As the traditional ground indicator diagram method is not suitable for fault diagnosis of submersible plunger pump

this paper proposes a new fault diagnosis method. Multiple Mutation Adaptive Genetic Algorithm(MMAGA)is used to optimize the weight and threshold values of Radial Basis Function neural network. Some characteristic parameters are then extracted from the operation parameters of the submersible linear motor and oil wellhead

which can reflect the working state of pump. These characteristic parameters are taken as input vectors of the new fault diagnosis model so as to realize the fault diagnosis of submersible plunger pump. The experimental results show that: the comprehensive diagnosis accuracy of MMAGA-RBF method is high when using less training samples

and the comprehensive misjudgment rate is lower than 4% when the training sample is above 1000 groups. The generalization ability is stronger compared to that of the traditional model. The MMAGA-RBF fault diagnosis method is suitable for the working condition characteristics and can meet the requirements of the comprehensive diagnosis accuracy of submersible plunger pump.