Fault Diagnosis of Hydroelectric Units Based on CSABO-VMD and Multi Strategy Improved BOA-BP.

In response to the lack of fault diagnosis methods for bulb tubular hydroelectric units, this study utilizes a rotor fault simulation platform to simulate typical faults within these units as the basis. This research combines the Subtraction-Average-Based Optimizer (SABO) with Chaotic Mapping, to enhance Variational Mode Decomposition (VMD) for decomposing raw data. The components obtained are then analyzed using multiscale sample entropy, with the results fed into a BP neural network optimized by a multi-strategy improved Butterfly Optimization Algorithm (BOA) for fault diagnosis and classification identification. The results demonstrate a diagnostic accuracy of 98.75%, significantly superior to non-optimized models. This research provides a valuable addition to the existing fault diagnosis methods for bulb tubular hydroelectric units. [ABSTRACT FROM AUTHOR]