ABSTRACT The global power industry is undergoing rapid transformation, with increasing integration of intelligent technologies and diverse energy sources. Converter stations, as critical nodes for power system monitoring and control, face growing demands for safe, reliable and efficient operation under complex and dynamic conditions. This study proposes a data‐driven hybrid framework for condition and risk assessment, integrating adaptive thresholding, a condition assessment algorithm and a multi‐factor two‐stage Weibull failure model. The adaptive thresholding method optimises alarm limits, balancing missed detections and false alarms. The condition assessment algorithm mitigates membership dilution, preserving latent warning signals from real‐time monitoring while quantitatively characterising state severity and evolution. The two‐stage Weibull model captures dynamic failure‐rate evolution, incorporating both natural ageing and maintenance effects, thereby providing accurate estimates of equivalent operating age and latent degradation. Case studies on line protection devices validate the framework, demonstrating improved assessment accuracy, enhanced sensitivity to early failures and a reliable basis for condition‐based maintenance and risk management.
Zou et al. (Thu,) studied this question.