基于SA-APSO-BP的多工况汽车刹车片磨损预测OA
Wear Prediction of Automotive Brake Pads Under Multiple Conditions Based on SA APSO BP
针对复杂工况下刹车片磨损预测精度不足的问题,提出一种融合模拟退火与自适应粒子群优化的BP神经网络混合模型.制动压力、初始速度和摩擦面温度为输入,通过Piecewise混沌映射初始化种群,引入距离控制因子与Metropolis准则自适应调参,增强全局寻优能力.实验结果表明,模型预测决定系数达0.993,平均绝对百分比误差为4.1%,平均绝对误差为0.0026,性能显著优于传统模型,为制动系统智能监测与维护提供可靠支持.
To address the issue of insufficient prediction accuracy for brake pad wear under complex working conditions,a hybrid model integrating a BP neural network with simulated annealing and adap-tive particle swarm optimization was proposed.Braking pressure,initial speed,and friction surface tem-perature were taken as inputs.The population was initialized via Piecewise chaotic mapping,and a dis-tance control factor,along with the Metropolis criterion,was incorporated to adaptively adjust parame-ters,thereby enhancing global optimization capability.Experimental results show that the proposed model achieves a coefficient of determination of 0.993,a mean absolute percentage error of 4.1%,and a mean absolute error of 0.002 6,significantly outperforming traditional models.This approach provides reliable support for the intelligent monitoring and maintenance of braking systems.
郭家瑞;侯俊
湖北汽车工业学院 汽车智能制造学院,湖北 十堰 442002湖北汽车工业学院 汽车智能制造学院,湖北 十堰 442002
机械制造
刹车片磨损自适应粒子群算法模拟退火算法反向传播神经网络
brake pad wearadaptive particle swarm optimizationsimulated annealing algorithmback-propagation neural network
《湖北汽车工业学院学报》 2026 (2)
14-19,6
汽车动力传动与电子控制湖北省重点实验室开放基金(ZDK12023B10)湖北省教育厅科学技术研究计划项目(B2024069)湖北汽车工业学院博士科研启动基金(BK202223)
评论