基于神经网络与PSO算法的发动机装配工艺参数优化

Optimization of engine assembly process parameters based on neural network and PSO algorithm

针对汽车发动机装配过程中缸体泄漏问题,结合Back Propagation(BP)神经网络及粒子群优化(Particle Swarm Optimization,PSO)算法,提出了一种发动机装配工艺参数优化方法。首先,使用BP神经网络建立了生产工艺参数与质量指标之间的非线性映射关系,并以此作为泄漏率预测模型。其次,根据实际生产需求,应用皮尔逊相关性分析法求解得到相关性最强的部分工位工艺参数,并以其作为后续优化对象。最后,以BP神经网络预测模型作为适应度函数,使用粒子群优化算法求解得到工艺参数的最优值。使用400台发动机的实际生产数据进行试验。试验结果显示,BP神经网络具有较准确的预测效果,结合粒子群优化算法得到了优化后的工艺参数值,显著降低了发动机的泄漏率,具有一定的指导意义。

Aiming at the problem of cylinder block leakage in the assembly process of automobile engine,combined with Back Propagation(BP)neural network and Particle Swarm Optimization(PSO)algorithm,an optimization method of engine assembly process parameters was proposed.Firstly,the BP neural network was used to establish the non-linear mapping relationship between the production process parameters and the quality indicators,which was used as the leakage rate prediction model.Then according to the actual production needs,the Pearson correlation analysis was used to obtain the process parameters of the most relevant part of the station as the subsequent optimization objects.Finally,the BP neural network prediction model was used as the fitness function,and the PSO algorithm was used to obtain the optimal value of the process parameters.The actual production data of 400 engines was used for experiments.The experimental results show that the BP neural network has a more accurate prediction effect.Combined with the PSO algorithm,the optimized process parameters were obtained,which significantly reduced the engine leakage rate and has certain guiding significance.