Application Of Intelligent Algorithm In Prediction Of Fatigue Life

Forecasting technology is an important aspect of scientific research and it can make scientific and effective decision according to the prediction results. The development of artificial intelligence technology has driven the research of forecasting technology, and the data driven modeling intelligent forecasting method is a widespread concern. As one kind of intelligent forecasting method, support vector machine building prediction technique is proposed to overcome the shortcomings of neural network forecasting method, such as slow convergence, solution instability and poor generalization.As the specific application of forecasting technology, life prediction has become a hot spot in the research of mechanical and electrical equipment. It is based on the data of the monitoring equipment to build a life model, and according to the model results to determine the status of the device and life. This paper is based on the support vector machine intelligent modeling and prediction technology, which is applied to the practical application of the rolling bearing as the research object, and the bearing life prediction. Using support vector machine model based on historical data to predict the life of bearing, that is to extract the vibration signal of the bearing and extract effectively, can reasonably predict the remaining life of the bearing and reduce the cost of bearing fatigue life test.In the process of model construction, the kernel function type and its parameters and the parameters of support vector machine has great impact on the model prediction results,so the genetic algorithm and particle swarm optimization algorithm are used to optimize the parameters, establishing the theoretical and model foundation for the bearing regression life prediction model. In the specific application of actual model, for the effective characterization of rolling bearing state changes in the running process, the principal component analysis are used to fuse the time domain and frequency domain feature extraction and EMD sample entropy, so a bearing declined performance was bulided.Compared the index in this article with the kurtosis index, the prediction results with two kinds of algorithm to optimize the support vector machine intelligent model show that the index in this article can characterize the running state of the bearing life better. In the concrete construction of the life model, the performance index and the remaining life of the target corresponding to the previous time points are used as inputs, and the output is the remaining life of the time points. Two kinds of algorithms are used to optimize the parameters, and the prediction of the higher precision of the bearing life of the bearing is achieved. For indicators of the status of the current running state of the bearing can not be obtained,this paper use the fuzzy information granulation method to forecast the change trend of the performance index and the variation range of the bearing signal.The predicted results are reasonable compared with the original data. In addition, the prediction of the residual life of the bearing is predicted, and the results are satisfactory.