A Reliability Accelerated Degradation Model for Turbine Wheel of Super charger Based on Neural Network Nonlinear Fitting Method

This paper establishes a supercharger turbo wheel reliability accelerated degradation model which is based on neural network nonlinear fitting. First of all, by the analysis of 120 hours supercharger structure test and 350 hours engine whole machine reliability test, the limit of durability of the material is determined as degradation degree and temperature and rotational speed are determined as sensitive stresses. Then, according to the method of neural network nonlinear fitting, the K418 material durable performance data are supplemented. Furthermore, from the above information, we propose two turbine wheel reliability accelerated degradation models, which include 120 hours supercharger structure test model and 350 hours engine whole machine reliability test model. Finally, the method of turbo wheel residual lasting life assessment is proposed with the help of the proposed models, and a conclusion that the 120 hours structural assessment test of the supercharger meets the 350 hours whole machine reliability assessment test is obtained.