Fault diagnosis of free-conducting particles within GIL based on vibration signals

Accurate quantitative diagnosis of free-conducting particle faults plays an important role in improving the reliability gas insulated line (GIL) system. However, existing fault methods cannot accurately identify with different quantities and sizes. Motivated by this, this paper proposes a novel method based on vibration signals, which integrates variational mode decomposition (VMD), self-adapting whale optimization algorithm-multiscale permutation entropy (SAWOA-MPE), deep forest (DF). First, raw signals are decomposed via VMD, reconstructed correlation degree. Afterwards, SAWOA is employed to optimize critical parameters MPE, optimized MPE further utilized extract features signals. Finally, extracted feature vectors trained tested construct valid DF classification model that identifies faults. The experimental results indicate identification accuracy proposed can reach 99.5%. Moreover, comparative tests various vector extraction models validate superiority method.