A Multi-Task Causal Knowledge Fault Diagnosis Method for PMSM-ITSF Based on Meta-Learning

In the process of diagnosing the inter-turn short circuit fault of the joint permanent magnet synchronous motor of an industrial robot, due to the small and sparse fault sample data, it is easy to misdiagnose, and it is difficult to quickly and accurately evaluate the fault degree, lock the fault location, and track the fault causes. A multi-task causal knowledge fault diagnosis method for inter-turn short circuits of permanent magnet synchronous motors based on meta-learning is proposed. Firstly, the variation of parameters under the motor's inter-turn short circuit fault is thoroughly investigated, and the fault characteristic quantity is selected. Comprehensive simulations are conducted using Simulink, Simplorer, and Maxwell to generate data under different inter-turn short circuit fault states; meanwhile, the sample data are accurately labeled. Secondly, the sample data are introduced into the learning network for training, and the multi-task synchronous diagnosis of the fault degree and position of the short circuit between turns is realized. Finally, the Neo4j database based on causality knowledge of motor inter-turn short circuit fault is constructed. Experiments show that this method can diagnose the fault location, fault degree, and fault cause of the motor with different voltage unbalanced degrees. The diagnosis accuracy of fault degree is 99.75 ± 0.25%, and the diagnosis accuracy of fault location and fault degree is 99.45 ± 0.21%.