Application Research Of Intellingent Fault Diagnosis Method For Marine Rolling Bearing

Rolling bearing of marine is an indispensable part of basic rolling part on warships,the fault of rolling bearings will lead to safety accidents on marine and affect normal tasks.Therefore,the fault diagnosis technology of marine rolling bearings is particularly important.In sight of engineering practical application,machine learning intelligent methods,decision tree and random forest,is been used to diagnose the single fault and multiple fault combinations of marine rolling bearings and an intelligent fault diagnosis application system for marine rolling bearings is been developed in this paper.The main research contents are as follows:Programed by Python,the simulation of decision tree and random forest using rolling bearing fault vibration test data published by DataCastle is been done.The comparison of simulation result and data from DataCastle proves the validity of decision tree and random forest in rolling bearing fault diagnosis.A platform of Marine rolling bearings vibration testing system with eight single faults and its combination is been built through vibration system,motor speed control system and data acquisition system in order to simulate the real faulty state.And the implementation scheme of the experiment has been worked out.Based on experiments results,the training of decision tree and random forest algorithmic model is been done and five-level binary decision rree classification model tree is been built.Through diagnostic results,when selecting the margin index,impulse index,peak index and kurtosis index of radial vibration data as characteristics,the highest fault diagnosis accuracy of decision tree has reached to 98.37%and when selecting axial vibration data the highest fault diagnosis accuracy of random forest algorithmic model has reached to 98.85%.Finally,according to the diagnosis conclusion of decision tree and random forest method of marine rolling bearings,an application of graphical user interface(GUI)of marine rolling bearings fault intelligent diagnosis system is been designed.The functional requirements and architecture requirements are been analyzed.Use two diagnostic pages,two diagnostic methods,three diagnostic directions and five diagnostic feature index selective architectures to realize the application significance of automatic diagnosis and manual diagnosis test.The rationality and practicability of the application program design have been verified.