th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
THE UPDATE OF THE SEISMIC PROBABILISTIC SITE HAZARD ASSESSMENT FOR LOVIISA NPP IN FINLAND 1 PENTTI VARPASUO
1 Fortum Nuclear Services, 00048 Fortum, Finland E-mail: [email protected]
ABSTRACT :
The Loviisa Nuclear Power Plant obtained an extended operation license for the period of 2007 -2027 in July 2007. One of the conditions set in the safety assessment document that accompanied the license extension approval was the requirement to update the seismic probabilistic risk assessment of the plant dated from 1991. This paper outlines the update of the seismic site hazard assessment of the said work. The purpose of the present work is the update of estimation of seismic hazard in the site of the Loviisa nuclear power plant in Finland. Because there are no registered strong motion acceleration recordings of earthquakes in Finland, the earthquake recordings from Saguenay and Newcastle regions from Canada and Australia were taken as sources of initial data because of their geological and tectonic similarity to Nordic countries. The code basis for the ground motion estimation in probabilistic seismic hazard studies stipulates the median spectra for mean return period of 100 000 years. The decision tree approach is used in the treatment of uncertainties in this study. The most important task in the update of the hazard analysis was the uncertainty quantification in the form of discrete distribution of the hazard at site. The actual hazard analysis carried out with the aid of Fortran Computer Program for Seismic Risk Analysis SEISRISK III developed by USGS and with the aid of the computer program EZ-FRISK developed by Riskengineering Inc. The resulting raw site hazard depicting the distribution of the seismic hazard was given in the form 32 curves equipped with appropriate weights. The end result of the analysis was given in the form of median, mean, 95% - fractile and 5% - fractile curves for horizontal peak ground acceleration amplitudes of 0.001g, 0.005g, 0.01g, 0.05g, 0.07g, 0.1g, 0.2g, 0.3g and 0.4g and for corresponding non-exceedance probabilities.
KEYWORDS: extended operation license, strong motion, probabilistic seismic hazard studies
1. Introduction
Finland is part of the Baltic shield seismic region. Baltic shield is low seismic activity area. The largest earthquake in the vicinity of the Finnish nuclear power plant sites is the magnitude 4.9 event in Estonian coast near the island Osmussaari in 1976. In the historical period about ten earthquakes with magnitude higher than 4.5 have been observed in Finland. There are four operating nuclear power plant (NPP) units in Finland: the TVO power company has two 840 MWe BWR units at the Olkiluoto site on the south-western coast of Finland and Fortum has two 500 MWe VVER 440/213 units at the Loviisa site on the southern coast of Finland. The units were commissioned between 1977 and 1982. A 1600 MW EPR (European Pressurized Water Reactor) is currently under construction at the Olkiluoto site. This paper describes the updating task of probabilistic seismic hazard assessment for Loviisa site. The probabilistic seismic hazard assessment is the first step in the seismic probabilistic safety assessment for either existing plant or for the plant in design phase.
2. Seismic design requirements applicable to Finnish NPP’s
The currently operating Finnish NPP units were built without any regulatory requirements on seismic design, and earthquake loads were not considered explicitly in the design. As a part of general probabilistic safety assessment of the existing plants seismic risk analyses were carried out in the 1990’s. These studies have revealed some seismic weak points in the plant design which have been removed with later plant modifications. th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
The seismic requirements on design of nuclear power plants valid at this time are set forth in the regulatory guide YVL 2.6 which has been accepted in 20011. The guides YVL 2.6 and YVL 2.8 also require that seismic events are included in probabilistic safety analysis (PSA)2. The design basis earthquake shall be determined so that its recurrence interval is 100 000 years on the 50 % confidence level. The design basis earthquake is characterized by a ground response spectrum. The ground response spectrum shall be presented and justified by the license holder or applicant, and it is reviewed by the safety authority STUK. The ground response spectrum for Finnish nuclear plant sites has been determined with the aid of a statistical procedure described by Varpasuo3 using an earthquake catalogue covering a period of about 600 years. The calculated horizontal peak ground acceleration is about 0.065 g at Loviisa and 0.085 g at Olkiluoto. The design PGA for new plants is taken as 0.1 g for both sites. Due to the low seismic activity, only the safe shutdown earthquake is relevant in Finland. The operating basis earthquake would be so weak that it would not result in any practical requirements for the plant systems. The requirements for design and the proof of adequate resistance against safe shutdown earthquake are stated in YVL 2.6. In addition to the deterministic seismic design requirements, a new NPP unit shall fulfill the probabilistic safety goals set forth in the Regulatory Guide YVL 2.8:
- the core damage frequency shall be less than 10-5/year and - the large release frequency shall be less than 5•10-7/year.
The “large release” is defined as 100 TBq of Cs 137 which corresponds to only about 0.015 per cent of OL3 Cs 137 inventory. This is also the release limit for severe accident management design defined in Government Decision on general regulations for the safety of nuclear power plants, VNp 395/19914.
3. Loviisa NPP seismic PSA
The seismic PSA for the Loviisa NPP was started in 1987 and submitted to the regulatory authority STUK in 1991. During the Loviisa operating license renewal process in 2007, the assessment of the seismic resistance was based on the seismic PSA from 1991. The total seismic core damage frequency was estimated to be 3.6•10-6 /year which is 4.3 per cent of the current total core damage frequency for the Loviisa NPP units. According to regulators review, the seismic PSA could be used show the plant‘s adequate seismic resistance to justify renewal of the operating license. However, the regulator required that the seismic probabilistic safety assessment is updated in the early part of the new operating license period. In particular, the regulator required the updating of the seismic hazard curves, since they were not based on the most recent available data.
4. Seismicity in Finland
The Baltic shield seismic province is the one of the quietest areas in the world. The plate movement from the North Atlantic Ridge in the NW-SE direction seems to be the major factor generating earthquakes in Finland. Other factors, such as glacial rebound, are also contributing to seismic activity in Finland. Earthquake recurrence rates are very low if compared with plate boundary regions worldwide. Scandinavian countries and Finland is an seismic region, at low earthquake recurrence rates and with small magnitudes. The earthquake catalogue for Northern Europe (FENCAT), maintained by the Institute of Seismology of the University of Helsinki, was used in this study5. The catalogue includes all documented earthquakes in Scandinavian countries, in Baltic countries, in north-western Russia and in Finland since 1375. Instrumental earthquake observations for detection seismology started in Finland in the 1920's and local short period recordings started in 1956. The seismic events in the vicinity of Finland have been predominantly instrumentally located since the mid 1960’s. The instrumental magnitudes are based on the Richter's classical local magnitude scale, ML, modified for the Nordic region. Figure 1 shows that southern Finland, which is the target area of this seismic hazard assessment updating study, is characterized by low seismicity. The most active zones of are the Swedish coast from the Bothnian Sea to the Bothnian Bay, western Lapland and the northern Bothnian Bay-Kuusamo region. th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
E
0
: n io t a t n 70 N e s e r p
e h t
f o
r e d r o B
Lapland
SWEDEN Kuusamo
65 N BB Oulu- j鋜vi
FINLAND RUSSIA Bothnian Sea
60 N
200 km
ESTONIA M = 4.5-6.1 M = 3.5-4.4 M = 2.5-3.4
M = 1.5-2.4 20 E 25 E 30 E
55 N
Figure 1. Distribution of the earthquake epicenters in northern Europe since 1375 according to FENCAT.
5. Updated seismic hazard curves for Loviisa NPP site The seismic hazard analysis described in this paper has been updated with the aid of SEISRISK III computer program6. The theoretical methods and the selection of input parameters has been described in references 7 and 8. The data on territory around Loviisa site with radius of 500 km are investigated. This source area is divided into six source zones. The number Nm of earthquakes having magnitude m or higher than m is described by the Richter equation log10 Nm = a – b*m. The curve is cut off at the maximum magnitude for the region denoted by Mmax. The parameters a and b were determined for each source zone and for the whole region. Results for the six source zones inside the 500km radius are shown in Table 1. Source zone b a arel Mmax 1. Russia 0.818* 1.515* 1.515 2.9 2. Å-P-P Zone 0.731 1.097 1.040 4.9 3. B-L Zone 0.782 1.459 1.213 4.6 4. Bothnian Bay - S. Kuusamo 0.829 1.121 0.383 4.7 5.SFQZ 1.166 2.065 1.229 3.2 6. Latvia 0.818* 1.033* 1.329 3.5 * Lack of data: a = arel. and b = brel Table 1. Seismicity parameters for the source zones of the Loviisa region. Usually parameters arel are scaled as to the seismicity parameter for the whole region in proportion to the zone surface area. Mmax = observed maximum magnitude. th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
The Dahle attenuation function form with minor modifications is adopted for the attenuation relationship used in this work. The main reason why the Dahle attenuation model in its original form was not accepted as such in this study was large standard deviation values over twice the commonly used values. Second reason was that the spectral attenuations were developed for pseudo-velocity spectrum and not for acceleration response spectrum, which was needed for the purpose of current task. Thirdly, the frequency band used in Dahle model was not sufficiently wide for the purposes of current study. It could be expected that spectral accelerations over 10 Hz would be of interest in developing the design spectra for geologic conditions where the bedrock outcrops were the expected level of grade at prospective site and the Precambrian rock had very high stiffness qualities. Because there are no registered strong motion acceleration recordings of earthquakes in Finland, the earthquake recordings from Saguenay and Newcastle regions in Canada and Australia were taken as sources of initial data because of their geological and tectonic similarity to Baltic shield area. The determination of attenuation functions in the Finnish seismic hazard analyses is described in detail papers by Varpasuo 7 and 8. Site effects are not relevant in Finland as the sites are located on solid bedrock. The logic tree structure was used to determine 32 combinations of input parameters and to describe the uncertainties involved in determining seismic parameters. For each combination of parameters in the logic tree the corresponding hazard curve was calculated. The 32 raw hazard curves are shown in Fig 2. This group of raw hazard curves was used to calculate the median hazard curve and the 5 % and 95 % confidence bounds levels are shown in Fig 3. The median ground response spectra for the Loviisa site representing the exceedance probability of 10-5/a for damping ratios of 0.5, 2, 5 and 10 % together with confidence bounds (5 %, 95 %) are shown in Fig 4.
Lov iis a raw s eis mic haz ard c urv es in PGA ; Seis ris k III analysis, December 2007
1,E-01
1,E-02
1,E-03
1,E-04
1,E-05
1,E-06
1,E-07
1,E-08 Annual frequencyAnnual of exceedance 1,E-09
1,E-10 0,0001 0,001 0,01 0,1 1 PGA amplitude (g)
L0_1l n L0_1l s L0_1tn L0_1ts L0_1Rl n L0_1Rl s L0_1Rtn L0_1Rts L0_5l n L0_5l s L0_5tn L0_5ts L0_5Rl n L0_5Rl s L0_5Rtn L0_5Rts O0_1l n O0_1l s O0_1tn O0_1t s O0_1Rl n O0_1Rl s O0_1Rt n O0_1Rt s O0_5l n O0_5l s O0_5t n O0_5t s O0_5Rl n O0_5Rl s O0_5Rt n O0_5Rt s
Figure 2. Raw hazard curves for Loviisa site. th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
Dis tribution of haz ard c urv es f or Lov iis a s ite; Seis ris k III analysis,December 2007
1, E- 0 2
1, E- 0 3
1, E- 0 4
1, E- 0 5
1, E- 0 6
1, E- 0 7
1, E- 0 8
1, E- 0 9 Annual frequencyAnnual ofexceedance 1, E- 10 0,0001 0,001 0,01 0,1 1 PGA amplitude (g)
5%fract ile median 95%fract ile
Figure 3. Median hazard curve with confidence bounds (5%, 95%) for Loviisa site.
Loviisa ground response spectra for seismic PSA for 1E-5 exc. probability
1
0,1 spec. acc. (g)
0,01
0,001 0,1 1 10 100 freq (HZ)
5%frac0_5D median0_5D 95%frac0_5D 5%frac2D median2D 95%frac2D 5%frac5D 95%frac5D 5%frac10D median10D 95%frac10D median5D
Figure 6. Ground response spectra for Loviisa site
th The 14 World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China
6. Conclusions In the updating task of the seismic hazard curves four alternative attenuation functions were used. One of them was based on longitudinal Saguenay recordings from eastern Canada; the other one was based transversal Saguenay recordings from eastern Canada; the remaining two alternatives were based on longitudinal and transversal recordings from Newcastle area from south –eastern Australia. The median level PGA with 100 000 year recurrence period was 0.048 g in the original probabilistic seismic analysis from the year 1991 and 0.056 g in the updated analysis reported in this paper. The relative increase in the hazard value is 16.6 %. In conclusion, the updated hazard curves do not differ significantly from those used in Loviisa seismic PSA from the year 1991.
REFERENCES
1 Nuclear legislation, Government Decisions on Nuclear Safety and YVL Guides (Nuclear Regulatory Guides) issued by Radiation and Nuclear Safety Authority - STUK are available on www.stuk.fi (English > regulations). 2 J. Sandberg, P. Välikangas and Y. Hytönen, Seismic Risk Analysis and Design in Finland - A Regulatory View, Proceedings of the CSNI Workshop: Specialist Meeting on the Seismic Probabilistic Safety Assessment of Nuclear Facilities, Jeju Island, Korea, 6-8 November 2006. 3 P. Varpasuo, Estimation of Seismic Hazard in the Territory of Southern Finland, Proceedings of the 8th International Conference on Structural Safety and Reliability (ICOSSAR’01). Los Angels, USA, 17-23 June 2001 4 Nuclear legislation, Government Decisions on Nuclear Safety and YVL Guides (Nuclear Regulatory Guides) issued by Radiation and Nuclear Safety Authority - STUK are available on www.stuk.fi (English > regulations). 5 Ahjos, T. And Uski, M. 1992, Earthquakes in northern Europe in 1375-1989. Tectonophysics, 207: 1-23. 6 Bender B. , Perkins David M. 1987. SEISRISK III: A Computer Program for Seismic Hazard Estimation, U. S. Geological Survey Bulletin 1772, United States Government Printing Office; Washington. 7 P. Varpasuo, Probabilistic seismic hazard assessment for OL3 plant site in Finland, Proceedings of the CSNI Workshop: Specialist Meeting on the Seismic Probabilistic Safety Assessment of Nuclear Facilities, Jeju Island, Korea, 6-8 November 2006. 8 P. Varpasuo, the Seismic Site Hazard Assessment for OL3 NPP in Finland, 18th International Conference on Structural Mechanics in Reactor Technology (SMiRT 18), Beijing, China, August 7-12, 2005, SMiRT18-KM01_2.