Diagnostic Method And Experimental Study Of Rotor/Shaft Typical Faults

Rotor and shaft system are important components of rotating machinery.In the course of work,the failure of rotating machinery and equipment due to the failure of rotor and shaft parts is not uncommon.Once the important mechanical equipment fails,the damage will be very serious.Therefore,timely detection and diagnosis before the failure of the rotor and shaft system is an important way to avoid or reduce losses.Aiming at the issue that most fault diagnosis methods of rotor and shaft systems adopt traditional shallow models and it is difficult to deal with big date at present,an improved fault diagnosis method based on Stacked Denoising Auto Encoder(SDAE)is proposed.And it is applied to diagnose the faults of rotor and shaft.Firstly,the SDAE model was initially verified using the rolling bearing fault data published by Case Western Reserve University,and then through simulating ten type faults of rotor and shaft on SQI mechanical fault comprehensive simulation platform,the fault signals have sampled with the data acquisition system based on LabVIEW.The Dropout mechanism is introduced to improve the performance of SDAE model,and then the Softmax classifier is used to classify and diagnose the faults of rotor and shaft.Through comparing with the traditional BP network,AE and CNN network which belongs to the depth learning model,it is found that the improved SDAE model has the highest correct identification rate for the faults of rotor and shaft.It is verified that the advantage of the improved SDAE depth model-based fault diagnosis method.