ML20213A438
Text
_ _. _
I 9
Report on Visit to Japan Japan Atomic Energy Research Institute and Science and Technoloay Agency by Clifford Anderson Generic Issues Branch Division of Safety Technology Office of Nuclear Reactor Regulation Nuclear P.egulatory Comission
(.
i October 13 - 17, 1980
~~
8702030277 870210 PDR FOIA MORROW 85-782 PDR 277 l _ - -- . - - . - . _ . ._ . . . _ ,
Abstract During the week of October 12, 1980, Professor Robert Scanlon of Princeton University (NRC Consultant) and I visited the Japan Atomic Energy Research Institute (JAERI) near Mito, Japan and the Science and Technology Agency (STA) in Tokyo, Japan. The purpose of this trip was to discuss the results of recent Mark II tests and analysis at JAERI and future Mark II related work at JAERI. We also discussed Mark II licensing efforts in Japan with personnel from the nuclear regulatory body in Japan, (STA),
and members of the Japanese Mark II owners. It was also our purpose to encourage information exchance related to Mark II containment design verification.
Some progress was made relative to resolution of the instrumentation questions raised by the USNRC and the US Mark II owners during a previous trip in June 1980. However, questions still remain in this area. He encouraged the JAERI people to resolve these questions prior to conducting additional tests.
Mr. Shiba of JAERI indicated that significant progress has been made towards resolving certain liark II containment pool swell related licensing issues in Japan. However, it was not apparent that much progress had been made towards development of Japanese licensing positions related to
. steam loads (i.e., condensation oscillations, chugging and lateral loads). In these areas, it appears that the US is still taking the lead in developing load specifications.
Dr. Nozawa, head of the Division of Reactor Safety at JAERI stated that recent progress has been made related to improve US access to Mark II test results for the tests conducted at JAERI.
6
-r
e Contents .
- 1. Chronoloaical Summary
- 2. Persons Contacted
- 3. Pool Swell
- 4. Steam Loads
- 5. Witness of Test 3104
- 6. Instrumentation
- 7. Shaker / Hammer Tests
- 8. Mark II Containment Licensing in Japan
- 9. JAERI/USNRC Relations
- 10. List of Documents Received C
h
. 1. Chronological Summary ,
Monday (10/13/80) we arrived at JAERI. We were introduced to the staff working on Mark II programs and toured the Mark II Containment test facility. We briefly discussed each of the week's discussion topics including JAERI instrumentation problems, future tests, pool swell loads, steam loads, structural modifications of the test facility, Fluid Structure Interactions (FSI), and future shaker tests. Comprehensive discussions were conducted relative to pool swell observations in the JAERI facility.
Tuesday we continued discussions related to JAERI pool swell observations and discussed air mass analysis for the JAERI tests.
Dr. Kukida presented observations related to condensation oscillations (CO) and Ch' u gging Loads observed in the JAERI Mark II test facility (CRT). On Tuesday night we witnessed test #3104 in the CRT.
Wednesday we discussed our concerns related to CRT instrumentation with Mr. Namatame. Mr. Takashita discussed the upcoming shaker tests of the JAERI test facility and their proposed structural modification of CRT to address FSI questions. We presented the steam loads under consideration in the US and the role of the JAERI tests in confirming these loads.
p' L
Thursday we traveled to the STA offices in Tokyo. I presented the USNRC position on steam (C0 and Chugging) loads and load combinations.
We also discussed the results of preliminary evaluation studies of US and Japanese Mark II Containments to pool dynamic loads. These discussions were conducted between personnel representin'g the USNRC, JAERI, STA and the Japanese Mark II owners and vendors Friday we discussed the status of Japan's Mark II licensing activities.
- and summarized the week's activities.
l 2. Persons Contacted .
Dr. M. Nozawa, Head, Reactor Safety Research, JAERI Mr. M. Shiba, Chief, Engineering Safety Lab I, JAERI Mr. K. Namatame, CRT Program Manager, JAERI Dr. Y. Kukita, CRT Research Leader, JAERI Mr. S. Takashita, CRT Research Engineer, JAERI Dr. Ushio, Science and Technology Agency (STA), Nuclear Safety Bureau, Atomic Energy Bureau Dr. Ebine, Ministry of International Trade and Industry, (MITI)
Mr. T. Enomoto, Asst. Mar., Nuclear Power Construction Dept.,
Tokyo, Electric Power Co.
Mr. T. Horiuchi, Mgr., Nuclear Power Plant Engineering Dept.,
Hitachi Ltd.
Mr. H. Aoki, Safety Engineer, Nuclear Engr. Dept., Toshiba Corp.
l l
- 3. Pool Swell Pool Swell Height JAERI tests were compared to the USNRC The US Mark II owner's pool swell model (PSAM) pool conservative yields swell criteria.results for the pool motion parameters of velocity, acceleration and maximum height. However, significant froth was observed in the tests.
Observations of the froth height varied considering the level probes used. Short probes indicated a greater amount of froth activity than the long probes. The short probes data is in question.
JAERI plans to use an increased number of long probes in future tests. The significant froth activity appears to be due to vent braces above the suppression pool surface. This bracing configuration is typical of plants in Japan but not in the US. Other reasons cited for the froth include Taylor instability in the JAERI 3D tests and the non-flat pool shape resulting from the tight vent cluster in the JAERI tests.
Diaphraam Uploads The JAERI tests exhibited several higher then anticipated diaphragm uploads. The diaphragm separates the drywell and the wetwell volumes. One test indicated an upload a little above 2.5 psi.
This upload occured during a test when the wetwell/drywell vacuum breaker failed in the closed position. The USNRC criteria upload criteria applied to CRT yields a value of 4.3 psi. Dr. Kukida questioned the basis for our criteria. Additional CRT tests may be conducted to investigate the effect of break size, vacuum breaker operability and pool temperature on diaphragm uploads.
- Mass and Energy Release (Liquid Break)
Pool swell loads for Mark II plants are determined from the Pool Swell Analytical Model (PSAM). They are a function of the input mass and energy release. Dr. Kukida raised questions regarding treatment of the subcooled inventory for calculation of recirculation line liquid breaks. His studies have indicated that the Henry -
Fauski critical flow model yields good comparison with the Marviken 3 data during the first few seconds following a LOCA and that this correlation should be used in pool swell calculations.
The US Mark II owners have stated in the past that the design basis break for pool swell analysis is a steam line break. Additional discussions were held regarding mass and energy release for liquid breaks during for the week of November 2,1980 between Dr. Kukida and the NRC staff (CSB).
- 4. Steam Loads Condensation Oscillation Loads Preliminary observations of the condensation oscillation (CO) phenomenon in the JAERI tests indicate that C0 phenomenon do not show the strong periodicity nor the strong vent to vent coherence anticipated. The US Mark II owners have conservatively proposed a aeneric C0 load that does not take credit for the potential load reduction that might result from these observations. The WPPS-2 plant in the US (the only US Mark II steel containment vessel) has indicated that they are considering a plant unique load that will reflect these JAERI observations. This new load is currently under devetopment by Burns & Roe and will be discussed with the staff in February 1981. There was considerable interest in Japan in the refinement of the US Mark II Long Term C0 load specification since the Mark II containment designs in Japan are all steel.
Chugaing Load Specification Dr. Kukida presented the results of his studies to characterize the chugging pheonmenon in the JAERI tests. This consisted of characterization of the chugaing oscillations according to magnitude, wave form and
([] interval. The character of chugging was observed to change significantly with vent mass flux. These observations are consistent with those made in other Mark II large scale tests conducted in the US and Germany. The US approach towards development of a chugging load is conservative in that only the worst characteristics of the observed chugging loads are used to construct a chugging load.
. Lateral Loads j Our discussions indicated that it is difficult to use the available
. JAERI data to. confirm the proposed Mark II dynamic lateral load specification proposed by the US Mark II owners group. The difficulties arise in several areas; insufficient number of strain aauges were used, the data sampling rate was too long and the JAERI bracing configuration resulted in large induced thermal stresses. Future tests are directed towards resolution of these problem areas.
l l S. Witness of Test 3104 Three tests were to be conducted this fall to address instrumentation questions posed in our June 1980 meeting with JAERI personnel.
Prof. Robert Scanlon (NRC consultant from Princeton) and I witnessed the second of these tests,#3104. This 75mm liquid break test was a l
repeat of a previous test. Significant low frequency shaking of the control room was observed. The control room was located about 60 feet from the test facility. The events at each of the 7 vents occured together at approximately 2 seconds intervals. Small differences (i.e.10-50 milli-seconds) in the time of individual vent chugging during a gross pool chug could not be distinguished with the. ear.
- 6. Instrumentation Questions regarding the JAERI instrumentation were raised at the June meeting between JAERI personnel, USNRC consultants, the US Mark JI owners and t5i e Japanese Mark II owners. The questions dealt with the type of pressure transducers used and the type of accelerometer cables used inside the test facility. Three additional tests were to be conducted this fall to address these questions.
Early tests at JAERI utilized cavity type transducers instead of the preferred flush mounted transducers.
The reason for this is: 1) the difficulty in getting flush nounted transducers produced by US manufacturers in Japan, and 2) temperature sensitivity problems associated with flush mounted transducers.
{ The use of cavity type transducers resulted in several problems.
Trapped air in the diaphrahm and oscillations associated with the water filled column make these cavity type transducers subject to spurious readings. Several modifications were made in the JAERI instrumentation in an attempt to address the instrumentation questions.
This included: the placement of a few flush mounted transducers on the test floor, a change in the orientation of the floor cavity
- transducers from horizontal to vertical and a change in the orientation of the vent exit cavity transducers from horizontal to about a 45*
l angle. In addition, the type of floor accelerometer cables was changed to eliminate the cable noise problems. A quick look at the preliminary results of test #3103 conducted in September 1980 indicated partial success in solving the instrumentation problem.
However, it appears that additional improvements should be made in this area. In addition to the JAERI full scale tests, a separate test system was used to investigate the effect of transducer orientation and trapped air on the cavity type transducer response. -
The results of these tests conducted since June 1980 were discussed with us. These tests shed some light on the orientation and air
, trapping questions, however, questions still exist regarding the l
use of the cavity type transducers.
i
- 7. Shaker / Hammer Tests JAERI personnel plan to conduct shaker tests followed by structural
, changes in the test facility to better understand the influence of fluid structure interactions (FSI) on pressure histories recorded in the facility. Mr. Takashita discussed the shaker tests scheduled for January 1981. Prof. Scanlan was not optimistic that the shaker tests would provide significant information about the natural modes of the facility. He suggested that thought be given to hammer tests, random vibration tests and tests using small explosive calibrated sources at the vent exits. He recommended that these tests be considered in addition to the planned shaker tests. The concern is that the mass of the shaker must be large to have an affect on the large test vessel. A mass this large would probably change the natural modes of the facility.
Future modifications in the facility were discussed to harden the facility to study the influence of boundary stiffness on observed pressure readings.
Several modifications are under consideration including the placement i of concrete into the vessel to thicken the flexible end walls and the placement of steel beams on the wetwell floor to stiffen the l'
(. floor. These modifications are scheduled for the period April -
June 1981. Professor Scanlan suggested that the proposed shaker /
hammer tests be conducted both before and after the structural modifications to assess the impact of these changes on the facilities natural modes.
- 8. Mark II Containment Licensing in Japan Japan has 7 plants that utilize the Mark II Containment design. At least one plant, Tokai-2, is currently in operation. The Japanese government is allowing their plants to operate for a limited time period during which time the issue of pool dynamics is studied.
They appear to have made significant progress towards the development of licensing positions related to the pool swell loads which would occur in the first few seconds following a loss of coolant accident. 1 They questioned us in depth about the basis for the USNRC load criteria for pool swell loads provided in our report NUREG-0487 of October 1978. They are working on a similar report that is now about 75 percent complete. It does not appear that they have made much progress toward the development of licensing positions-for steam related pool dynamic loads (i.e., condensation oscillations, chugging and vent lateral loads). Thus, they are following US load development I and licensing efforts in this area with considerable interest. In contrast to the US Mark II plants, the Japanese plants have not evaluated their plants with respect to dynamic steam loads. They use static loads. In addition they do not combine loss of coolant accident (LOCA) loads with seismic loads.
9-
- 9. JAERI/USNRC Relations Mr. Shiba and his staff have continued to cooperate closely with the USNRC and with the US Mark.II owners to resolve issues related to Mark II containment pool dynamic loads. Dr. Nozawa stated that as a result of his discussions with STA during our visit, information flow related to the JAERI Mark II tests would be improved. He requested that the NRC send a letter to JAERI formally requesting documents related to the JAERI Mark II containment test program.
- 10. List of Documents Received A number of handouts related to the JAERI Mark II tests were given to us during this trip to Japan. A list of these handouts is attached. They are available through C. Anderson at X29424.
(.'
h>
j l
i
,,..w--- rm - - --r-,, ----*-.--e- - -e-, -
JAERI Trip Handouts
- JAERI nemo 9068, Test 3201
- JAERI memo 8987, Tests 0002-3102, blowdown analysis
- JAERI Annual Report
- Report on fluclear Safety Administration in Japan, August 1979
- Preliminary Characterization of Chugging in JAERI CRT, Part 1:
Categorization of Chugging
- Preliminary Characterization of C0 in JAERI CRT, Part 1 and Part 2 Present status of JAERI Full-Scale Mark II CRT
- JAERI Outline of Pool Swell Discussion Topics
- Water level probe slides
- JAERI evaluation of air mass history Pool Swell Slides
- Preliminary results of test 3103
- Dynamic Response test of Pressure Transducers
- Phase 1, 2, & 3 instrumentation modifications
- Proposed structural modification for the CRT facility (N.
h
^
Report on Trip to JAERI .
October 10-19, 1980 -
R.H. Scanlan (with C. Anderson of NRC)
OUTLINE Friday, October 10, 1980. PanAm Flight 801, JFK-Tokyo Narita.
Saturday, October 11, 1980. Tokyo Shimbashi-Dai-Ichi Hotel.
Sunday, October 12, 1980. Tokyo-Mito (JNR train).
Nato-Sannomaru Hotel (October 12, 13, 14, 15, 1980)
Monday, Oetober 13, 1980. Visit at JAERI sith Drs. Shiba, Takeshita, Kukita, Namatame, Nozawa.
Tuesday, October 14, 1980. Discussions at JAERI. Witnessing of Blowdown Test #3104 at JAERI test facility.
Wednesday, October 15, 1980. Continued discussions at JAERI.
Thursday, Octuber 16, 1980. Tokyo Shimbashi-Dai-Ichi Hotel (October 16, 17, 18, 1980). Discussions at Science and Technology Authority, Tokyo, with C JAERI and Japan Utilities personnel, Dr. M. Shiba presiding. Presen-tation and discussion by C. Anderson on HK II CO and chugging accep-tance criteria. ,
Friday, October 17, 1980. Discussions in Tokyo on CO, chug, criticisms of JAERI tests.
- e. Saturday, October 18, 1980. Free day, Tokyo.
Sunday, October 19, 1980. PanAm Flight 800'to JFK. .
~
i i
l -
ACTIVITIES REPORT The following vill briefly review the principal activities of technical nature during the October 10-19, 1980 trip.
1 Monday, October 13, 1980. Introduction to JAERI staff Tokai Research Establishment. Discussions with Kukita, Namatame, Takeshita on
- 1) pressure transducer characteristics, JAERI CRT facility; 2) pro-posed further test program and measurement modifications; 3) inter-pretation of CO data to date by Kukita; 4) pool swell; 5) impli-cations of possible containment reinforcement in Japanese plants as
- a result of CO and chugging loads.
Tuesday, October 14, 1980. Continued discussions on several of the above topics, plus: 1) pool swell; 2) air mass inventory analysis;
- 3) FSI. Evening: Witnessing of JAERI test #3104 (prepurged blow-down) in CRT facility (during side' effects of typhoon).
Wednesday, October 15, 1980 Continued discussions on the above topics, plus presentation (mainly by C. Anderson) of status of USNRC chugging, CO, and lateral load specifications. Review of USNRC viewpoint on JAERI blowdown facility test results to date. Discussion of implica-tions for Japanese plants.
Thursday, October 16, 1980 Return to Tokyo, accompanied by JAERI per-({} sonnel. Presentation (by C. Anderson) at Science and Technology -
Authority (Japan) of USNRC position on definition of CO and chugging loads. Commentary by Scanlan. Present: Representatives of Japanese MK II owners' group -- Tetsuo Horiuchi (Hitachi, Ltd.),, Toshiaki Enomoto (Tokyo Electric Power Co.) and H. Aoki (Toshiba Corp.).
Representative of STA (MITI): T. Ebine. Representatives of JAERI (K. Namatame, Y. Kukita, and M. Shiba (presiding)).
e.
Friday, October 17, 1980 Summarizing meeting. Response to questions.
General interchange. Problem review. *.
Some comments on the above are:in order:
- 1. General position' and attitude of Japanese. They are not monolithic in their approach to the JAERI blowdown tests, interpretation of them, and the ultimate use of the results. JAERI personnel are not licensing people and tend to take a more " scientific" view of the tests, leaning more toward " understanding the phenomena" than answering immediate practical questions for nuclear plant design. Japan licensing 'eople p have largely used static loads in designing existing plants (except for earthquake concerns). Thus they are slightly at a loss as to what to do, immediately, with dynamic pressures obtained as JAERI test data.
_ .- - . _ , . . - _ ,y- ,
Finally, Japan Mark II owners would, broadly speaking, like to see the whole issue of CO and chugging loads fade away. All agree, however, that they have not received open-armed coopera- '
tion from G.E., in spite of'the quite open Japanese response.
Cliff Anderson encouraged JAERI personnel to try to direct.
(and redirect) the JAERI blowdown tests to more exactly meet the needs of HK II plant evaluation.
Overall, there was the most cordial and open exchange possible between all Japanese met and the visitors Anderson and Scanlan.
- 2. JAERI test #3104 and transducer calibration. It certainly pays off on general perceptions and " feel" to witness a blowdown test. Anderson and Scanlan were the first foreigners to witness
,such a test at JAERI according to our hosts' accounts. We stayed inside the instrumentation shack due to the high typhoon winds and rain outside. Hence we missed the possible view of tank shaking that we might have had during test #3104 Nonetheless, even forewarned, we were rather impressed by 1) the general synchrony of all chugs (7 downcomers) "together" into one big
" whomp" about every two seconds or so during the chugging phase. (maybe 50 chugs) of the blowdown; and 2) the way the building shook from each chug. It was not an earthquake, but the floor got a solid jolt each time -- mostly vertical, to my sense.
Thinking back over the impressive s.ights and sounds of the chugging that night, the undersigned got to thinking about $he effectiveness of the pressure transducers during such dynamic events, where FSI may even involve rocking of the whole facility, etc. As a result, a list of comments was drawn up and passed on very informally to Namatame. In thinking back now, these reflections really con-stitute quite a serious indictment of the credibility of JAERI pressure readings to date. For not only are the previously criticized floor pressure tranducers vulnerable to accusation; Cliff and I now believe that any one of the large-pot, silicone oil-filled transducers is likely to be vulnerable to spurious readings induced by dynamic shock, which activates the transducer diaphragm against the oil-filled tube. The details of this trans-
! ducer are discussed in a draft report *by JAERI personnel that l
accompanies this.nemo. As an example, transducers on flexible side walls and downcomers are highly susceptible to shock-induced responses caused by transducer acceleration, not just
. local fluid pressure.
The implications of this may well be far-reachings as regards the interpretations made in various places to date both of timing (phasing of chugs) and magnitude of pressure traces. The best l
l l *See Cliff Anderson for a copy of this draft report.
l l
l
. . . , l 1
that can be said at the moment is that it is to be hoped that the JAERI tests will be halted until improvements in facility '
(stiffening walls and bottom), in instrumentation (better pressure transducers), and in test design (more Mark II represen-tative and repeatable tests) may be effected.
- 3. Japanese attitude toward U.S. know-how. This is still respectful, but is beginning to shade, with the realization, for example, that the JAERI blowdown facility is really potentially superior to things like G.E.'s 4T, stateside. (They are already accus-tomed to the situation in automobiles!). They are also rapidly coming off the attitude that G.E. (o'r NRC) knows best in all departments. I shared with them my feeling that, based on eventual good JAERI data, they could come up with their own
.C0 and chugging load definitions that could be more rationally based than the piecemeal job done by G.E. to date on the same items. We also explained that the role of NRC is not automa-tically to develop and endorse "best in the world" methods, but instead to 0.K. an acceptable method put forward by Mark II owners, no matter what superior one might theoretically be avail-able. The feeling does linger, nevertheless,that Japanese HK II owners would still be happy to "get the word", from any U.S. source, on how to get CO and chugging out of their hair.
O
- h. . _,
R.E. Scanlan RHS: par l
i l
i i
[
_. . -_. .-- . - - - -