锻压技术

基于PSOBP的超声滚挤压轴承套圈表面加工硬化程度预测

英文标题：Prediction on degree of work hardening for surface of bearing ring by ultrasonic rolling extrusion based on PSO-BP

作者：朱其萍徐红玉王晓强刘志飞刘东亚

分类号：TG376.1

出版年,卷(期):页码：2021,46(11):190-196

摘要：

为了提高轴承套圈表面质量、延长轴承使用寿命，对超声滚挤压加工参数与轴承套圈表面加工硬化程度之间的影响规律进行分析。提出采用PSOBP神经网络模型进行预测，建立以加工过程中4个主要参数为输入、加工硬化程度为输出的神经网络模型，采用粒子群优化算法对BP神经网络模型的权值和阈值进行优化，并对该模型进行了验证。结果表明：采用PSO算法优化的BP神经网络模型可有效地避免网络陷入局部最优的问题，具有更好的泛化能力，预测精度高，预测相对误差在0.5%以内，预测平均绝对百分比误差降低了0.378%。

In order to improve the surface quality of bearing ring and prolong the service life of bearing, the influence laws of ultrasonic rolling extrusion parameters on the degree of working hardening for the surface of bearing ring were analyzed, and PSO-BP neural network model was proposed to make prediction to establish a neural network model that takes four main parameters in the processing as input and the degree of work hardening as output. Then, weights and thresholds of BP neural network model were optimized by the particle swarm optimization algorithm, and the model was verified. The results show that the BP neural network model optimized by PSO algorithm can effectively avoid the network falling into the local optimal problem, which has better generalization ability and high prediction accuracy. It is shown that the relative error of the prediction is within 0.5%, and the average absolute percentage error of the prediction is reduced by 0.378%

基金项目：

国家自然科学基金资助项目(U1804145)；国家重点研究专项（2018YFB2000405）

作者简介：朱其萍（1995-），女，硕士研究生，E-mail：zhuqiping1995@163.com；通信作者：徐红玉（1972-），男，博士，教授，E-mail：xuhongyu@haust.edu.cn

参考文献：

[1]王晓强, 徐少可,崔凤奎,等.轴承套圈表面超声滚挤压加工硬化模型[J].塑性工程学报,2019,26(3):231-237.

Wang X Q, Xu S K, Cui F K, et al. Ultrasonic roller extrusion hardening model for bearing ring surface [J]. Journal of Plasticity Engineering,2019, 26(3):231-237.

[2]叶寒, 赖刘生, 李骏, 等. 超声滚压强化7075铝合金工件表面性能的研究[J]. 表面技术, 2018，47(2):8-13.

Ye H, Lai L S, Li J, et al. Surface properties of 7075 aluminum alloy workpieces after ultrasonic burnishing processing [J]. Surface Technology, 2018，47(2):8-13.

[3]Li G, Qu S G, Guan S, et al. Study on the tensile and fatigue properties of the heattreated HIP Ti6Al4V alloy after ultrasonic surface rolling treatment[J]. Surface and Coatings Technology，2019,379：124971.

[4]郑建新, 任元超. 7050铝合金二维超声滚压加工表面完整性综合评价[J]. 中国机械工程, 2018，29(13)：1622-1626.

Zheng J X, Ren Y C. Comprehensive assessment of surface integrity in two dimensional ultrasonic rolling 7050 aluminum alloys[J].China Mechanical Engineering,2018，29(13)：1622-1626.

[5]杨兴宽, 刘颖鑫, 武小鹏, 等. 机车车轮复合超声滚压表面强化工艺研究[J]. 铁道技术监督, 2018, 46 (8):40-43.

Yang X K, Liu Y X, Wu X P, et al. Study of ultrasonic surface rolling extrusion process for locomotive wheels[J]. Railway Quality Control, 2018, 46 (8):40-43.

[6]秦毅男, 廖晓辉,赵庆治.一种基于粒子群优化算法的神经网络训练方法[J].河南师范大学学报：自然科学版, 2007, 35(3):169-171.

Qin Y N, Liao X H, Zhao Q Z. A neural network training algorithm based on particle swarm optimization[J]. Journal of Henan Normal University：Natural Science Edition, 2007, 35(3):169-171.

[7]Lahiri S K,Ghanta K C.Development of a hybrid artificial neural network and genetic algorithm model for regime identification of slurry transport in pipelines[J]. Chemical Product and Process Modeling, 2009,4 (1):1-32.

[8]沈艳, 郭兵,古天祥.粒子群优化算法及其与遗传算法的比较[J].电子科技大学学报,2005,34(5):696-699.

Shen Y, Guo B, Gu T X. Particle swarm optimization algorithm and comparison with genetic algorithm[J]. Journal of University of Electronic Science and Technology of China,2005,34(5):696-699.

[9]Rumelhart D E, Mcclelland J L. Parallel Distributed Processing Explorations in the Microstructure of Cognition:Volume 1, Foundations of Research[M].Cambridge (Massachusetts):MIT Press, 1986.

[10]Ding X L, Guo Y B, Liu T, et al. New fault diagnostic strategies for refrigerant charge fault in a VRF system using hybrid machine learning method[J]. Journal of Building Engineering,2021,33:101577.

[11]Piotrowski A P, Napiorkowski J J, Piotrowska. A E. Population size in particle swarm optimization[J]. Swarm and Evolutionary Computation,2020,58:100718.

[12]洪文鹏, 陈重.基于自适应粒子群优化BP神经网络的氨法烟气脱硫效率预测[J].动力工程学报,2013,33(4):290-295.

Hong W P, Chen Z. Efficiency prediction of ammonia flue gas desulfurization based on adaptive PSOBP model[J]. Journal of Chinese Society of Power Engineering, 2013,33(4):290-295.

[13]张德丰. MATLAB R2017a 人工智能算法[M]. 北京：电子工业出版社，2018.

Zhang D F. MATLAB R2017a Artificial Intelligence Algorithm[M]. BeiJing: Publishing House of Electronics Industry, 2018.

[14]王建国, 孟娜,殷鑫.基于粒子群优化算法的凝汽器真空预测模型[J].动力工程学报,2012,32(10):815-819.

Wang J G, Meng N, Yin X. Prediction model of condenser vacuum degree based on particle swarm optimization algorithm[J]. Journal of Chinese Society of Power Engineering, 2012,32(10):815-819.

[15]杨文宇, 谢应明,闫坤,等.基于灰色关联BP神经网络的压缩式蓄冷系统中的水合物生成量预测[J].化工进展，2021，40（2）：667-670.

Yang W Y, Xie Y M, Yan K, et al. Prediction of hydrate production in compressive cold storage system based on grey relational BP neural network[J]. Chemical Industry and Engineering Progress，2021，40（2）：667-670.

[16]Boostani M, Karimi H, Azizi S. Heat transfer to oilwater flow in horizontal and inclined pipes: Experimental investigation and ANN modeling[J].International Journal of Thermal Sciences,2017,111:340-350.

服务与反馈：

【文章下载】【加入收藏】