Unanticipated fault diagnosis of large telescopes based on IGWO-iForest

A method for unanticipated fault diagnosis based on IGWO-iForest (Improved Grey Wolf Optimizer-Isolation Forest) is proposed to address various unpredictable problems faced by large telescopes in extreme environments. First, the random forest feature selection algorithm is used to identify the features of the original dataset and eliminate redundant features. Secondly, the differential evolution strategy is introduced into the GWO (Grey Wolf Optimizer) to improve the local search efficiency and accuracy, and the Levy flight strategy is introduced into the GWO to improve the global search ability of the algorithm. Then, the improved IGWO is used to optimize the parameters of the iForest model. Finally, the performance of the model is verified through data collected from a fault diagnosis and self-healing hardware-in-the-loop simulation platform. The experimental results show that the IGWO-iForest algorithm achieves a fault diagnosis accuracy of 99.1%, which demonstrates its higher sensitivity to a small number of unanticipated fault data compared with other anomaly detection algorithms, proving the effectiveness of this method in accurately diagnosing unanticipated faults in telescopes