the design of psb-vver experiments relevant to accident ......journal of power and energy systems...

Journal of Power andEnergy

Systems

Vol. 2, No. 1, 2008

371

The Design of PSB-VVER Experiments Relevant to Accident Management

Alessandro DEL NEVO**, Francesco D’AURIA**, Marino MAZZINI**, Michael BYKOV***, Ilya V. ELKIN**** and Alexander SUSLOV*****

** UNIVERSITY OF PISA, Via Diotisalvi 2, 56126 Pisa, Italy

E-mail: [email protected] *** EDO GIDROPRESS

Podolsk, Moscow Region, Russian Federation **** EREC

Electrogorsk, Moscow Region, Russian Federation ***** RRC Kurchatov Institute Moscow, Russian Federation

Abstract Experimental programs carried-out in integral test facilities are relevant for validating the best estimate thermal-hydraulic codes (1), which are used for accident analyses, design of accident management procedures, licensing of nuclear power plants, etc. The validation process, in fact, is based on well designed experiments. It consists in the comparison of the measured and calculated parameters and the determination whether a computer code has an adequate capability in predicting the major phenomena expected to occur in the course of transient and/or accidents. University of Pisa was responsible of the numerical design of the 12 experiments executed in PSB-VVER facility (2), operated at Electrogorsk Research and Engineering Center (Russia), in the framework of the TACIS 2.03/97 Contract 3.03.03 Part A, EC financed (3). The paper describes the methodology adopted at University of Pisa, starting form the scenarios foreseen in the final test matrix until the execution of the experiments. This process considers three key topics: a) the scaling issue and the simulation, with unavoidable distortions, of the expected performance of the reference nuclear power plants; b) the code assessment process involving the identification of phenomena challenging the code models; c) the features of the concerned integral test facility (scaling limitations, control logics, data acquisition system, instrumentation, etc.). The activities performed in this respect are discussed, and emphasis is also given to the relevance of the thermal losses to the environment. This issue affects particularly the small scaled facilities and has relevance on the scaling approach related to the power and volume of the facility.

Key words: PSB-VVER, Integral Test Facility, VVER-1000

1. Introduction

The Project R2.03 of TACIS-97 (3) (started in January 2004 and ended in July 2006) program “Software Development for Accident Analysis of VVER and RBMK Reactors” consisted of two independent parts: • Part A: development of Accident Management (AM) procedures (4) on the test facility *Received 7 Sep., 2007 (No. 07-0521)

[DOI: 10.1299/jpes.2.371]

Journal of Power and Energy Systems

Vol. 2, No. 1, 2008

372

“PSB-VVER” at Electrogorsk, see Fig. 1; • Part B: development of a code system for severe accident analysis in RBMK reactors.

In both parts of the Project, University of Pisa (UNIPI) was the western contractor with EC. Part A of the project was focused on the development and optimization of accident management procedures for VVER1000/320 (Russian pressurized water reactor) Nuclear Power Plant (NPP). The objective was fulfilled using of Thermal-Hydraulic (TH) system codes and through the provision for the experimental data relevant for demonstrating (or better confirming) the code capabilities in predicting VVER accident management relevant scenarios measured in the PSB-VVER facility. The following Russian organizations were involved, with different extent, in the project: Electrogorsk Research and Engineering Center (EREC) owner of the PSB-VVER Integral Test Facility (ITF), EDO Gidropress (EDO-GP) designer of the VVER reactors and Kurchatov Institute (RRC KI). The Unit 3 of Balakovo NPP was selected as the reference unit and the specialists of the plant were also involved in the project. The design, the execution and the analysis of the experiments, selected in the final Test Matrix (TM), as well as the application of different thermal-hydraulic system codes, constituted the largest effort inside the Project. For sake of completeness, the experiments actually executed were sixteen instead of the foreseen twelve in order to provide additional information and data on the quality of the experimental database.

Fig. 1 PSB-VVER general view of the facility. The experiments were initially proposed as “ideal” experiments (4): “ideal” means that

they were selected taking into account: • the scaling issue and the Counterpart Tests (CT) performed in other Pressurized Water

Reactors (PWR) facilities to demonstrate the PSB-VVER capability; • the relevance for the accident management in the reference plant;


Vol. 2, No. 1, 2008

373

• accident management strategies of Balakovo NPP, i. e. optimal recovery procedures and Critical Safety Functions (CSF) restoration procedures, results of Balakovo Emergency Operating Procedure (EOP) analytical justification and intention to support the development of missing EOP. The first list of tests was proposed in the Terms of Reference (ToR) of the project. This

list was updated on the basis of the “ideal” test matrix proposed by UNIPI and RRC-KI (4). The final test matrix included the results from the discussion about the tests proposed in the “ideal” test matrix and took into account the suggestions of the other consultants (the Balakovo and EDO-GP specialists and EREC scientists). The final test matrix, in fact, considered that the boundary and initial conditions for each test must be adapted to the PSB-VVER facility. This can introduce some modifications that could affect the relevance of the test for the accident management in the reference plant, because of the PSB-VVER facility intrinsic limitations, which make impossible to perform some kind of tests (e.g. maximum cladding temperature has to be in the range 600-800 °C). Another relevant aspect, embedded in the final test matrix, was related to the code capability to reproduce the phenomena occurring during the selected tests. It has an important role in the demonstration of the code capability to realistically reproduce the plant behavior during an accident.

The list of experiments actually carried out in the frame of the Project R2.03/97, part A is presented in Table 1. Preliminary pre-test calculations in accordance with preliminary scenarios developed for experiments of Table 1 were executed using both CATHARE2/V1.5B and RELAP5/Mod3.3 codes. The final scenarios were, therefore, the result of a systematic methodology which considers the results of preliminary pre-test calculations, the characterization tests provided by EREC, the analyses performed at UNIPI, the discussions at the Working Group meetings in Pisa, as well as the further deeper analyses based on the final pre-test calculations and the NPP analyses preformed with RELAP5/Mod3.3 and CATHARE2/V1.5B codes. This methodology, described and discussed in the text, was the base for the execution of the largest experimental programme performed in the integral test facility PSB-VVER and suitable for assessing the code capabilities in predicting relevant scenarios occurring in VVER-1000 NPP.

2. The Procedure for Test Design

In the area of system thermal-hydraulics, the PSB-VVER (Fig. 1) is one of the largest test facilities put into operation with a power and volume scaling factor equal to 1/300. The volume-to-power scale, full-pressure, time-preserving scaling laws were adopted for the design of the test facility. PSB-VVER is the most qualified facility for the study of the VVER-1000 and the only one in operation. The facility includes all the main components of the primary system of the prototype NPP, till the connections between the steam and feed-water lines with the Steam Generators (SG). Emergency Core Cooling Systems (ECCS), are part of the facility, as well as all components and systems on the secondary side needed for the simulation of the concerned accident management related scenarios. The process of test design applied was based on the following three key topics: a) the scaling issue and the ‘reproduction/simulation’, with unavoidable distortions, of the expected performance of the reference NPP, b) the code assessment process involving the identification of phenomena challenging the code models, c) the features of the concerned integral test facility (scaling limitations, control logics, data acquisition system, instrumentation, etc.). The procedure adopted, within the present context (see Fig. 2), includes the logical steps hereafter listed

1) Overview of previous experiments executed in PSB-VVER, part of the OECD Project (5) in progress, of similar experiments performed in other test facilities and of data available from NPP. This implied the selection of counterpart and of similar tests (6), (7), (8), (9).


Vol. 2, No. 1, 2008

374

2) Consideration of thermal-hydraulic phenomena of the OECD/CSNI code validation matrix (10), (11) including the VVER matrix.

3) Consideration of phenomena specifically relevant for accident management, e.g. natural circulation, depressurization, discharge of saturated water from one vessel to another.

4) Consideration of the list of experiments indicated by the Term of Reference and, again, the related relevance to the accident management.

5) Consideration of the features of the PSB-VVER (4), (12): e.g. steady state conditions (low power, low flow at full pressure) at the test start.

Pre-test calc. T.# (Id) - Test type – Description

C R K T

T.#1 (LFW-25): Total LOFW – Failure of HPIS pumps and operator actions on secondary circuit depressurization.

X X

T.#2 (LFW-28): Total LOFW – Failure of HPIS pumps, operator actions on secondary and primary circuit depressurization.

X X

T.#3 (PrzVS-01): PORV stuck open – Similar to Zaporozhye NPP. X X T.#4 (CL-0.7-08): SBLOCA – Delayed AM similar to BETHSY9.1b. X X ●

T.#5 (SL-100-01): MSLB & PRISE of 42 mm (to be scaled) – HPIS pumps failure and primary circuit depressurization through the PORV valve.

X

T.#6 (LFW-27): Total LOFW – Feed and bleed procedure (pressurizer relief valve and HPIS pumps available).

X X

T.#7 (BO-05): SBO – Secondary circuit depressurization for passive feedwater supply from deaerator

X X

T.#8 (CL-0.5-03): SBLOCA – HPIS pumps failure (SIT initial pressure and coolant volume optimizations and evaluation of the CBA that cause primary pressure stabilization).

X X ● ●

T.#9 (PSh-1.4-05): PRISE of 100 mm (to be scaled) – BRU-A failure and stuck open in the affected SG and AM actions.

X

T.#10

(NC-6): NC – NC flow-rate and regimes established when draining PS coolant and keeping available the SG heat sink. DNB/DO occurrence, primary side refilling and observing hysteresis.

X, ▲

X

T.#11 (CL-0.7-12): SBLOCA – HPIS pumps failure, cooldown procedure and recovery of one HPIS train in affected loop (to define time of HPIS recovery).

X X

T.#12 (CL-0.7-11): SBLOCA – HPIS & LPIS pumps failures. Water supply by normal operation systems to primary circuit.

X X

T.#13 (BO-06): SBO – Single variant of T.#7: core power decreased of 100KW.

▲

T.#14 (PSh-1.4-07): PRISE of 100 mm (to be scaled) – Single variant of T.#9: no BRU-A stuck open.

▲

T.#15 (CL-0.7-13): SBLOCA –– Single variant of T.#4: opening of the PORV vale at 4000s.

▲ ●

T.#16 (*) – SBLOCA – Equal to T.#12. Demonstration of repeatability. Performed pre-test analyses legend: X = UNIPI ▲ = EREC ● = EDO-GP C = CATHARE2 R = RELAP5 K = KORSAR T = TRAP

Table 1 Summary of the PSB-VVER experiments executed and of the pre test calculations performed.


Vol. 2, No. 1, 2008

375

6) Availability of information and data of components of the Balance of Plant including, for instance the configurations of feed-water and steam lines and the characteristics of the installed valves.

7) Selection of codes: CATHARE2/V1.5B and RELAP5/Mod3.3. 8) Development of VVER-1000 and PSB-VVER (4) nodalizations for both

CATHARE2 and RELAP5 codes. 9) Qualification of VVER-1000 and PSB-VVER (4), (13), (14) nodalizations. 10) Ideal test matrix, to accomplish steps 1 ) to 3). 11) Actual test matrix to combine step 10) with steps 4) and 5), including the

specification for main test goals. 12) VVER-1000 calculations adopting the VVER-1000 nodalizations from step 9) to

substantiate, as far as possible the findings from step 3), including the consideration of the scaling issue (see also below).

13) PSB-VVER preliminary calculations adopting the nodalization from step 9) to address the scaling issue and the integral test facility features from step 5).

14) Evaluation of calculation results from step 13) and feedback with steps 3), 4) and 5). The result of this step is constituted by the final test designs.

15) Execution per each experiment, of pre-tests at UNIPI (“generic” test facility configuration) and of pre-tests at EREC (“specific-detailed” test facility configuration).

16) Calculation of uncertainty bands for pre-tests of two experiments. The process of test design continued till the evaluation of the experimental database. At the end of the process of experimental data analysis, any designed test was accepted, considering the correspondence between measured data (i.e. targets of the experiments) and the outcomes of steps 11) and 12) indicated above. The above listed steps have been also reported in the Fig. 2 which outlines the role of the facility and of the reference NPP in the process, particularly in the preliminary and final pre-test and post-test activities.

TOR requests

Preliminary list of transient

Preliminary BIC

Final BIC

Phenomena evaluation

Facility capability evaluation

Preliminary pre-tests

Update list of tests

Final pre-tests

Post-tests

Accuracy evaluation

VVER-1000 data

Preliminary (initial) NPP calculations

Russian comments

Preliminary (final) NPP calculations

VVER-1000 simulations

Scaling evaluation

Tests execution

A

JG

N

B

C

D

E

F

H

I

K

LM

O

P

Q

R

Fig. 2 Procedure for test design: flowchart of the methodology adopted in the framework of the TACIS 2.03/97.

For sake of completeness, the created experimental database consisted of four key parts: a) the integral test facility description including test specific configuration and description


Vol. 2, No. 1, 2008

376

of components added for the execution of individual experiments; b) the results from the characterization or shake-down tests (pressure drops, heat losses, volume vs. height, etc.); c) the logic of imposed events in each experiment; d) the resulting sequence of main events and the time trends of a significant number of quantities.

3. The Nodalizations Developed and Qualification

The description of the nodalizations used in the framework of the present activity requires a long section that can be considered outside of the scope of the present paper. Nevertheless these computational tools were fundamental also during the design activity. For this reason this section has been included but the information here are limited to the list of the tools used. The design activity was supported on the application of four codes (different versions) and nine nodalizations, developed by three institutions, as hereafter summarized:

a. UNIPI PSB-VVER nodalization by CATHARE2 (Fig. 3) (13); b. UNIPI PSB-VVER nodalization by RELAP5 (Fig. 4) (14); c. EREC PSB-VVER nodalization by CATHARE2; d. EREC PSB-VVER nodalization by RELAP5; e. EDO-Gidropress PSB-VVER nodalization by KORSAR; f. EDO-Gidropress PSB-VVER nodalization by TEACH; g. UNIPI Balakovo unit 3 VVER-1000/320 nodalization by CATHARE2; h. UNIPI Balakovo unit 3 VVER-1000/320 nodalization by RELAP5; i. EDO-Gidropress Balakovo unit 3 VVER-1000/320 nodalization by TEACH.

A complete description of the nodalizations, as well as of all other activities, is reported in the published final report of the Project (4).

Fig. 3 CATHARE2/V1.5B nodalization: general scheme.


Vol. 2, No. 1, 2008

377

Fig. 4 RELAP5/Mod3.3 nodalization: general scheme.

4. The Design of the PSB-VVER Experiments

The methodology developed to design the 12 scenarios starting from the contractual requirements (3) is outlined in Fig. 2. The details about the procedure represented are hereafter discussed including sample results of the activities. The complete description of the activity is discussed in section 5 of Ref. (4), pp. 260-488.

a) Starting from the general description of the 12 scenarios to be simulated in the experimental facility (Term of Reference of the Project) (3), a detailed design of every scenario was needed (e.g. definition of the scaled core power, diameter of the valves throttles suitable for scaled mass flow rate, diameter of the throttle simulating the break, etc.). The pre-tests analyses were scheduled by UNIPI with CATHARE2 code (committed) and in parallel with RELAP5 code in order to have additional support information. Additional contributions were provided by EREC (RELAP5 code) and EDO-GP (KORSAR and TRAP codes) performing additional pre-tests of selected transients, see Table 1.

b) The main phenomena connected to the general scenarios were addressed preliminarily given their relevance in demonstrating the thermal-hydraulic system codes capabilities. With regards to this issue, it was stated that physical phenomena of VVER-1000 accident management scenarios are analogous to those discussed in the validation matrices for VVER Small Break (SB) Loss Of Coolant Accidents (LOCAs), Intermediate Break (IB) LOCAs and transients (see Ref. (10)). In fact, scenarios with simulation of accident management are initiated by the same initial events. Large Break (LB) LOCAs and related physical phenomena were not considered because at the time PSB-VVER was not able to simulate these accidents. It was also observed that procedures of “feed and bleed” type in primary and secondary circuits are important elements of accident management. Therefore the phenomenon “flow rates through valves” was included into the validation matrix for accident management. Moreover the “hydro-accumulator behavior”, not included in the validation matrix (10) specific for VVER, was considered and included between the relevant phenomena as it is in the validation matrix for PWR (11).

c) The capabilities of the facility were investigated with EREC staff in order to


Vol. 2, No. 1, 2008

378

characterize and to demonstrate its performance. In this connection, the analysis was focused on the relevant phenomena that can be induced, reproduced and observed by the actual configuration of the test facility and by its measurement system. The information provided was: the relevance of the phenomena in connection with the selected scenarios and with item b); the capability of the facility to induce the phenomena defined in item b) and the capability of the measurement system to detect and supply data relevant for code assessment. Moreover the characterization tests (already available) were collected (e.g. Refs. (15) and (16)), analyzed and compared with the standard practices of other European facility (LOBI, BETHSY, SPES, PKL, etc.). The characterization tests may be subdivided in four categories: hardware and/or geometric characterization tests such as volume versus height distributions, main components simulation (number of loops, pumps, number and location of ECCS, etc.); physical characterization tests (e.g. pressure distribution, primary to secondary heat transfer, heat losses, etc.); components characterization tests like pumps, valves, special devices (e.g. break simulator), etc.; shake-down tests such as long lasting transients performed to check the overall system performance (e.g. instrumentation and data acquisition response, test conduct capability, etc.). Special care was paid to the scaling criteria of the facility and its performance in comparison with the others well known PWR integral test facility: for instance, the performances of the steam generators (with the helicoidal tubes) in transient conditions were investigated in order to demonstrate the correctness of the design criteria adopted with respect to the prototype VVER-1000 NPP (12).

The analysis of the counterpart test executed in PSB-VVER facility, in the framework of the OECD Project, allowed to draw some conclusions based on the experimental data (Fig. 5) and the comparison with those obtained in other facilities. In fact, notwithstanding the differences among the considered facilities regarding the geometry, the scaling factor, the prototype plant that they represent, the overall similarity for this type of accident was evident. Furthermore, the phenomena that characterize the transient were replicated in all the facilities. This means they were not affected by the scaling. For sake of completeness it should be noted that Figure 5 includes experimental data of low power and of full power for two facilities, i.e. LOBI and SPES in which the different amount of core power did not affect the key phenomena experienced during this transient and (practically) it did not move the timing of their occurrence. This was a confirmation of the impossibility to extrapolate directly

-500. 0 500. 1000. 1500. 2000. 2500. 3000.Time (s)

100

150

200

250

300

350

400

450

500

550

600

Tem

pera

ture

(°C

)

XXX SB_EXPPSB

X XX

XX

X

X X X X X X X X X X X

X

X

YYY EL34 LOBI low power

Y YY Y

Y

Y

YY Y Y Y Y Y Y Y Y

Y

Y

Y

ZZZ EL44 LOBI high power

Z Z

Z ZZ

Z

ZZ Z Z Z Z Z Z Z

Z

Z

Z

Z

VVV SB03 SPES low power

VV V

VV

VV V V V V V V V

V

V

V V V V

JJJ SB04 SPES high power

J

J J JJ

JJ J J J J J J J J J

J

JJ

J

HHH Bethsy

H HH

H H H H H H H H H H H

H

H

H

HH H### LSTF

#

#

##

# # # # # # # # # ##

#

#

#

#

(a) Rod surface temperature (high level)

-500. 0 500. 1000. 1500. 2000. 2500. 3000.Time (s)

0

20

40

60

80

100

120

% P

Smas

s (-)

XXX SB_EXP1PSBX

X

XX X

X X X X X X X X X X X X X X X

YYY EL34 LOBI low power

Y Y

Y

YY

Y YY

Y Y Y Y Y Y Y Y Y YY

ZZZ EL44 LOBI high power

Z Z

Z

Z

ZZ Z Z

Z Z Z Z Z Z Z Z Z Z

Z

VVV SPES low power

V

V

V

VVV VVVV V V

V VV

V

VVVV

JJJ SB04 SPES high power

J

J

J

JJ

J J J J J J J J J J J J J JJ

HHH Bethsy

H

H

H HH H H H H H H H H H H H H H H H

### LSTF#

#

##

# ## # # # # # # # # # # # #

(b) Non-dimensional primary mass inventory

Fig. 5 The scaling issue, the SBLOCA CT – PSB-VVER, LOBI, SPES, BETHSY, LSTF

experimental data.

PSB-VVER

PSB-VVER


Vol. 2, No. 1, 2008

379

the time trend of any parameters from the trends measured in another installation.

Moreover in the OECD database was also available a Natural Circulation (NC) test (17) carried out through stepwise reduction of the primary coolant inventory. The analysis of these experimental data (18) allowed the understanding about the relationship between natural circulation modes and the mass inventory and to provide information on the PSB-VVER natural circulation performance (and indirectly the VVER-1000 performance) using the natural circulation flow map and the database available at UNIPI (7). Figure 6 shows the natural circulation map created by the envelop of ten experimental data obtained in six PWR simulators and putting them into a phase space (core mass flow rate over the core power versus residual Primary Side (PS) mass over PS volume). In Fig. 7, the Residual Mass (RM), in percentage, at the dryout occurrence versus the volume scaling factor (Kv) of different facilities (essentially PWR simulators) is reported. The comparison confirmed that there are not correlations between Kv and the residual mass at the dryout. In the PSB-VVER facility the dryout occurred at the lowest value of the residual mass demonstrating that it has a greater capability of heat removal, in natural circulation conditions, then the equivalent (in terms of volume scaling factor) PWR simulators.

d) In parallel, the data related to the reference NPP, Balakovo Unit 3 (19), were analyzed, and used to set up the NPP nodalizations. The RELAP5 code model was derived by an existing “generic” VVER-1000 nodalization developed in several years at UNIPI and used in various international projects (20). On the basis of the RELAP5 nodalization, of the available data and the information directly provided by the Balakovo NPP specialists, a VVER-1000 model was developed and qualified also for CATHARE2 code (21).

e) A first list of generic transients was selected. The test matrices constitute the basis for experimental campaigns in integral test facility. A test matrix includes the target experimental scenarios, the relevant phenomena expected and the experiments identification. Two “ideal” test matrices and one Anticipated Transient Without SCRAM (ATWS) test matrix were proposed by UNIPI and by RRC-KI. These were derived from a review of needs in the area, considering all past and present experimental programs involving integral

800 600 400 2000

5

10

15

20

W/P

[ K

g/ M

W.s

]

RM/V [Kg/m3]

Fig. 6 The scaling issue, the natural circulation test: PSB-VVER experimental and calculated data compared with the natural circulation flow map.

Fig. 7 The scaling issue, the natural circulation test: a comparison of the Residual Mass (RM)

fraction at core dryout for different installations as a function of the volume scaling factor.

EXP DATA NOT

VALID

PSB-VVER data

RELAP5 simulation


Vol. 2, No. 1, 2008

380

test facilities (e.g. LOBI, LOFT, BETHSY, PACTEL, LSTF, PMK, PKL, etc.) and focused on phenomena rather than on accident management relevant scenarios. As examples of tests part of the above two types of matrices (4), it can be mentioned the test T.#7 of the UNIPI “ideal” test matrix aimed at maximizing the impact of thermal-hydraulic conditions during the Pressurized Thermal Shock (PTS) event and the test T.#7 of the ATWS matrix aimed at characterizing the phenomenon of sub-cooled boiling during fast transients. The “ideal” test matrices constituted the background to derive the final test matrix. The process involved the consideration of the Term of Reference experiments and the needs in the area of accident management. Twelve experiments (number part of the Term of Reference requirements) were designed. Later on, three additional experiments were performed and one more “repeated” experiment was included in the database. Therefore, the final Project matrix included sixteen experiments (see Table 1), each of them designed with specific objectives, such as T.#2 of Table 1 to test the effectiveness of the steps of the accident management strategy reference for the Project (“feed and bleed” procedure); T.#4, analogous to Bethsy test 9.1b, to address the scaling issue and to evaluate the similarity in the response between VVER-1000 and PWR; T.#5 to investigate complex NPP scenario involving multiple failures and simultaneous primary-to-secondary heat and mass transfer; etc.

f) The Boundary and Initial Conditions (BIC) of the selected tests were identified as well as the Imposed Sequence of the Events (ISE). These BIC constituted a first attempt to be evaluated by the application of the computer codes, item g).

g) Preliminary pre-test calculations adopting the PSB-VVER nodalizations by RELAP5 and CATHARE2 codes were carried out. The nodalizations of the facility were previously qualified (13), (14), (22) using the data already available in the framework of another international project (the OECD PSB-VVER project (5)). The preliminary pre-tests were useful to check the relevance of the selected scenarios and the correctness of the related boundary and initial conditions with reference to three main aspects:

• the effectiveness of the accident management procedures and their relevance; • the aspects challenging for the codes; • experimental relevant remarks. h) The results from the activity of the previous step caused the modification of some

tests (e.g. general aspects, imposed events, initial conditions, etc.). An updated list of tests (mainly concerning the boundary and initial conditions, but also the imposed sequence of main events) was defined (4).

i) A new set of pre-test calculations (4) was performed taking into account the new boundary and initial conditions and imposed sequence of main events. The final pre-tests (sample results are in Fig. 11a and 12a) were executed not only for checking the final scenarios, but also for addressing the issues related to the design of experiments. During this phase the experimental data of three experiments (T.#4, T.#11 and T.#12) were already available, providing relevant information and feedbacks. In particular, the correct compensation of the heat losses was an important issue, given the long duration of many planned experiments (e.g. Station BlackOut, SBO; Loss Of FeedWater, LOFW, etc.). As the heat losses of the PSB-VVER facility are up to 2% of nominal power (16), the core model power would decrease quickly to the heat losses level if they are not compensated. This would preclude continuation of the experiments simulating accident management procedures. The experiment T.#13 (Fig. 8a and 8b) was designed with the objective of verifying the quantity of the thermal losses of the test facility during and at the end of high pressure transients. In T.#13, the power supplied to the core, when steam generators secondary sides are empty, is equal to the power released to the environment because the heat losses, while the experiment T.#7 was designed with the objective to simulate a station blackout scenario with an accident management procedures. In conclusion, the curve “core model power versus time” for PSB-VVER facility was defined in order to include the


Vol. 2, No. 1, 2008

381

following contributions, using the same procedures already done at UNIPI for experiments carried out on LOBI and PIPER-ONE facilities: • curve of power decay heat after

scram for VVER-1000; • compensation of PSB-VVER

heat losses (to be evaluated depending upon the transient condition). The heat losses compensation was then subdivided in two contributions: 1) the pressurizer heat losses compensated through the use of the pressurizer heaters and 2) the Reactor Pressure Vessel (RPV), the Reactor Coolant Loops (RCL), and the steam generators secondary side (depending upon the transient scenarios) heat losses that were compensated increasing the core power.

On the contrary, the pre-tests of the experiment T.#3 (designed as counterpart test of Zaporozhye NPP Unit #1 incident in 1995) (23) highlighted the scaling distortions of a small facility with respect to a real NPP unit. In this case, the ratio between the masses of structure metal and of the water of the coolant is larger in the facility than in the VVER-1000. This causes a larger amount of energy released from the metal to the coolant in the experiment (as it is demonstrated by the measuring systems in Fig. 9), that was taken into account in the design of the core power (reduced to zero at the beginning of the depressurization phase, without complete compensation of the heat released by the structures).

j) Preliminary calculations were performed adopting the VVER-1000 NPP nodalization to highlight the phenomena expected and of interest in the plant.

k) Russian comments were received concerning the implementation of the boundary and initial conditions and imposed sequence of events applied to the NPP transient scenarios.

l) A new set of NPP calculations, aimed at the verifications of the up-dated BIC and ISE of the scenarios, was carried out including the results of the step k): samples results are summarized in Fig. 11b and 12b.

m) The final set of tests and related boundary and initial conditions and imposed sequence of events were defined, considering the results from steps i) and l), and are the basis for the executive plan of the experiments performed in the facility. Further final pre-test calculations were needed on the basis of new information on PSB-VVER special device set-up (e.g. main steam line break for T.#5, primary to secondary leak system T.#9); of the scaling distortions discussed at the end of item i) (T.#3) and of the results of T.#13 (which showed larger heat losses in very long lasting transients, like T.#6, T.#1 and T.#2).

-.5 0 .5 1 1.5 2 2.5X 104Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

Pres

sure

(MPa

)

WinGraf 4.1 - 09-01-2007

XXX T#7_EXP3 YC01P17

X X X X X XX XX

XX X

X

XX

X X X XYYY T#7_EXP3 YB01P01

Y YY

Y Y Y YY

YY YY Y Y

Y Y Y YY

ZZZ T#7_EXP3 YB02P01

Z Z ZZ Z Z Z Z Z Z

ZZ Z Z Z Z Z Z Z

VVV T#13_EXP3YC01P17

V V V V V V V V V V V V V V V V V V V V

JJJ T#13_EXP3YB01P01

J JJ J

J J JJ J J J J J J J J J J

JJ

HHHT#13_EXP3YB02P01

H HH H

H H HH H H H H H H H H H H

HH

(a) Primary and SG1, 2 pressures

-.5 0 .5 1 1.5 2 2.5X 104Time (s)

0

200

400

600

800

1000

1200

1400

1600

1800

Pow

er (k

W)

WinGraf 4.1 - 09-01-2007

XXX T#7_EXP1 *YC01N01X

XX X X X X X X X X X X X X X X X X

YYY T#13_EXP1*YC01N01Y

Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

(b) Core power

Fig. 8 Relevance of the heat losses compensation in SBO scenarios (T.#7 and

T.#13): recoded experimental data.

T.#7 T.#13


Vol. 2, No. 1, 2008

382

Those pre-tests constituted the basis for the executive plan of the just mentioned experiments performed in the facility.

n) Execution of tests in the facility included three additional tests consisting in a single variant parameter of the accident management test matrix scenarios and the repetition of the T.#12 (demonstration of repeatability), see Table 1 and Fig. 10a.

The repeatability of an experimental test is a critical issue in any integral test facility and it is connected with the “control” of the boundary conditions. Such issue is more critical in case of long lasting transients because the difficulties in keeping under control causes such as: heat losses (absolute value); heat losses (spatial and timing distribution); very small leakages (e.g. valves, primary to secondary side leakages); minor changes in pressure drop due to chemical processes (e.g. corrosion, erosion, deposition); valves position repeatability; timing of action when manual operations are considered in the test. Given its relevance in connection with the evaluation of the code results the repeatability and consistency of the experimental data was deeply investigated, comparing also the experimental data of the others tests having similar boundary and initial conditions (see Fig. 10b). The execution of the experiments, then, confirmed the considerations about the core power compensation for long transients, see item i) and Fig. 8. Each experiment was provided with a suitable documentation consisting of a Test Analysis Report (TAR) and an Experimental Data Report (EDR). Samples data recorded during the experiments T.#5 and T.#7 are shown in Fig. 11c and 12c.

o) The availability of the experimental results made possible the execution of the post-test calculations in order to demonstrate the availability of qualified computational tools. The codes applied in this phase were at different extent CATAHRE2 (F), RELAP5 (US-NRC), KORSAR (R) and ATHLET(D). The post-test analyses were performed taking into account the actual boundary and initial conditions and imposed sequence of events of the experiments.

p) The codes capabilities and the qualification level were checked using a standard

-200. 0 200. 400. 600. 800. 1000.Time (s)

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

Pres

sure

(MPa

)

WinGraf 4.1 - 03-23-2006

XXX PSpres NPPX

X

X

X

XX

X

XX

XX X X X

XX

X XX X

YYY ET03c PSpres PSB

Y Y Y Y

Y

Y

Y

Y

Y

Y

YY

YY Y Y Y Y Y

Y

Fig. 9 The scaling issues of the counterpart test to Zaporozhye NPP: plant and

PSB-VVER measured primary pressures.

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

0 2000 4000 6000 8000 10000

Time (s)

Pres

sure

(MPa

)

Test 12Test 12-2

(a) T.#12 & 16: primary side pressure (all transient)

4.0

6.0

8.0

10.0

12.0

14.0

16.0

0 200 400 600 800 1000 1200 1400 1600 1800

Time (s)

Pres

sure

(MPa

)

Test 11Test 12

(b) T.#11 & 12: primary side pressure (0-1800s)

Fig. 10 Demonstration of repeatability

Zaporozhye NPP

PSB-VVER ITF


Vol. 2, No. 1, 2008

383

procedure (24), (25), aimed at evaluating the accuracy of the available computational tools in predicting the experiments. The evaluation of the accuracy was fulfilled from the qualitative and the quantitative point of view, see Refs. (4) and (27).

q) The experiments were also simulated using the VVER-1000 nodalizations of Balakovo 3 NPP CATHARE2 as well as RELAP5 (4), to check the expected phenomena in the plant.

r) Considerations on scaling (4) were also carried out through a systematic comparison between the experimental data, the post test analyses and the simulation of the scenarios with the NPP nodalizations (CATHARE2 and RELAP5).

-2000. 0 2000. 4000. 6000. 8000. 10000.Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0Pr

essu

re (M

Pa)

WinGraf 4.1 04 12 2006

XXX tpsbtr_aR05_A01i p156250000

X XX

X

XX

XX

XX X X X X X X X X X X

YYY tpsbtr_aR05_A01i p608010000

Y YY Y Y

YY

YY Y Y Y Y Y Y Y Y Y Y Y

ZZZ tpsbtr_aR05_A01i p908010000

Z

Z

Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z

(a) PSB-VVER pre-test calculation

-.25 0 .25 .5 .75 1 1.25 1.5 1.75 2X 104Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

Pres

sure

(MPa

)

WinGraf 4.1 04 12 2006

XXX tpsbtr_aR07_01 p156010000

X X X X XX X X X X X X X

X

XX

X

XX

XYYY tpsbtr_aR07_01 p608010000

YY Y Y Y Y Y Y Y

Y

Y Y Y Y Y Y Y Y Y Y

ZZZ tpsbtr_aR07_01 p708010000

Z ZZ Z Z Z Z Z

Z Z Z Z Z Z Z Z Z Z Z Z

(a) PSB-VVER pre-test calculation

-1000. 0 1000. 2000. 3000. 4000. 5000. 6000.Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

Pres

sure

(MPa

)

G a 0 006

XXX t5 p30070000

X

X XX

XX

X

XX

X

X X X X X X X X X X

YYY t5 p600010000

Y YY Y Y

Y YY

Y

YY Y Y Y

Y Y Y Y Y Y

ZZZ t5 p900010000

Z

Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z

(b) NPP pre-test calculation

-.2 0 .2 .4 .6 .8 1 1.2 1.4 1.6X 104Time (s)

0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

22.5Pr

essu

re (M

Pa)

G a 0 006

XXX t7 p30070000

X XX

XX X X X

XX

XX X

X

XX

X X X

YYY t7 p600010000

YY

Y Y Y YY

Y Y Y Y Y Y Y Y Y Y Y Y

ZZZ t7 p700010000

ZZ

Z Z Z ZZ

Z Z Z Z Z Z Z Z Z Z Z Z

(b) NPP pre-test calculation

-1000. 0 1000. 2000. 3000. 4000. 5000. 6000. 7000. 8000.Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

Pres

sure

(MPa

)

G a 0 006

XXX ET05dYP01P01X

X

X

X

X X

XX X X X X X X X X X X X X

YYY ET05bYB01P01

Y Y Y Y Y YY

Y Y Y Y Y Y Y Y Y Y Y Y Y

ZZZ ET05cYB04P01

Z

Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z

(c) PSB-VVER experimental data

-.2 0 .2 .4 .6 .8 1 1.2 1.4 1.6X 104Time (s)

0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

Pres

sure

(MPa

)

G a 0 006

XXX T_#7_EXP4YP01P01

X X X X X X X X X

XX X

X

XX

X X X X

YYY T_#7_EXP3YB01P01

Y YY

Y Y Y YY

Y Y YY Y Y

Y Y Y Y Y

ZZZ T_#7_EXP3YB02P01

Z Z ZZ Z Z Z Z Z Z

ZZ Z Z Z Z Z Z Z

(c) PSB-VVER experimental data

Fig. 11. T.#5 – (SL-100-01): Primary side pressure, SG 1 and 4 pressures.

Fig. 12 T.#7 – (BO-05): Primary side pressure, SG 1 and 2 pressures.


Vol. 2, No. 1, 2008

384

5. Conclusions

The experimental programme, performed in the framework of the Project TACIS 2.03/97, constitutes a relevant database for thermal hydraulic system code validation. It has been executed at PSB-VVER facility that is the most suitable and qualified for the study of the VVER-1000 NPP and the only one in operation. The final scenarios of the experiments were the result of a systematic methodology developed at UNIPI and described and discussed in the text. All the fulfilled steps of this methodology are highlighted, starting form the scenarios foreseen in the final test matrix until the execution of the experiments. The design process considers three main key topics: a) the scaling issue and the simulation, with unavoidable distortions, of the expected performance of the reference NPP, b) the code assessment process involving the identification of phenomena challenging the code models, c) the features of the concerned integral test facility (scaling limitations, control logics, data acquisition system, instrumentation, etc.). The steps of the design procedure, with particular regard to the boundary and initial conditions and imposed sequence of events are presented, as well as the tools used (codes and nodalizations of reference plant and of PSB-VVER facility). The final pre-test calculations were based on the results of the preliminary calculations, following the contributions of RRC KI and Balakovo NPP specialists, the information derived from the characterization tests, the scaling analyses, etc. The comparison between the various analyses related to the reference plant and to the expected experimental results (pre-test), showed that generally the transient in the experimental facility is slower than in the real plant. This issue, connected mainly with the effect of the heat losses, required specific considerations in order to evaluate the power compensation for the different scenarios. In this connection a specific additional experiment (T.#13) was executed. However, the relevant phenomena envisaged in the reference power plant were expected also to be observed (and occurred) in the experimental facility.

In conclusion, the methodology described was developed and applied for executing a valuable experimental database suitable for analyzing the phenomena and assessing the code capabilities in predicting scenarios occurring in VVER-1000. This was a necessary step for the qualification of the analytical tools used to evaluate and set up the accident management procedures in the reference unit#3 of Balakovo NPP.

Acknowledgement

The overall activity described in the paper was performed with the contributions of a lot of researchers, among which one shall mention D. Araneo, N. Muellner, M. Cherubini, W. Giannotti and G. M. Galassi (from University of Pisa) as well as the Russian Scientists of EDO-Gidropress, EREC and RRC Kurchatov Institute.

References

(1) Addabbo, C., D’Auria, F., Maintenance of European LWR Integral System Test Thermal-Hydraulic Databases - Status Report,. (2001), European Commission Report , CERTA/SC/D6.

(2) Melikhov, O.I., et. Al., Report about PSB-VVER description (including measurement system), PSB-VVER REPORT, PSB-03, OECD NEA, (2003).

(3) EC Project R2.03/97, Software Development for Accident Analysis of VVER and RBMK Reactors, Annex II: Terms of Reference – TACIS 1997 Programme, Russian Federation, Nuclear Safety” Aidco/A/5 TOR for R2.03/97 version 240603.

(4) D'Auria, F., et. Al., Accident Management Technology in VVER-1000, (2006), pp 0-1440, Industrie Grafiche Pacini Editore.

(5) Melikhov, O.I., Melikhov, V.I., Parfenov, I.V., PSB-VVER tests priority for the OECD PSB-VVER Project, PSB-VVER REPORT, PSB-01, OECD NEA, (2003).


Vol. 2, No. 1, 2008

385

(6) D’Auria, F., Karwat, H., Mazzini, M., Planning of counterpart test in LWR experimental simulators, Proc. of 25th National Heat Transfer Conference, (1998).

(7) D’Auria, F., Frogheri M., Use of a natural circulation flow map for assessing PWR performance, Proc. of Single and Two-Phase Natural Circulation EUROTHERM Seminar, No. 63, (1999), pp. 407-416.

(8) D’Auria, F., et. Al., Scaling of natural circulation in PWR systems, Nuclear Engineering and Design, Vol. 132, (1991), pp. 187-205.

(9) Belsito, S., D'Auria, F., Galassi, G. M., Application of a statistical model to the evaluation of counterpart test data base, J. Kerntechnik, Vol. 59 No. 3, (1994).

(10) OECD/NEA, 2001, Validation Matrix for the Assessment of Thermal-Hydraulic Codes for VVER LOCA and Transients. NEA/CSNI/R(2001)4, June 2001.

(11) OECD/ CSNI, 1996, CSNI Integral Test Facility Validation Matrix for the Assessment of Thermal-Hydraulic Codes for LWR LOCA and Transients. NEA/CSNI/R(96)17, July 1996.

(12) Cherubini, M., et. Al., Comparison between PSB and VVER 1000 steam generator modeling, Proc. of 6th Int. Seminar on Horizontal Steam Generators, (2004), CD-ROM.

(13) Del Nevo, A., et. Al., Nodalization qualification process of the PSB-VVER facility for CATHARE2 thermal-hydraulic code, Proc. of 5th Int. Conf. on Nuclear Option in Countries with Small and Medium Electricity Grids, (2004), CD-ROM.

(14) Cherubini, M., Scaling Analysis for Russian WWER Type Reactor, Master Degree Thesis in Nuclear Engineer , University of Pisa, 2004.

(15) Melikhov, O.I., et. Al., Guidelines for performance of PSB-VVER tests, PSB-VVER REPORT, PSB-02, OECD NEA, (2003).

(16) Elkin, I.V., Dremin, G. I., Lipatov, I. A., Galchanskaya, S. A., Report about PSB-VVER heat losses, PSB-VVER REPORT, PSB-04, OECD NEA, (2003).

(17) Melikhov, O.I., et. Al., Post-test Full Experimental Data Report (Test 2), PSB-VVER REPORT, PSB-15, OECD NEA, (2003).

(18) Cherubini, M., . Al., RELAP5 simulation of a natural circulation test in the PSB-VVER test facility, Proc. of the 11th Int. Topical Meeting on Nuclear Reactor Thermal-Hydraulics (NURETH-11), (2005), CD-ROM.

(19) RISKAUDIT, Input Data Base for Analyses on VVER-1000 type NPP (Balakovo), Final Report, Report No 87, (1997).

(20) Cherubini, M., et. Al., Scaling of small break LOCA in VVER-1000 system, Proc. of 4th Int. Conf. Safety Assurance of Nuclear Power Plants with WWER, (2005), CD-ROM.

(21) Araneo, D., et. Al., Scaling of small break LOCA in VVER-1000 system, Proc. of Nuclear Energy for New Europe 2005 Int. Conf., (2005), CD-ROM.

(22) Del Nevo, A., et. Al., 11% Upper plenum break: application of RELAP5-3D© and comparison with other codes, Proc. of RELAP5/ATHENA Int. Users Seminar, (2004).

(23) Romanyuk, S. S., Kocharyants, O. R., Analysis pf temperature field of pressure vessel of VVER-1000 under accidents with RELAP5/MOD3.1, Proc. of the 2nd Conference on Safety Assurance of NPP with WWER, (2001)

(24) D'Auria, F., Bousbia-Salah, A., Petruzzi, A., Del Nevo, A., State of the art in using best estimate calculation tools in nuclear technology, Nuclear Engineering and Technology, vol. 38, Num. 1, (2006), pp. 11-32.

(25) Prosek, A., D’Auria, F., Mavko, B., Review of quantitative accuracy assessment with Fast Fourier Transform Based Method (FFTBM), J. Nuclear Engineering & Design, Vol 217, N. 1, (2002), pp. 179-206.

(26) Del Nevo, A:, et. Al., Thermal-hydraulic system code assessment for transient analyses involving accident management procedures in VVER-1000 reactors, Proc. of 12th Int. Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-12), (2007), CD-ROM.

the design of psb-vver experiments relevant to accident ......journal of power and energy systems...

Documents