基于二阶粒子群算法优化的神经网络再制造工件疲劳寿命预测

再制造工件多元异质材料特性及工艺参数对疲劳寿命的影响,使得传统的疲劳寿命计算方法无法适用于再制造工件,针对此问题建立了再制造工件疲劳损伤预测修正模型,并通过疲劳试验分析了不同熔覆厚度和宽度条件下对试件疲劳强度和可靠性寿命的影响,同时获取了寿命预测修正系数;进而采用二阶粒子群算法优化的反向传播(back propagation,BP)神经网络,构建了材料性能参数、应力水平及再制造工艺影响因素与疲劳寿命之间的关系模型,针对再制造工件进行寿命预测。结果表明,神经网络的预测结果与试验数据相符,优于数值计算预测模型,为实现再制造工件的疲劳寿命预测提供了一种新的方法和手段。

Fatigue Life Prediction of Neural Network Remanufactured Based on Second-Order Particle Swarm Optimization

: Aiming at the problem that the traditional fatigue life calculation method can not be applied anymore due to the influence of the multivariate heterogeneous material characteristics and process parameters on the fatigue life of the remanufactured workpiece. This paper establishes a fatigue damage prediction correction model, analyzes the effects of different cladding thickness and width on the fatigue strength and reliability life of the specimen, and obtains the life prediction correction coefficient. Furthermore, BP neural network optimized by second-order particle swarm optimization algorithm is used to establish a model to predict the life of remanufactured workpieces, this model contains material performance parameters, stress levels and the relationship between the factors affecting the remanufacturing process and the fatigue life.The results show that the prediction result of neural network is better than that of numerical calculation model, which provides a new method and means for fatigue life prediction.