Fault diagnosis model of MMC high-frequency oscillation electromechanical equipment based on adaptive fruit fly optimisation algorithm

Electromechanical equipment plays a pivotal role in improving manufacturing efficiency and driving the national economy. However, with the increase of its usage, various failures are more frequent. Efficient diagnostic methods are necessary to enhance equipment operation and reduce time and cost. This study focuses on diagnosing faults in high-frequency oscillation electromechanical equipment, specifically in the Modular Multilevel Converter (MMC). Therefore, a novel fault diagnosis system model is proposed, combining Back Propagation Neural Network (BPNN) with Adaptive Fruit Fly Optimisation Algorithm (AFOA). This model consists of modules for information acquisition, fault monitoring and equipment control. The study utilises the access, aggregation and core layers to establish the overall structural model. Through simulation experiments, the proposed method demonstrated high localisation accuracy (>0.94) and fault diagnosis accuracy (>97%) within 60 minutes. Compared with other algorithms, it exhibits superior accuracy, stability and practical value in electromechanical equipment fault diagnosis.