| Gearbox which is a constant transmission ratio, high transmission efficiency,transmission ratio range of driving parts, is widely used in various mechanical devices.However, because of the structure of complex and harsh working environment factors it is sovulnerable to damage and failure that affecting the normal operation of the whole equipment.There is an important practically significance to increase the fault diagnosis of gearbox, tounderstand running state about equipment, timely to detect abnormal.Containing a lot of information of the gearbox vibration signal, a key factor is can extractsignal containing of fault features from the noisy vibration signal. Use independentcomponent analysis as a signal preprocessing step to get mechanical state source signal fromthe many mixed signals, and set an ARMA model based on “pure” signal to improve thevalidity of fault diagnosis from model’s feature.Firstly, I do analysis common failure mode of gearbox in my research topic, especiallythe gear and bearing failure. Secondly, I combine both independent component analysisalgorithm and quantum behaved particle swarm optimization algorithm based on studying ofthe theoretical. By selecting a certain objective function, I use the quantum behaved particleswarm algorithm as the optimization algorithm of independent component analysis to formQPSOICA algorithm. Thirdly, in JZQ250gearbox’s experimental object, build test systemand acquainted vibration signal. The independent component analysis algorithm is applied tofault diagnosis of gearbox, and some results are achieved. Finally, develop a gearbox faultfeature extracting software based on ICA algorithm on the MATLAB software platform,which consists of acquisition data module, ICA algorithm simulation module and extractinggearbox fault diagnosis feature module. The system not only simulates the real-time acquisition data, but also succeeds using ICA algorithm and ARMA model the fault diagnosisof gearbox. |
PDF Full Text Request