Abstract:
As an important driving equipment, motor has been widely applied in industrial production. Rolling bearing is one of the most important parts of the motor. Once a fault occurs, the production process will be seriously affected. So, the diagnosis of motor bearing fault has great significance for ensuring the safe and normal production. In this paper, the probabilistic neural network (PNN) is introduced to achieve the fault classification of motor bearing. In order to deal with the shortcoming that the smoothed factor, one of the most important parameters of PNN, which usually needs to be artificially set from experience or repeated attempt, we propose an adaptive probabilistic neural network (SPNN) based on sine cosine algorithm (SCA). It is noted that PNN will get the fixed recognition results for the same training set and test set, which may reduce the generalization ability of PNN to some extent. Hence, this paper proposes an AdaBoost-based integrated learning model (ASPNN) using SPNN as a weak classifier. Moreover, the output results of the strong classifier are obtained by means of the linear combination of the output probability of SPNN. Finally, the proposed ASPNN model is applied to the fault diagnosis of motor bearing. Simulation results show that the proposed method has better performance in motor bearing fault diagnosis than PNN, SPNN and multi-classification support vector machine (MSVM) .