Study On The Model Decision Trees C4.5 Algorithm For Fault Diagnosis Of Rotor's Vibration Based On Fuzzy Logic

In the paper, the experiments which conducted the rotor system fault diagnosis were presented. The data mining method and fuzzy inference system (FIS) were used in the fault diagnosis method.The vibration tests include:normal running of rotor (GOOD), rotor mass un-balance (RMUB), rotor with moving and still parts crashing and friction(MSCF) and rotor with oil whirl (RWOW) for three rotating speed (2000,4000,6000 rpm) in research room of our school,and each rotor system fault condition were test for 50 times. The features of vibration signal were extracted from descriptive statistic parameters. The C4.5 algorithm of decision tree was used to select pertinent features from data set. Decision tree was built to produce the crisp if-then rules and membership function sets. Then C4.5-FIS classification model were set up based on the crisp sets using the MATLAB Fuzzy Logic Toolbox. The accuracies of decision tree classification and fuzzy logic rules were tested by using the testing samples,then the produced fuzzy inference engines conditions were transmitted to the SIMULINK toolbox of MATLAB in order to get the final simplified model.The decision tree was built by using of the vibration signals on the rotor speed of 4000rpm, then the if-then rule and membership function sets were produced. In order to evaluate the proposed C4.5-FIS model, the data sets obtained from vibration signals of the rotor were used.Then the classification accuracy of the fault diagnosis model for the speed of 2000 rpm and 6000rpm was evaluated respectively. The experiment verifies that C4.5-FIS model could be useful for vibration fault diagnosis with good accuracy. The model was expected to be applied in the field of fault diagnosis of steam turbine generator.