基于PSO算法改进BP神经网络的氟金云母点磨削工艺参数优化

发布时间：2018-01-01 04:22

本文关键词：基于PSO算法改进BP神经网络的氟金云母点磨削工艺参数优化　出处：《中国机械工程》2016年06期 　论文类型：期刊论文

【摘要】：通过氟金云母的高速点磨削试验,测试了加工表面硬度和表面粗糙度,分析了表面硬度和表面粗糙度随工艺参数的变化规律。基于单因素试验值和PSO算法改进的BP神经网络,利用最小二乘拟合,建立了氟金云母点磨削表面硬度和表面粗糙度关于各工艺参数的一元模型,以相关系数检验模型,证明模型具有较高可靠性。分析一元模型,提出表面硬度和表面粗糙度分别关于工艺参数的多元模型。基于正交试验值和PSO算法,对模型进行优化求解,并通过试验证明了模型具有较高可靠性。利用PSO算法对两个多元模型进行双目标优化,求解得到一组较为合理的工艺参数值。
[Abstract]:The hardness and roughness of machined surface were tested by high speed spot grinding test of fluorophlogopite. The variation of surface hardness and surface roughness with process parameters is analyzed. Based on the single factor test value and the improved BP neural network based on PSO algorithm, the least square fitting is used. The univariate model of surface hardness and surface roughness of fluorophlogite point grinding with respect to each process parameter was established. The model was verified by correlation coefficient, and the model was proved to have high reliability. The univariate model was analyzed. A multivariate model of surface hardness and surface roughness is proposed, which is based on the orthogonal test results and PSO algorithm. The PSO algorithm is used to optimize the two multivariate models and a set of reasonable process parameters is obtained.
【作者单位】：东北大学秦皇岛分校;东北大学;
【基金】：国家自然科学基金资助项目(51275083)
【分类号】：TG580.6
【正文快照】： 0引言氟金云母属于硅酸盐类人工云母晶体,化学式为KMg3(AlSi3O10)F2,具有绝缘性好、高频介质损耗低、耐高温、抗腐蚀等优良的性能,广泛应用于机械、电子、生物等领域[1-2]。较高的表面质量能够显著改善机械装备的可靠性,提高机械零件的使用寿命。传统方式通过热处理和改变材料

【二级参考文献】