核能電廠結構元件地震機率式風險評估及易損性分析方法研究

Seismic Probabilistic Risk Assessment and Fragility Analysis Approach for Structural Components in Nuclear Power Plants

楊淳任(Chun-Jen Yang)

指導教授 ： 黃尹男

國立臺灣大學/工學院/土木工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202003469

摘要

地震機率式風險評估 (Seismic Probabilistic Risk Assessment，SPRA) 的主要目的為估算因地震事件引致的核能電廠事故發生頻率，如爐心受損、輻射外洩等，其流程主要可概括為三個部分，包含地震危害度分析、結構與設備元件易損性分析、以及事故序列分析，並整合以上結果計算風險。本研究著重於結構元件之易損性分析方法的探討，並進一步推導不同的易損性分析結果對於總體失效事件風險所造成的影響。
研究中以案例核能電廠輔助廠房之鋼筋混凝土剪力牆為案例結構元件，分別以傳統方法 (分離變數法)、以及由非線性歷時分析結果建立易損性曲線的三種統計方法 (線性迴歸法、最大似然估計法、增量動力分析法) 對其進行易損性分析。所得結果顯示以非線性歷時分析方法建立之易損性曲線在對數標準差上相比傳統分離變數法明顯較小，為探討箇中緣由，本研究挑選了數個可能的影響因子，包含：1) 水平向最大反應之隨機性、2) 變異性參數取樣方法、以及3) 破壞層間變位角之變異性，觀察在其餘條件維持不變的前提下個別改變該些因子之設定對於非線性歷時分析方法求得之易損性曲線中位數與對樹標準差的影響，並與分離變數法進行比較。
本研究進一步以 1) 傳統SPRA將危害度與易損性曲線連續積分的方法、以及 2) 新型態SPRA以蒙地卡羅模擬法決定元件損壞狀態後，將失效機率與危害度曲線進行離散迴積的兩種方法，針對不同的易損性分析結果計算其風險，探討兩種風險評估方法對於相同條件下所得之風險計算結果的差異，以及不同易損性分析之影響因子對於最終失效事件風險所造成的影響。

關鍵字

地震機率式風險評估 ； 易損性分析方法 ； 非線性歷時分析 ； 拉丁超立方取樣法 ； 蒙地卡羅模擬法

並列摘要

The main purpose of seismic probabilistic risk assessment (SPRA) is to determine the annual frequency of unacceptable performance in nuclear power plants (NPPs), such as core melt and release of radiation. The risk-assessment procedure involves the establishment of accident sequences and integration of plant fragility data and seismic hazard curve. The study focuses on methodologies for developing fragility curves and risk- assessment.
The fragility analysis of a sample auxiliary building studied in the paper was performed using four different methodologies, including conventional separation of variable method (SOV), and three statistical methods using data from nonlinear time-history analysis: regression analysis method (LR), maximum likelihood estimation method (MLE), and incremental dynamic analysis method (IDA). The results show that the logarithmic standard deviation from SOV is much larger than those from the other methods. To gain more insight into this observation, this research studied the impact of three parameters, namely, randomness of horizontal direction peak response (HDPR), the variance of damaged drift ratio (DD), and latin hypercube sampling (LHS), on the results of fragility analysis, and found that the first two parameters have significant impact on the median and logarithmic standard deviation of fragility curves, respectively.
In order to identify the influence of HDPR and DD on the results of risk-assessment, SPRA procedures were performed using the above-mentioned fragility curves. Besides, the conventional SPRA and a new procedure for SPRA which involves nonlinear time-history analysis and Monte Carlo procedure were both conducted in the study to determine the difference between the results of the two risk-assessment methods.

並列關鍵字

Seismic probabilistic risk assessment ； Fragility analysis method ； Nonlinear time-history analysis ； Latin hypercube sampling ； Monte Carlo procedures

參考文獻

ACI (American Concrete Institute). (2008). ACI 318-08: Building code requirements for structural concrete.

American Society of Civil Engineers. (2014). Seismic analysis of safety-related nuclear structures and commentary. American Society of Civil Engineers.

ANSI, (2008). Probabilistic Seismic Hazard Analysis. American National Standards Institute, American Nuclear Society. ANSI/ANS-2.29-2008, 33.

ATC‐58. (2009). Guidelines for seismic performance assessment of buildings.

Barda, F., Hanson, J. M., Corley, W. G. (1976). Shear Strcngtli of Low-Rise Walls wit11 Boundary Elements. In ACI Symposium, Reinforced Concrete Structures in Scistnic Zones, SP-53, Detroit, Michigan.