A method based on Hilbert marginal spectrum and sparse autoencoder(SAE)-deep neural networks(DNN)is proposed to recognize partial discharge(PD)types. Firstly

PD signals are dealt with variational mode decomposition(VMD)

and these obtained modes are used to construct corresponding Hilbert marginal spectrum by Hilbert transformation. Secondly

a Hilbert marginal spectrum of PD signal is taken as an input vector

and SAE can learn the inherent features and extract the succinct expressions from original data automatically. Thirdly

the results obtained by SAE are used to initialize DNN which is trained by a large number of samples. In the meanwhile

in order to speed up the convergence in the processes of learning for SAE and DNN

the network is optimized with the adaptive-step learning rate and updated with the weight parameters. Finally

DNN is trained well to identify the PD types of samples. Besides

compared with the results based on BP neural networks and support vector machines

the results based on SAE-DNN can achieve a higher accuracy.