Online monitoring of nuclear power plant (NPP) operational status is critical for ensuring safe and reliable operation. Given the complexity of NPP systems and the large number of variables to be monitored, conventional monitoring approaches which rely heavily on operators’ experience are facing significant challenges in fault detection and identification. This research focuses on developing and applying an improved reconstruction‐based contribution (RBC) analysis for NPP status monitoring and fault variables identification. Although traditional PCA‐based RBC methods are widely employed for identifying fault variables in multivariate transient deviations, there are still several limitations that compromise diagnostic accuracy: optimizing one statistic may lead to an increase in another; “dragging‐tail effect” causing nonfault variables to be identified; and statistics converging to local optima after partial fault direction reconstruction. To address these issues, this paper proposes an improved approach through the introduction of a joint target function with the weight coefficient empirically determined to achieve optimal balance, as well as a variable contribution sorting and truncation method to enhance the accuracy and reliability of fault variables identification.
Liu et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: