Prediction of Cylinder Fatigue Lifetime with Kernel Density Estimation Theory

摘要

As a part of power transfer process, cylinder plays an important role in pneumatic system. Its failure can cause mechanical equipment downtime suddenly and gas leak, so that production and personnel security will be in danger. Cylinder lifetime prediction has been an important topic. In this paper, an adaptive method based on Kernel Density Estimation is put forward for predicting the cylinder lifetime and getting the reliability function of cylinders. Kernel Density Estimation is a nonparametric estimation method of statistics. It can make full use of samples data without assuming distribution model. In the end, a comparison is made on the cylinder experiment between the proposed method and the common used parameter estimation method, Weibull distribution, and the results show that the proposed method has a more satisfactory performance.