Research On Gearbox Fault Diagnosis Technology Based On OOB-SSA-SVM

Marine gear box is the main component of Marine power system,with many advantages,mainly reflected in high reliability,strong transmission efficiency,very compact structure and relatively long application life,etc.Even in the limited space,the torque output is very high,so it is widely used in the Marine industry,play a huge application value.In ship design,gear box as an indispensable important driving equipment,has a large operating strength,and is operating in bad environment,easy to produce a high failure rate.Therefore,whether it has good reliability and stability will have a direct impact on the overall power system to the performance and safety of the whole ship.In this study,the gear fault of gear box is studied deeply.Combining with the existing fault diagnosis technology,a gear box fault diagnosis method based on out-of-pocket data(OOB)estimation and Sparrow algorithm(SSA)optimization support vector machine(SVM)method is proposed.The main research work is as follows:First of all,because in the collection of vibration signals,and can not directly collect the vibration signal of the gear,the acceleration sensor is generally placed in the gear box,so the original signal will be affected by the transmission path and environmental noise.Therefore,the minimum entropy deconvolution(MED)method is proposed to eliminate the influence of transmission path and environmental noise on the original signal.Secondly,the feature information of vibration signals is extracted in multiple dimensions,including time domain feature,frequency domain feature and time frequency domain feature.Dimensionality reduction and simplification of the extracted high-dimensional feature data set is carried out based on the random forest bag external estimation method,which does not affect the accuracy of subsequent fault diagnosis.When judging the specific fault types of gear faults,a classifier model based on parameter optimization is constructed,and the optimal parameters of support vector machine(penalty parameter c and kernel parameter g)are determined by different parameter optimization methods.Compared with particle swarm optimization algorithm Cuckoo optimization algorithm and Sparrow optimization algorithm,Sparrow optimization algorithm has the best optimization performance.Among the 200 test samples,only 10 test samples predicted the fault type incorrectly,and the accuracy of fault pattern recognition reached 95%.Finally,from the perspective of engineering practice,the software system of Marine gear box fault diagnosis is designed by using MATLAB programming,which provides an effective idea for fault analysis and diagnosis of internal gear box in practice.