Prophase Failure Prediction of the Mechanical Transmission Systems Based on the Biological Evolutionary Algorithm
There are several types of the mechanical transmission failure, such as gear tooth broken, fatigue, pitting etc. The deterioration pattern of each failure varies according to the different environment. Furthermore, setting up the fault prediction model is quite difficult, especially at the early stage of the fault. In order to predict the prophase failure of the mechanical transmission systems depends on the condition monitoring signal, this paper researched on the biological evolutionary algorithm combined with other artificial intelligence algorithm. As the case study, the typical failure of the gearbox test-bed, for example gear tooth broken or fatigue at the test-bed was monitored by several sensors. An improved support vector machine (SVM) optimized by genetic algorithm (GA) was chosen to predict the prophase failure of the gear, due to its self-adaption and self-learning ability. The prediction results showed that it simulated the failure pattern well on the condition of a few sample data.