in the absolute values of the normal force
coefficients is observed in the inner part of the field where a maximum
of CN = -0.27 was recorded at array B2 for WD = 30°. Significantly
lower values are observed for southerly winds.
Figure 12 compares the values of ICNI obtained in the unfenced
field in the present study for H = 1.5 ft with those obtained in the
unfenced field previously with H 2.0 ft. In general the H = 2.0 ft
values are larger. The only significant difference is at array NC2.
(The arrays denoted by 1,2,3,4,5,6,7 and 8 in the previous report are
being denoted now by Al,Cl,A2,C2,A4,C4,A7 and C7.)
The use of a 5-ft high fence, with a corner fence, has drastically
reduced the values of the normal force coefficients on the array facing
the wind. The fence was particularly helpful when the wind was normal
to it. It was not very effective, however, in reducing side winds. The
normal force coefficients on the east side were, in some cases, even
higher than in the unfenced field probably due to the effect of the cor
ner fence. The fence has also increased the values of some of the
coefficients inside the field.
Figure 13 compares the normal coefficients measured in the present
study with those measured in the previous study (H = 2.0 ft). Some of
the values from the previous study were measured without a corner fence.
These are denoted by an asterisk.
The order of magnitude and the distribution of the two sets of
data appear to be quite similar. One sees, for example, that the increase
in the values of the coefficients from array C2 to array C4 and array C7
9
for wind direction 45° exists in both cases and that the relatively
high value of ICNI at C7 was not due to an experimental error.
Figures 14 and 15 show the maximum value of ICN[ obtained at each
array for all wind directions for the unfenced field and for the fenced
field.
3.2.2 The Pitching Moment Coefficients
The pitching moment coefficients in the unfenced field were
generally low, see Figure 16. The maximum value recorded was
[cM21 = 0.09 at the southern edge of the field for wind directions 180°
and 210°. The values of CM2 in the fenced field were practically zero.
3.2.3 The Yawing Moment Coefficients
The distributions of the yawing moment coefficient are shown in
Figure 17. Very small values of [CMS[ were recorded for WD = 0° and
180°. Larger values were recorded at the west edge of the field for
WD = 210° and very large moments were recorded at the same edge for
WD = 225°. The reason for these large moments is that only one side of
these arrays is exposed to the wind. A similar effect is noticed in the
eastern edge for WD = 45°, but apparently the protection provided by
upwind arrays in this case is larger.
The fence appears to reduce to one-half the large moments recorded
on the west side for WD = 225°, as well as the rest of the large
moments in the field. However, in some cases small amplifications of
ICMSI are observed, as in arrays SFB2 and SFB3.
E-15
10
4. CONCLUSIONS
The measurements made in the present study appear to be consistent
with those made in the previous study and a clear picture of the load
ings in the field emerges.
In the unfenced field one can divide the field into two parts; the
edges of the field where the values of ICNI are larger than 0.30 and
reach a maximum near 0.8, and the inner part of the field where
ICNI < 0.30. The yawing moment is also large only in the arrays at the
edges of the field.
The introduction of a fence with a corner fence reduces the high
loadings on the outer arrays to a maximum of ICNI = 0.37 and also
reduces the very large yawing moments in the field. However, the
fence slightly increases the loadings on some inner arrays. Thus, the
loadings on the entire fenced field appear to be approximately of the
same order of magnitude for design purposes.
E-16
11
REFERENCES
1. Poreh, M., J. A. Peterka, and J. E. Cermak, Wind-tunnel study ofwind loads on photovo1taic structures, Colorado State UniversityReport No. CER79-80MP-JAP-JEC12, September 1979, 89 p.
2. Kei1y, G. K., Unpublished report describing a project preparedunder the supervision of Professor J. A. Peterka.
E-17
12
FIGURES
E-18
13
Figure 1. A view of the array field model and fence in theMeteorological Wind Tunnel (configuration NFA4).
E-19
0"Iu
rnH
JOIf
)(Men,
,;H',
uy
/~III
<>
17~':
.••
•••••••••••••••••••••
II•
•,
f-~
lLf-
••
••
•••••••••••••••••••••
••
f-•
••
f-•
••
••
••••••
•••••••
••
••
••
••
••
f-~
B"I.
.f-
f-•
••
••••••••
..•••
•••
••
~'(
...
••
•••••••••••••••
•••
•••
••
f-l-
E•
••
••
••
•••
••
••
••
••
••
••
••
••
••
l-• I
••
••
••••••••••
•••••••••••
••
f-
7
~12
'·1·
36''~4+4'+4'+4,.j.4'+-161-----l
I"C
ubes
,II
%A
rea
Cov
erag
e
t.>J
,{, o•
••
••
•
••
•
••
•••
•••
••
•••
••
1/2
"P
egs,
4.5
Peg
s/f
t
1/2
"P
egs,
9P
eg
s/ft
1/2"
Cub
es,6
%C
over
age
m ""II
..... ~
Fig
ure
2.
Ro
ug
hn
ess
co
nfi
gu
ra
tio
nin
the
Mete
oro
log
ica
lW
ind
Tu
nn
elfo
rg
en
era
tin
ga
1/7
pow
erla
w(N
EWBY
L)n
ear
the
metr
icarr
ay
.
15
LOG-LOG PROFILE PLOTS
190.0
a NEWBYL
o OLDBYLa
aa
til
a
In i.lPE_ a
""a
cv aI- 10.0 III:I:(!)HW:I:
lJ3
lD
n
10.91.9UMEAN/U(50)
I.9L...---I---I.--I.---4lu......L..L.UL...--_-1..---L.--L.-I-.L..L-.L..L.J
.I
Figure 3. Velocity distribution in the test section.
E-2l
ME
AN
VE
LO
CIT
YP
RO
FIL
ES
TU
RB
UL
EN
CE
PR
OF
ILE
S2
02
0
oNE
llBYL
CINE
llBYL
oOL
DBYL
oOL
DBYL
15C
D1
5Cl
O
ZZ
HH
dJ
~,.l
ClO
I 1-\
0C
ClO
-i:'J
II
1C>
C>I
I.....
I\)
HH
'"I\
)
WW
IG
IC
5l:D
5llI
1-
lIII
oCI
[Jl
ClO
-C
00
00
ClO
CIO
00
00
.0.2
.4.6
.81
.ra1
.20
10
20
UM
EA
N-
No
rm
oI
fzed
LO
CA
LT
UR
BIN
T-
X
Fig
ure
4.
Mea
nv
elo
cit
ies
and
turb
ule
nt
inte
nsit
ies
inth
elo
wer
secti
on
of
the
bo
un
dar
yla
yer.
17
Figure 5. A view of the 6-component balance systemand the 1:24 scale model.
E-23
18
see alsofigure 7
/A2=6.75
~
y
~dl-7""'----L-.--__X
xro()
SOUTH•
H
Figure 6. Conceptual low-cost support for a photovoltaic array.
E-24
19
Panel i ," FN<O\
ES·..c..~2 /\
\ ,\
FX<O "
a) Northerly Winds
NortherlyWindDirection
b) Southerly winds
Y
SoutherlyWindDirection
•x
MX (North)
MYPanel i,c "ES·-___.\ FY<O2 FN>O"'.(',,,,,
ELEVATION
Figure 7. Directions of forces and moments for northerlyand southerly winds.
E-25
20
!Northerly Wind
S
c-DMS
EZ·..Q..2
cES·- 1\ Z2C V
MZ
a) Northerly Winds
S
( DMS
'\ Zc 1ES·.k tV
~ 2 MZEZ·~
2
b) Southerly Winds
iSoutherly Wind
® Action Point of the Normal Force FN
PLAN
Figure 7 (continued).
21
I .eCCN (198 \) *eN (1979)
.5 ~ VI • ~ ~.....
8cQ)-
~0 .0- ~4-4-
8Q)0
(.)
-.5 ~
8.
, e 8 a-1.0
0 45 90 135 180\.lind Dired ion - (Alpna)
I .eOCMZ( 198 j) *CMZ(1979)-r
~
.5 I- -.....C
*Q)
*- ~ * C *0 C • • VI iii *.e c-4-4-
Q)0
(.)
-.5 ~ .
18045 90 135\.lind Direc~ion - CAlpna)
DIMENSIONLESS FORCES AND MOMENTS
-I .0 L.... --I- ....I- .L...- -'
o
Single Col lec~or, Be~a = 35
Figure 8. The values of CN and CMZ in the present and theprevious study (1979), H a 2.0 ft.
E-27
22
1.0DCN CH=2.0fD *CN CH=I.SfD
.5 f- i i Vi 1
.....c 8q)-0 .0 8- 8l+-
I+-q) if0
u-.5 -
8• e e e
-1.0 ,
0 45 90 135 180\Jind Direction - CAlpha)
1. 0DCMZCH=2. 0fD *CMZCH=1 .5fD
.5 f- -.....cq).-
W • * 8 ~0 .0 ];I IOl - - *.-l+-I+-
q)0
u-.5 -
18045 90 135\Jind Direction - CAlpha)
DIMENSIONLESS FORCES AND MOMENTS
- I. 0 '-- --1' ..1...-' '-- --'
o
Single Col lector, Beta = 35
Figure 9. The values of CN and CMZ for H = 2.0 ft andH = 1. 5 ft.
E-28
23
l.eOCMS(H=2.0ft) *CMS(H=I .Sft)
.5 ... -.....~c • i• ill- iu .0 ~ • i- aCt- 8Ct-
~• e0u
-.5 ~ -
18045 90 135Yind Direction - (Alpha)
-I." L- ....L...- --L. ---l ---l
o
Figure 10. The values of eMS for H = 2.0 ftand H = 1.5 ft.
E-29
24
c B AWD = 0
171 181172 151174 1411for NE corner
1 9 n1114 91112 71 2
WD = 180 ! I 7 711 3 711 7 71 3for SW corner • •
• I 111 C~=~~~~~ 110 n1 4ICNI x 100 8
(::" fence fence I r---------j ,..--------~,.-------,
". r~ : I : I ,5"" ________..J .... ________ -' L ________ j
II-105ft ,..---------. ~--------.,---------1: :1 I: 161. _______ -' L.. ________ J '- ________J
312 1111 1 110 1 1114141 :--------1121 1917L. ________ .J
216 4118 1 117 21
Ila 611 52 811 51 81A B C
C B AWD = 30
163 1511 1179 11 I1for NE corner 62 11
I12 611 27 7 1138 121 2WD = 210 /'. I 911 9 1128 3013for SW corner 11 16
ICNI I 10 I[~~~~~~~8] 1 27 3714x 100 ]0
Ina fepce fence I ../ ,...--------1 r--------., ,..--------.I " " 15L_______ J L _______ JL _______ J
II - 1.5 ft r--------'r--------'r--------,l II II 16________ J ,________ J L _______ oJ
31 25 2711 4 211 2 311 14 13 I[=~~~~~~J 134 30 17
21 20 1511 3 411 5 21
II 53 811 44 311 46 41A B C
Figure 11. Distribution of ICNI x 100 in the field(left number, without fence; right number,with fence).
E-30
25
c B A
cB
WD = 225for SW corner
ICNI x 100
bo fence fencJ
H • 1.5 ft
WD = 45 I II II Ifor NE corner~5=1===1=1 :::5::2===9 79 14 I
115 71117 101145 271 2
111 91 h4 2611~38=====2613111 uI G:~=~~~;]139 d4
r---------1;--------1r-------15.. J ... ... I. ~
1"'---------1 .. --------.. j--------,___--. L J L JL. l631 31 2611 2 711 1 4 1113 201 [=~===~J 146 B!7
~:::::::::::21 30 27 II 3 2 II 3 1 0 II L6 '61135 311 35 2 1
A
•••
ICNI x 100
H • 1.5 ft
IDo fence fence I
WD • 60 I C II B II A Jfor NE corner 32 8 31 9 54 201 I
I 8 81116 24 1144 14/ 2
.. I 6 161115 28 1139 1413
/' I 6 141 [~i~~==to] 139 n!4..-- ------, r--------., 1'------- ..., " II 15L .J ~ oJ L J
r--------,~--------'r--------~I I' t I 16____ .. OJ L J l. J
31===~11~==::::;11 II 6 141 C=~~:~~J 141 1J 7
21 II II 1II II II I
A B C
Figure 11 (continued).
E-31
No fence
ICNI x 100~____
H =Cl:5 2.0!ft
B A
===1174 7711
~~II 12
~_II 13
26
1· ..C
WD = 0171 7111for NE corner
I 31119
WD = 180 ... i I IIfor SW corner
21 II II 7 71
I 148 6111 1151 591A B C
C B AWD =30 / .. I 61 5911 Ib 9111for NE corner
I 12 3311 II 12WD =210 ..j 1 II II 1
3for SW corner
21 II II 7 41 ICNI x 100
I I 48 6211 1151 511H =[ii 2.0\ft
A B C
C B AWD = 45 GJ 4611 II 79 831 1for NE corner
/- 1,5 zzll II 12N
I II II 13·WD = 225 "/for SW corner
21 II II 3 41 ICNI x 100
I 146 5111 II 35 351H = 11. 5 2.01ft
A B C
Figure 12. Comparison of values of ICNI x 100 forH = 1.5 ft (left) and H = 2.0 ft (right)without a fence.
E-32
27
c
c
B A
WD = 30for NE corner
C B A
JZ II 111110*11 1112
II II
*Denotes a fence without a corner modification
Figure 13. Comparison of values of ICNI x 100 for fenced fieldswith H = 1.5 ft (left) and H = 2.0 ft (right).
E-33
H
KD = 225f or S~\T corner
C:-; x 100
1'-'1::.;...:5'-- .::.2.:..;:.01 f t
fenced
28
C B A
12* II 1114
9 II 1127II 1122
A B c
C B A
II 11209*11 1114
11 IIC:-; x 100
H 1l.5 2,01 ft
fenced
31 II21 IIII II
A B
N
/•
c
Figure 13 (continued),
E-34
29
C B AWD = All
In nil 1179 9111for NE corner 72
115 3311 27 1145 12
WD = All111 II 16 Ib9 13
for SW corner • ••
111 I:~;-----: b9 14ICNI x 100 '- _______ J
ks 1ft r---------·. r--------j r-------..,H 2.0 • : : I II ,5
~________..J L. ________ -II_________ .J
• ,...---------j r--------., .---------1no fence • : II I, 16.... _______ -1 '-________ J '-________J
3hl II 4 II 2 11141:--------1146 17L ________ .J
2130 II S II 7 I
1153 6211 52 II 51 IA B C
Figure 14. Maximum values of ICNI x 100 without a fence for allwind directions for H = 1.5 ft (left) andH = 2.0 ft (right).
1=18==1~911~lS====:::;1120 lsi 1
1::::11====1=;31124 lin 3312
116 Ib Ibo 13
1=14=====17::;1 G]~~~~:~] 1~3=7=====29=14,..--------, r---------.., j--------,l II q 15~. J l- J '- J
~--------,,--------'r--------,I II II 16.... J l ll.. .J
1120 271 [=:~~~~~J boI1
WD = Allfor NE corner
WD = All •for SW corner •
•ICNI x 100
H 11.5 2.01ft •• •fenced
3126 117 II 4
2127 25114 1110
I h6 diS II sA B C
c B A
Figure 15. Maximum values of ICNI x 100 with a fence (and acorner fence) for all wind directions for H = 1.5 ft(left) and H = 2.0 ft (right).Note: Higher values for H = 1.5 ft are usually from
the WD = 30 which is missing in the H = 2.0 fttests.
E-35
30C B A
WD = 014 1114 dh 111for NE corner
!14 0114 dis 01 2
WD = 180 It dlo 0110 01 3for SW corner • •• 10 Ir-------'I d4ICMZI x 100 O:J.. J'O_____ .J
Ina fence fence I r---------, ~--------~~-------,
". t: II :. ,5..________J ~________ -J L ________ ~
r---------· ,.--------, ---------1H = 1. 5 ft I l; I: '6L_______ -, L. ________, '-________1
310 111 2 1110 1110 11 :--------111 11 7~________..J
210 1111 0112 01119 0118 1118 01
A B C
c B AWD = 30 I II II Ifor NE corner 2 a a a 3 a I
13 0111 a114 01 2
WD = 210 I 11 II II a13for SW corner • a a 1 ..l3====;. 10 11 [l~~~===i] I~ 114,..--------, ,..--------, r--------.
I " II 15L .J 1.. .JL ,
r--------,,..--------,,.--------,I It II 16t.. J 1 l .... J
----'1110 3IC==~~~~JI1 21701
H = 1.5 ft
3]r 1110 21102h 1110 all a118 0119 0119
A B C
Figure 16. Distribution of ICMZI x 100 in the field (leftnurnber--without fence, right nurnber--with fence).
E-36
31c B A
01c
WD = 45 I II II Ifor NE corner 2 13 11 1 I
12 Ilh 211z 11210 1111 1112 01310 01 [~====_-.9JIz 014r---------i r--------"1 ,--------jL j L JL J5,..---------. r--------, --------,: II I: 161.. -1 '- ,I '- 1
111~llr--------ll~ 117======..~__....=..i L. -',,;:;.. =-.
11
WD = 225for SW corner
ICMZI x 100Ina fence fence I
H = 1.5 ft
312 11 h 2110213 olb 1110'is 1117 01 h
A B
C B AWD = 60 12 0111 0114 111for NE corner
10 1112 2113 01 2
• 10 01 h 0113 113•
ICMZI x 100 ",- 10 01 rl-------~113 014L..: ________ ~
Ina fence I ,..--------1 i--------, r--------1fence • : II II 15
•~ _______ .J l._______ .J L _______ J
• r--------,r---------'r--------,H = 1.5 ft f I' II 16... _______ ~ I________ J L. _______ .J
31 II II 110 1---------1 017o L_____ :.. __i 3
21 II II III II II I
A B C
Figure 16 (continued).
E-37
32C B A
WD = a14 a 116 0118 111for NE corner
WD = 180
lh a110 0110 112
for SW corner 10 aIII a lb d3• •ICMSI x 100 • h aI:3-------;: 14 21 4"' _______ .1
Ina fence I r---------i ,..--------... ,--------.,fence '. ·1 : :; :. 15... ________..J '- ________ -' L ________ ~
H = 1.5 ft,.---------j ,..--------, .---------1: ,; I, '6... _______ -1 '-________...... ________1
31 8 211 2 111 1 1111 zl r--------11s 217L.. _______ ..J
21 5 111 1 111 1 11II 3 zll 3 aII 5 11
A B C
C B AWD = 30 Is 21110 21117 dlfor NE corner
10 all 1 aII 2 31 2WD = Z10 / .. IQ J II JJ 611 ul3for SW corner 6
jCMSj x 100 16 31 r-------;119 ul 4L.3________
Ina fence I ,...--------1,...--------, r--------.fence ' " " '5L_______ J L _______ .J L _______ J
H = 1.5 ft r--------'r--------'r--------,I II II 16L _______ j ,_ _______ l L _______ J
3(; 2511 6 dl 6 31 h 21 [~~~~~~~J 110 5172135 nil 5 511 2 zl1[;) 51117 01121 21
A B C
Figure 17. Distribution of ICMSI x 100 in the field (leftnumber--without fence, right number--with fence).
33
c B A
zl
WD = 4S I II II Ifor NE corner 1 1 1 3 Zl S I
Iz 311z8 91116 zI2III 9111z sll17 81 3
I 6 61~~==~~~-.!JI17 714r---------·. ,---------.., .-------.,: :: :' 151- -' '- -' L .J
r---------i r--------, .---------1: I, I, '6L.. -I L.. J '- J
----,sll 41:--------1Izo 11 7_ .. 6 _ . L. -' """ .....
41
H=1.Sft
ICMSI x 100Iuo fence fence}
WD = ZZSfor SW corner
3140 2111 6 911 62146 zoll 7 1311 4I/JO 101118 61113
A B C
c
oo
o
B
~r=N~Ocorner lIS C 411 7 B 311 8 A 71 1
119 lOlls 6111s 21 2
00 In 7116 411n 31 3
/ In 51 [~:~~:~~~] 113 J 1 4,...--------, r--------, r--------l, " " '5L J L .J L J
r--------'r--------'r--------,I II II 16:.... J I l L J
===11---1114 0IC~~~~~~~JI=14_~217
~=II I~_II I
ICMSI x 100Ino fence fencel
H = 1.S ft
31 II21 IIII II
A
Figure 17 (continued).
E-39
34
17
C B A
NOC1 II NOB 1 II NOA1
NOC2 II NOB2 II NOA2
NOC3 II NOB3 II NOA3
~===;ll~===12~===;13
,--.c.;N..;c0..;cC4,--_1 C=~~~~-_j '-NOM 14r---------- ~---------,.-------.,: 11 ;, 151. -' .... -' L .)r---------i r--------, .---------1; II If 1 6L.. -I L.. ~ 1-. •
=::~ll'--~-l r,'--------','1_ _NOC7. '- -'. NOA7
I
WD = 0,30,45,60for NE corner
• ••
WD = 180,210,225 ••for SW carner •
31 SOA3 II SOB3 II SOC3
21 SOA2 II SOB2 II SOC2
I I SOA1 II SOB1 II SOC1A B C
Notation for the field with no fence.
AB
II NFB1 "NFA1 ] I==~II NFB2 II==NF=A2==!2
II NFB3 II NFA3 13NF'1\4 -.-J 4
c
•••
WD = 0,30,45,60~for NE corner ~C1
j NFC2
I NFC3
I NFC4
WD = 180,210,225 ••for SW corner •
31 SFA3 II SFB3 II SFC3
2 1 SFA2 II SFB2 II SFC2
II SFA1 II SFB1 II SFC1A B C
Notation for the field with fence.
Figure lB. Notation used for array location and field test files.
E-40
35
APPENDIX A
AERODYNAMIC COEFFICIENTS FORPHOTOVOLTAIC ARRAYS
H = 1.5 ft(see Figure 18 for file notation)
and
S20 Single array at H 2.0 ft
SIS Single array at H = 1.5 ft
E-41
36
H'T!'< fOR FILE : S 20
F.U ~J CllNF bJU,[' CII CtlZ ES ells EZ
292 35 ".Y -'.79 .e;,3 .0-4 01 -.Ct2
Z~3 35 15. ¢ .'. e 1 · ¢4 .05 - . t 2 15
294 7" 30 . (. -'.83 .05 .06 -.24 .29,,: .~
295 3S 45. (I -.8t .10 .12 -.30 .35
29€· 7" €.(, . (. -.eo .<0, .15 - .11 18v'.'
297 3S 75. (: -'.27 .02 . ';9 ·Qe -.23
z~e ~" 9(· . (, •.. 1)8 - · (,l) - (, ., · 10 - 1. 1'1,./ "M' . ,'oJ
":'Ie Q ~" 1 OS. £;: .05 · <>2 . .3~ · 2:~' 3.e?,;. J" r .... '.'
3~) () 7" 1 2(0 . Ii .29 .ot - .21 .31 1. 08•.! .~.
3(' 1 35 135. £;: .H .07 -.It .34 .72
3(,12 35 15¢.¢ .50 · (,t -.12 .25 .51
3C~ 3' 7" If.5.(' .51 .07 - .13 .14 .2e~ '.'
304 35 180.0 .4e .Ot - . 13 · (,2 .04
E-42
37
()I\TA FOR FILE ; SIS
Fun COHF /,J I HC' Cll CI1Z ES CMS EZ9';1 35 i,'< e- o' 75 - (leo - .0 to' .03 - 04
1 ~) to'! 35 15 C' - 76 - . (1\;. - . ,) 0 - (> " C,g.' ..., ~~ 30 :::. o· 78 (, 1 ;) 1 - 1 g 23.:. •.... , "'"1 1 :2 -~ 45 I) o· 81 ('6 .08 - ,- 34.;,' .J . -(
1 :~~ 3 Z5 r; ,> :~! ,. 58 e'9 1 5 - . ;) 9 1£,
g-4 35 75 e- o' 30 (>2 .06 .05 - 17
1(05 35 9 '.:' ~, 0'. ')6 - 00 - .01 .or> - 1 ~' ?
1~t E:. 35 1 "5 C' f.' i' e'2 - .25 18 - i' 1~.
107 35 120 C' 3') 07 - -, .30 9~.... -g.g 35 135 (0 41 .08 -.2() .27 65
109 35 1Sti 6 4S ('8 - 18 19 43
1 i " 35 16S ~l 4'l (-9 - 19 1)9 19
1 i 1 35 '1 £0 () 413 e'9 - 19 .02 1)3
E-43
38
lil'<TI'< fOR FILE I NOl'<1F: II ~, (,0 ~,F ltlltU;' (.N CI'!Z ES (.I\S EZ
1~ 8 :;S r, . (I -'. H - . c;.3 -.(,04 · t) e -. 1 c;.
14 " 35 31) . (, -'.79 - . (13 -.04 - . 17 .21
15 (, 35 45.c;. -'.79 · ('1 «1 - .21 .2€,
1S 1 7~ H.c;. -'. S4 · ('4 . « I) -«I) .15oJ·..•
t'RTR FOR FILE , IIFRl
RUN COHF YINI> CII CIIZ ES CMS £Z
147 35 Cl • (I -'.14 - . ('1 - .Of. · 0 1 -.04
IH. 35 3~( . (, -'. 11 - . YO -.04 · I) 1 - . 07
145 35 4S. (. -'.14 · ('1 .05 - .05 .3G
144 35 b~( . (. -'.21) · (, 1 .1)3 .07 -.34
1)1'< TA FOR FILE : tl OR 2
RUN (,0 NF ~I! til) CN CI'IZ ES CMS EZ
14 (. 35 C;. . (. -'. 12 - . c;.5 -.37 .(1) -.c)3
141 35 3c;' (, -'.38 - . ('4 -.OS' .02 -. ¢4
142 7~ 4S . (,1 -.44 - . Y2 -.04 -.IE; 7-v·oJ .~(
143 35 t·¢ . (, -'.44 · (,3 . <H. - . 15 .3~
NiTA FOR F!LE , II FA 2
RUN CO tlF WINe, CII CI'IZ £5 CKS £Z
138 35 O. (. _.. (.\ 7 - . ('0 -.04 - . 0 1 .15
i 3 7 35 3() . (. _.. 12 - . (Ic) - . (q) -.ti3 .27
13 b 35 45. (, ·'.27 · (,1 . tJ 2 .1)2 -.1)9
i 3 5 35 ~o . (I - . 14 - . ('0 -.()O .1)2 -. 13
£-44
39
!)ATA FOR FILE tlOf\3
j;'U N CONF blINf) Cli CMZ ES eMS EZ
1 7. (. 35 {) e, .' l\ :' .00 · 'J 1 - .0:2 31
1:2 ':J 35 30 . C' .'.2S - 03 - · 1 1 •. · ~) f., 21
12g 35 45 e· .. 3S - (12 - · v5 - 17 45
1. ~:- -:>=; f.e, ~I 39 03 .07 - 13 33...., .~
C'A TA FOR FILE : NFA3
RU ti COiF WIN C, eN OIZ ES ells EZt ~ 1 ,,, 0 C' •. CoS - c, I) - .04 - .01 17L ...... <.! '-'
132 35 3¢ "•. 30 OCo · 0 1 - 1 1 37
1 ~ '7' 7" 45 (0 •. 2t; - 1)0 - .02 · (J 8 .' 29L ......... ~~
134- ,,, H) (, 14 - (·1 -.05 · V·3 - o! .,::."-' '-'
~IA TA r 0 j;: rILE : NOli?
FUH COr·!F wnw Cll CI'IZ ES eMs EZ
234 35 0 C' o. 21 C'1 .05 - .05 23
233 35 30 0 •. 34 - C'1 - .04 •. 1 t\ 30
232 Z5 45 e- .. 4f, - C'1 - 0" - .20 43· ~
:?3 i 70= f, t) C· 41 (·3 · () 13 - 14 35~"
r'H TA FOR FlLE : ifF'" 7
,'U t; C(] f\f III fi I) CN. CI'lZ ES eM S EZ
~~3 5 35 (; C· .. 19 - (, 1 - .07 - · ·,2 1·3
=:::. 'E. 35 3(' 0 - 30: - 02 - · (,6 .Co5 .' 18
237 .:; 5 45 C' . 14 ('1 .05 - .01 07
23 (I "e: IE, e, ~, •. 14 (.1) · 0 1 - .02 14.., '-'
E-45
40
['HTA FOR FILE ; HOSI
FUN CONF ~II N[, eN Cl'IZ ES CKS EZ
1B £. 35 () · c' "-.72 - · C'4 -.1'>5 .06 - . f"\ 8
1e5 35 3Q. C, ..·.62 - · ('0 - . 0 C, - .10 1£
1f) 4 35 45 . ~. ··.52 - . C'3 -.05 .<)1 -.Q2
..... ~ 35 bel. (I ... 31 - .01 -.Q4 .07 -.241 c' .:.'
I)A TA FOR FILE , life 1
"u i'i eOHF WItW eH el'lz ES ells EZ
17'1 35 0 .0 _.. 15 .Ot) .01 -.(;(1 .03
180 ~" 3(~ . C, •. 11 - .00 -.04 -.02 1~~~
lU 35 4S. ¢ •.• (lOJ - . C'I -1(1 -.03 .28
182 35 6(1.0 •. 0" - . (1) -.1)3 · 03 -.30
t'kTA FOR FILE , NOB2
~:UN COHf ioJHlt- CN el'lz ES ells E2
l?e 35 0 · C, - 14 - . (04 - .31 .0(0 -.01
1 "'!' .., 35 30 · (I -'.27 - · (, 1 -,03 · 0 1 -.02
i7~ 35 45. C' _.. 17 - . (, 1 -.04 .28 -1.b€.
175 35 bl'( · C, - . 1b - . (12 - .12 .05 -.28
t-ATA FOP. FILE , HF82
it: lJ ~i COHF WI tiC, eN ellz ES ells EZ
171 35 o . c- •. (I" .01 . I) " - • (J (I . t> 5
172 35 30.0 •.. 07 - .0(1 -.03 · 0 (I -.06
1-- 35 45. (, - . 1(I - . (02 - .17 .09 -.89( ~
174- 35 60. (I -.24 .02 .07 .o Ii, -.25
E-46
~Rl'A f'OR FILE: NOB3
41
RU~i CONF
Ib9 35
Lb 8 35I.'.., -,=1 ~ t ~....}
b.i IN Cr
45 (,
CN
- 03
- 16
•. 14
•. 15
CMZ
(OV
- (0 i
ES
.v5-.v(o
- .•HI
CMS
-.Vl1 1
12,H
.v""
EZ
3£.
- 69
- 135
- 40
~ATA fOR fILE : "fB3
164
1·' t::., t- '.'
I Ii, 1;.
GONf7<,~ ".'
7<,~ '.'
7<oJ '.'
1,)1 "I)
C; ("
45. (,
... C; (0
- O~
• 26
•. 28
CMZ
(d;
- (, 1
• (, 1
ES
.02
-.c;9
-.04
.01
eMS
· ,) Ii,
.tiS
EZ
- I) 1
- 7<;0
•. 1e- 14
DATA fOR FILE HOB4
RU~! CONF CN CMZ ES eMS EZ
ib2 35
15 ill :3 5
45 (,
•.. oS
.. 23
•. 1b
- 16
• e'i
- (0(.
- eo 1
.38
-.03
-.V3
-.07
-.03
.03
14
.05
59
- 12
- e7
- 33
DATA fOR FILE' MFB4
RUN GONf WIND GN
155
156
158
35
7<..' '.'
(; • (>
10 (r
45 (0
-, 15
.. 1e33
CliZ
- (01
• . (> 1
- (> c;
• (, 1
ES
- . ,) 4
-.04
- . ()3
CNS
- , ();2
· () 5
· CIl
,02
EZ
1 1
_. 27
- 1 <;,
42
DATA FOR FILE 1 NOC 1
RUN CONF WINI> CN CI'IZ ES CMS EZ
227 35 0.0 -.71 -.04 -.05 .04 -.1)5
22' 35 30.0 -.63 - .02 -.03 -.05 .I)~
225 35 45.0 -.51 - .02 -,03 .01 -,1)2
224 35 60.1) -.32 - .02 -.08 .15 -.46
t>ATA FOR FILE : life 1
RUN CONF lolli'll> CN CPlZ ES CPlS EZ
220 35 0.0 -.1'J - .01 -.Of, -.00 .01
221 35 3~. (I -.15 - .00 -.01 -.02 . 11
222 35 45.0 - . 11 -.01 - .10 .01 -.06
223 35 60.0 -.08 .00 .03 .04 -.4f-
C'ATA FOR FILE : NOC 2
RUN CaliF LI I II D CN CtlZ ES eMS EZ
211 35 0.0 -.09 -.04 -.43 .01 -.07
2i2 35 30. ¢ ". 12 -. (.] -.23 -,00 .04
213 35 45. (. -.15 - . ('2 - .13 -.02 .13
214 35 ';0 .0 -.08 . ('0 0';' .19 -2.a
l>ATA FOR FILE , Ii Fe 2
RUN CONF WINl> CN CtlZ ES CPl S EZ
217- 35 0.0 -.11 - .00 -.Ci4 .00 -.01
217 3S 3(; . 0 -.Of, - . or, -.02 -00 .00
216 35 45.0 -.07 -.01 -.Ci,} .03 -.47
215 7" f,0.0 -.1)8 - .01 -.08 .10 -1.27~'J
E-48
43
DA TA 1'0 I" !'lLE : Noe 3
RUN CONF wnw CN C112 ES el1s E2
21 t) 35 (;< • (I ·'.07 - . (, 1 -.07 - .0(' .05
2tl ';t 3S 3(( . (I •.. 11 - . (, 1 - . 12 -.00 .01
~;(d3 35 45. (, -. 11 · (, 1 .os .11 -1. 0('
;2('17 25 6;) . (, -.0' .00 .01 .13 -1. %
!)t'!TA FOR FILE : NFO
!"tHI eOHF WINC' eN CI1Z ES CI1S EZ
2l) 3 35 o . (- _.. 07 .(01 .09 -.00 .05
2(;,4 '" 30 . (> -.09 · (-0 .03 · (11 -.14~v
205 35 45 . (> -.09 - . (>1 - . if; · 09 -.9&
2t.) f· 35 H. (> -. If, - . (>0 -.00 .07 -.041
C'ATA FOR FILE : NOC4
RUN eONF I,)IND CN CI1Z ES CI1S EZ
195 35 (} . (I -.oe .00 .01 - .01 .Oil
196 35 3(;i . <> _.. 10 - .00 -.02 .t>li. -.1\('
197 35 45. (, - . 11 .00 .01 .06 - ." ~I
19£: 35 t,« . (I -.0£ - . ('0 -.04 · 13 - 2.21
!)A TA FOR FILE , tiFe 4
FU fi CO fir WIN!) eN eftZ ES ells EZ
2C ;2 35 t; . (I - . 11 · 00 .03 -.00 .02
2(' 1 3S 30.(0 - . 1(0 .(01 .10 .03 -,35
200 3S 45 . (> -.12 · (>(> .02 · (, '" -.52
199 35 f·t\ . (0 -'. 17 - . (>0 -.00 .1)5 -.31
£-49
44
riA TA FOR FILE NOC?
RUN CONI' loll Nt' CN CMZ ES C"S EZ
194 35 <,\.1) -.14 - .1)1) -.<.\2 - . I) 1 .<.\6
347 35 3<.\ . I) -.14 .1)0 .00 -.03 -.22
1 (\ ? 35 45.-:- •.. 13 - .00 -.<.\2 .<.\6 -.5('
18 .~ 35 f. ,,' . (I -.06 .00 .02 .14 - 2.13
Col'! TA 1'0 R fILE : HFt?
RU H CONI' WIN" CN C"Z ES C"S EZ
34 g ,~ 0 .0 - 14 - . (>0 .02 -.02 .02<oJ .~.
349 ,~ 30 . (. -. 17 - .(>0 .02 . ':.4 -.25'" .~.
1 c , ,~ 45. (, ·'.20 .01 .04 .04 -.22.' , vv
is 7 35 ~o . (, -.14 . ('0 .03 .0(' -.03
45
!)A TA FOR FILE SOAI
RUN co NF Id ItIC' CN CMZ ES CMS EZ
252 35 18t).t\ .48 · (19 - .18 .1>3 , (.\ 7
25i 35 21<).1> .53 · ('8 - .16 -.23 -.43
25 ,. 35 225. (. .46 · (IS - .18 - .3(1 -'.66
c'ln f\ FC R FILE , srI< 1
F: IJ N CONF IJI til) CN CMZ ES CMS E2
253 35 leo.O . (;6 - .(>0 <;1 · (,:2 . 31
254 35 210. (, .oe - · (I c) · () 1 -.05 _. 57--::J:;I:;; ,~ ., ? c: (, 16 - · (01 .¢6 - . 1') -'.62.:. .... ~ ~~
__ -.t •
C'ATA FOR FILE , SOA2
>:lI N COtlF wItIC' CN CMZ ES CKS EZ
247 35 18') . (. _.. (,'7 - · "() - · ,) 5 -.05 .83
248 35 210,(' .21> · (, 1 -.06 - .35 -1.76
24') 35 225. (I .30 (I ., - .09 - .46 -1.52· ~
e, f\ TA fOR FILE , SFA2
RUN CO ~~F ~J It! Ii eN CI'IZ ES eMS EZ
24€· 35 1S¢ . (, -'. ¢4 - · (>1 -.21 - .01 .4¢
245 ,~ 21 ¢ . (, 15 - .01 · I) 7 - .17 - 1 . 12'-' '-'
24.4 35 225. (,: .27 - · (,,;, .1)1 _..2':' -.7:;
C-ATIl FOR FILE : S0113
RUN CON!' WINI) CN CI'IZ ES CMS EZ
24(:\ 35 1 21') . (I ~t" -· (10 · 15 - · t'q] -3.27
239 35 .2: 1 t') . t\ .25 · (, 1 - · r) 2: - .31 - 1 . :2 3
23f: 35 :2 25 . Co" .31 · (1;2 - . t) 5 - · 4 ~\ -1. :2 9
E-51
46
C'E TA fOP F 1LE SFA3
Pl! ~, CONF II,' TNr, GN CHZ ES CMS EZ
24 1 35 1 80 (> (, 1 - (, 1 - 65 () ;2 9 99
24 2 ,~ 2 1 c' (;1~, - (01 (; 5 - 25 - 93"-' .~' ~(
,:4 <. 35 2: 25 (;. 26 -" 1
() 2 - 2 1 " 7"[; ~. A FOP f I L E SOb i
RUN CO fiF biT NC- CM OIZ ES CMS £2
:~ 7 i 3"5 < E: ~) :) 52 (18 - 1 5 - <)3 _. () 6. 0 :2 5 1') I) H (19 - :2 e· - i ? 3 9~ c· "
27 .3 ::: 5 225 (, 35 ('7 - 2 C' - 1 8 - 5 ~>
H' TA FOR F ILE , SFe 1
PLiN GO ~iF ~; I ~I r, ell el'lz ES eMS EZ
27(,: ::: ~ 1 80 " Ol> 01 - (; e 0(, 0 1.~
2~· 9 35 <: 1,) (, (;J 00 - (;5 - <)0 - 08
2€· 8 ", - "« (, -' 08 00 C; 1 - Of, 68~ .~. " t!..~
!:.~ I H FO h' F i LE $092
FUr·j CONF 1IJ, N ~I eN 012 ES eMS EZ,
2:; 4 35 i BO ,;, ~;. :2 - (, 1 - li,2 0 1 - 3b
2f. ::: 35 2 1:,) (, ;)3 (.:.) - 1 0 - 05 -1 9~1
2:;2 35 225 (, {,\;2 <>2 - 'H - t)7 - 2 67
r'ATA FOR F i LE , SF62
pur; GO NF ~.!I Nr, ell eMZ ES CKS EZ
~:Ii. 5 .~ 5 1 8') (, (,1 - 0(,· - 36 <) 1 9 99
~.: l~_ i. ~ 5 2 1(;. (, (;t4 YO (; (;. 1,5 1 3(1
2i 7 ~ 5 225 ,. 02 (, 1 Ii 1 1 .:' 9 99
E-52
47
C'R TA FOR FILE , S083
RUN CO tHo WINC, Cll CHZ ES CMS EZ
259 35 1et,) (. •. t'l - (,,2 -2 · () tl .02 9. 99
26 t;, 35 :210 C, .os (I() - I) ., - .~, 6 - 7£
261 35 225 C· ~l1 C'1 - .98 · () 6 9 9'~
C'" TI'< fOR fiLE SFB3
RUN CONF IHINIl ell CMZ ES Cf1S EZ
258 35 I B<). C, (01 - C'l - · c; 5 · 0 1 9. 99
257 7~ 210 ~, 01 - 02 1 .3' - 12 ')I. 99" '-'
25€· 7~ --~ ~, <).( - (;2 .28 - · 09 - 1 25........ t!t!..r
I:l~ "fA FOR -I LE , SOC 1r
RUN CONF IoJ I II C' Cll CHZ ES eMS EZ
288 3S 1eo t;, 51 OS - 15 - 05 - le'
287 35 a1 t'< C' 46 C'9 - 19 - .21 - 45
;;::8 f, 35 225 C· 35 (.( - .2'1 - 13 - 38
r,nA FO R FrL E : S Fe 1
FU t1 COH I. Itdi CH CI'lZ ES CliS EZ
289 35 18() ~. ell1 ¢Ii - 01 · 0 1 1 <).
29· (~ 35 21¢ ¢ ,;.4 Col - .22 - Ii? - 6'>· ~
2:'31 :3'5 225 r" ()2 - (Ie) - 12 - ,02 .,( "
!)ATA FO~: FILE , SOC 2
EU t4 CONF bJ IIH' CII CMZ ES CKS EZ
.:.~ 7 'e i Bt.) <:' •. . ~\ 7 - . 1~12. - .25 .01 - 14... '-1 _' .j .J
2&4 35 :2 1 (. C' •. c)5 - ('0 - .02 - t) 2 3<:'
285 35 225 C, •. 03 (10 · () 8 - · ~'4 1 52
E-53
48
I)I<TI< fOR fILE , SfC2
FU t! COHf ~n!Co CII CI'lZ ES ells EZ
2&2 35 18(' . (. - . .,2 - . .,., - . .,8 · ., 1 -.304
2&1 3S 210.¢ _.. 02 -.(0" - . .,4 - · ., 2 .77
26 \\ ::15 --.. ¢ .. 10 · (,1 .08 - . <) 4 .H'::='--.r.
CIATA FOR FILE : SOC 3
F:U N CO t~F '" INC' CII CIlZ ES CIIS EZ.-, "'f .- 35 1 8t~ . (, .IH\ - . C'l) 1. li5 -.0\\ 9.99~r t:.
??5 -~ 21 C< . e, ·'.02 - · <10 - . (.'t6 -.06 2.80.J ....:
274 35 225. e, . C'l · Coo - .15 -.06 ~. 'HI
CoATA fO.; f I LE , SFC3
FllN COHf ~ I HC' Gil GI'IZ ES C"6 EZ
277 '" 18<; .(0 . (,1 - ('1 ·'1 .41;; · Q I 9.9~.~ .~.
278 ~ .. 21 () . (, _.. OJ - · (01 - .17 -.03 .IH·~ ".'
27'?- 35 22S . (~ .'. Q4 - .¢1 - .22 - . 05 1 . 19
E-54