Fault Diagnosis For AMT Based On Experimental Data Mining

AMT is a kind of mechanical-electrical-hydraulic system integrated with thecharacteristics of non-linear, multi-parameter and time-varying. The control processof AMT is complex due to numbers of independence control parameters.Consequently we have to conduct lots of experiments to enhance the reliability andstability of the AMT system. Massive experimental data will be collected by means ofthe long-term tests. Because of the limited to analyze or excavate potential andvaluable rules for the traditional data analysis and management techniques, it issignificant to build up a professional data analysis system to analyze the faultsautomatically and provide the basis to optimize the performance of AMT.Based on the design criteria of Microsoft Access2010and LabVIEW2012development platform, this paper establishes the fault diagnosis system for AMT,which is responsible for deeply analyzing data for fault diagnosis of AMT with thedata mining methodThe fault diagnosis system can automatically analyze the collected data,diagnose the faults and analyze the failure reasons in the processes of vehicle shiftingand braking. By mining the key data to diagnose the failures, the efficiency of faulttreatment is improved. The software which includes data uncompressed module, dataextraction module and data analysis module is built up to analyze data packages. Thedata uncompressed module can calculate the actual values of the sensors and actuators,and then return the actual operation situations of vehicle. The data extraction modulecan extract the characteristic data of vehicle shift efficiently. The data analysis modulecan be used to diagnose the faults of vehicle shift and analyze the failure reasons. Thedisplay module shows analysis results with different faults as well as key data infigure, and also provides summary, statistics and screening functions. In addition, theconfigure module makes it possible to deal with various types of AMT data.Finally, the data analysis system is verified by a lot of data analysis tests, and theresults show that it works stability and reliably to meet the development requirementsof AMT system.