1 September 2010 in Computer-Aided Design
5 January 2021 in IEEE Access
Because machine learning has been widely used in various domains, interpreting internal mechanisms and predictive results of models is crucial for further applications of complex machine learning models. However, the interpretability of complex machine learning models on biased data remains a difficult problem. When the important explanatory features of concerned data are highly influenced by contaminated distributions, particularly in risk-sensitive fields, such as self-driving vehicles and healthcare, it is crucial to provide a robust interpretation of complex models for users. The interpretation of complex models is often associated with analyzing model features by measuring feature importance. Therefore, this article proposes a novel method derived from high-dimensional model representation (HDMR) to measure feature importance. The proposed method can provide robust estimation when the input features follow contaminated distributions. Moreover, the method is model-agnostic, which can enhance its ability to compare different interpretations due to its generalizability. Experimental evaluations on artificial models and machine learning models show that the proposed method is more robust than the traditional method based on HDMR.
22 February 2022 in Kerntechnik
CPR1000 severe accident management guideline (SAMG) is a product of the Westinghouse Owners Group (WOG) SAMG and the French SAMG (called GIAG) combination. It adopts the structure of WOG SAMG. However, it follows the approach of GIAG performing reactor coolant system (RCS) depressurization immediately after entering SAMG. This is a special feature of CPR1000 SAMG. The purpose of this paper is to evaluate this RCS depressurization strategy. The simulations of a station blackout (SBO) accident are performed for CPR1000 with RCS depressurization strategy adopted at both core exit temperature (CET) of 650 and 1100 °C together with opening different number of safety relief valves (SRV) using an integral severe accident analysis code. The optimized RCS depressurization strategy is provided to delay the reactor pressure vessel (RPV) failure time. And some comments are provided for future improvement.