Induction motors, the key equipment for rotating machinery, are prone to compound faults, such as a broken rotor bars and bearing defects. It is difficult to extract fault features and identify faults from a single signal because multiple fault features overlap and interfere with each other in a compound fault. Since current signals and vibration signals have different sensitivities to broken rotor and bearing faults, a multi-channel deep convolutional neural network (MC-DCNN) fault diagnosis model based on multi-source signals is proposed in this paper, which integrates the original signals of vibration and current of the motor. Dynamic attenuation learning rate and SELU activation function were used to improve the network hyperparameters of MC-DCNN. The dynamic attenuated learning rate can improve the stability of model training and avoid model collapse effectively. The SELU activation function can avoid the problems of gradient disappearance and gradient explosion during model iteration due to its function configuration, thereby avoiding the model falling into local optima. Experiments showed that the proposed model can effectively solve the problem of motor compound fault identification, and three comparative experiments verified that the improved method can improve the stability of model training and the accuracy of fault identification.

In view of the insufficient representation of single source signals for multi-point compound fault and the nonlinear strong coupling of vibration signals in motor transmission system, the inverse graph Fourier transform and improved multi-head convolutional neural network (IGFT-MHCNN) diagnostic model–based decoupling and denoising of multi-source vibration signals is proposed. The MHDCNN is constructed by improved multi-head convolution network and multi-label decoupling classifier, in which they extract multi-source signal features and decouple compound fault label information, respectively. Due to strong noise in compound fault and mapping capability of graphic signal processing method for complex data, a signal reconstruction method based on IGFT is established to remove residual noise from multi-source signals and enhance the main feature components. Two experimental cases show that the constructed model can effectively reduce vibration noise and achieve multi-label decoupling and identification of motor compound faults with strong coupling.