Demonstration of Clyde BergemannWater Cannons at Miller Unit 1

Charles BoohakerMike CarlisleBrian Mead John SorgeSouthern Company

EPRI Heat Rate Improvement ConferenceJanuary 28 – 30, 2003 Birmingham, Alabama

Background

z Alabama Power Miller Unit 1y 660 MW, Commercial operation October 1978y B&W opposed wall-fired furnace

x B&W DRB low NOx burners (56), B&W MPS 89 mills (7)x Compartmentalized windboxx Tube material: SA-210-A1 (medium carbon steel)

z Started using PRB coal in mid-1990s (Unit 1 - 1999)y Change in furnace slagging characteristics

x Installed water lances for water wall cleaning (Diamond Power)• Furnace walls - 20 IK-4M-WL and IK-4M-PA• Upper rear wall - 2 IK-545 Selective Pattern• Furnace division walls - 6 Selective Pattern (3 per side)

x Continuous water wall cleaning (~ three hour cycle time)x Decreased water wall life due to quench crackingx Increased condensate makeup water requirements

Quench Cracking

z Quench cracking dependent ony Velocity of water stream used to clean surface (pressure)y Progression velocity of water streamy Cross-sectional area of water stream (nozzle diameter)y Slagging condition at time of cleaning (clean / dirty)y Tube material (carbon-steel vs. low-chromium alloy)y Cleaning cycles

x Can have multiple temperature cycles per cleaning cycle

Refs:Water Blowing of Fireside Deposits in Coal-Fired Utility Boilers (EPRI CS-4914)

Quench Cracking

Circumferentialcrack leading to steam leak

crazing patternon the surface

tube at nose arch2.75" OD x 0.290" MWT, SA210 A-1cleaned by water lance

cracks inprogress

Primary Goals

z Extend life of water wall sections by mitigating quench cracking

z Reduce plant condensate makeup requirements by using filtered water instead of condensate for cleaning

z Possible other benefitsy Improved boiler efficiencyy Reduced NOx Emissions

Program Schedule

z Installation - Nov. 2001 - Feb. 2002z CBI conducted testing - Feb. 2002z Program Testing - Mar. 2002 - Feb. 2003z Reporting - Mar. 2003

Installation Summary

z Late October 2001 project approvalz Installation during previously scheduled outage

y Mechanical ~ 3 weeks / 6 days week / 10 hour daysx 4 water cannon bent tube openings (~18’ x 50’, 8 tubes) x 24 heat flux tube sections

y Unit back on-line late December 2001 y Electrical ~ 4 weeks; most with unit online

z Operating in program (manual) mode week of Feb. 8z Operating in auto (feedback) mode week of Feb. 18z Continue optimization and debugging of systemz Installation of auto-tuning modules (3 Qtr 2002)

Cleaning Zones

OutsideFront Wall

OutsideRear Wall

Outside LeftSide Wall

Outside RightSide Wall

Water CannonExisting Sootblowers

Heat Flux Sensor

Operations / Reliability to Date

z Running mostly in auto mode since February 18z Very little operator interaction requiredz Some sections being cleaned too often è adjustment of

targeted heat flux levelsz Failure of supply line to cannon(s)z Failure of data acquisition panel (1 of 2)z Adjustment of jet progression velocity and pressure to

limit tube temperature rise

Testing and Assessing Benefits

z Testing mostly passive – unit runs as normalz Track water consumption of lances / cannons

y Frequency / duration of cleaning cyclesy Peak use / annual use

z Cracking of waterwall tubesy Visible inspection of waterwall tubes in area of previous damagey Tube samples in area of previous damagey Use of CBI fast data & faster sampling of waterwall

temperaturesy Use of tube failure models

Testing and Assessing Benefits

z Furnace Heat Absorption and Efficiencyy Gas temperatures (furnace / economizer exit )y Spray flowsy Boiler steaming capacityy Steam temperatures (superheat / reheat)y Boiler efficiency

z Generation Capacity / Unit Availabilityz Emissions

y Flue gas characteristicsx NOx, temperature

Clyde BergemannPerformance Testing

z Conducted Feb. 18-21, 2002z Elements

y Temperature and gaseous composition (NOX, CO, O2) at nose of furnace

y Data logging through DCS/PI and CBI systemy Furnace wall emissivity

z Unit full load during entire periodz Several modes of operation tested

Clyde BergemannPerformance Testing

z Findingsy Improvement in heat transfer distribution (more consistent)y Slight reduction in FEGT è greater overall energy transfer in

waterwallsy Reduction in NOx emissions (10%) as measured at nose of

furnacey No observable change in boiler efficiency

Testing Nov. 22, 2002

z Goaly Verify CBI collected temperatures y Faster sampling (~15 ms vs. 200 ms)y Higher precision

z Used spare thermocouples on CBI heat flux sensorsz Sampling system was Agilent 34970Az Four runs to date

Run 1Comparison of Surface Temperatures

400 450 500 550 600 650 700700

720

740

760

780

800

820

840

860

880

900

Tem

pera

ture

(D

egF)

Time (Seconds)

Run 1 / Nov. 22, 2002

CB Fast DataDatalogger

400 450 500 550 600 650 700650

700

750

800

850

900

Tem

pera

ture

(D

egF)

Time (Seconds)

Run 1 / Nov 22., 2002

CB Daily FileCB Fast DataDatalogger

Zone 415Channel A09 / 9Water Cannon # 4West furnace wall

Data logger trend shifted in time and biased

Run 1Surface vs. Subsurface Temperature

Zone 415Channel A09 / 9Water Cannon # 4West furnace wall

0 20 40 60 80 100 120660

680

700

720

740

760

780

800

820

840Run 1 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)

TC 1TC 2

40 42 44 46 48 50 52 54 56 58 60660

680

700

720

740

760

780

800

820

Run 1 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)

TC 1TC 2

TC1 – SurfaceTC2 - Subsurface

TC1 - SurfaceTC2 - Subsurface

Run 2Surface vs. Subsurface Temperature

Zone 410Channel A10 / 10Water Lance IK WL 09West furnace wall

0 10 20 30 40 50 60 70 80 90 100660

680

700

720

740

760

780Run 2 / Nov. 22, 2002

Time (seconds)

Tem

pera

ture

(D

egF)

2 4 6 8 10 12 14 16 18 20

670

680

690

700

710

720

730

740

750

760

770

Run 2 / Nov. 22, 2002

Time (seconds)

Tem

pera

ture

(D

egF)

TC 1TC 2

TC 1TC 2

TC1 - SurfaceTC2 - Subsurface

Run 3Surface vs. Subsurface Temperature

Zone 308Channel B09 / 21Water Lance IK WL 20East furnace wall

0 20 40 60 80 100 120 140700

710

720

730

740

750

760

770

780

790Run 3 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)

TC 1TC 2

30 35 40 45 50 55 60 65700

710

720

730

740

750

760

770

780

790

800Run 3 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)

TC 1TC 2

TC1 - SurfaceTC2 - Subsurface

0 20 40 60 80 100 120700

710

720

730

740

750

760

770Run 4 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)Run 4Surface vs. Subsurface Temperature

50 55 60 65 70 75700

710

720

730

740

750

760Run 4 / Nov. 22, 2002

Time (Seconds)

Tem

pera

ture

(D

eg F

)

Zone 303Channel B10 / 22Water Cannon #3East furnace wall

TC1 - SurfaceTC2 - Subsurface

TC1TC2

TC1TC2

Long-Term Data

z Data sourcesy Plant DCS / PI system

x Normal plant data such as excess O2, load, gas temperature

y CBI daily log filesx Few plant parameters plus water cannon flow information such as

flow, pressure, cannon, zone, heat flux, surface temperature, etc.x 30 second sample

y CBI fast data filesx Waterwall surface temperatures (only when cleaning that zone)x Nominally 200 millisecond sample

Typical Surface Temperature Profile Over 24 Hours

HEATFLUX_1_SURFTEMP

770

780

790

800

810

820

830

840

11/30 04:48 11/30 09:36 11/30 14:24 11/30 19:12 12/1 00:00 12/1 04:48 12/1 09:36

Source: Daily log file for December 1, 2002

Example of Surface Temperatures During Cleaning

Sensor01

640

660

680

700

720

740

760

780

800

820

840

1 38 75 112 149 186 223 260 297 334 371 408 445 482 519 556 593 630

Tem

pera

ture

(D

egF)

Source: fast data file for December 1, 2002

Distribution of Temperature Droop and Gain(April 2002 - November 2002)

Temperature Droop

Temperature (Deg F)

Fre

quen

cy

0 100 200 300 400 500

010

0020

0030

0040

00

Temperature Gain

Temperature (Deg F)

Fre

quen

cy

0 20 40 60 80 100

020

040

060

080

010

0012

00

Droop

Gain

Source: fast data files April 2002 – November 2002

Distribution of Cleaning Cycle EventsTemperature Droop > 0°F, < 50°F(April 2002 - November 2002)

20020405 20020525 20020707 20020815 20020920 20021027

Cleaning Events with Droop > 0 DegF

FastFile

Cou

nt

050

100

150

20020405 20020525 20020707 20020815 20020920 20021027

Cleaning Events with Droop < 50 DegF

FastFile

Cou

nt

050

100

150

Source: fast data files April 2002 – November 2002

Distribution of Cleaning Cycle EventsTemperature Droop > 100°F, > 200°F(April 2002 - November 2002)

20020408 20020508 20020607 20020801 20020825 20020914

Cleaning Events with Droop > 200 DegF

FastFile

Cou

nt

010

2030

4050

6070

20020405 20020525 20020707 20020816 20020921 20021028

Cleaning Events with Droop > 100 DegF

FastFile

Cou

nt

010

2030

4050

6070

Source: fast data files April 2002 – November 2002

Boiler Tube Life

z Failure = Cracking = f (TInitial, TDroop, material, frequency)z Statistical approach

y Identify temperature transients - worst, best, typicaly Determine historical (12 months) use of water cannonsy Compare to water lance induced transients

z Identify tube failure model(s)y Direct calculation (non-FEM)y Determine sootblowing impacts on quench cracking

z Run temperature data through models

Summary

z Installation y No major problems / should have done more with unit online

z Operations / Reliability to Datey Relatively few problemsy Plant staff have positive impression and are installing on other unitsy Little training to date, more required for operations, I&C

z Performancey Balanced heat flux in furnacey Still open questions on NOx and efficiency y Many cleaning events have temperature drop greater than 100°Fy Some zones cleaned considerably more than other zones