基于PSO-LSTM的碳纤维增强聚合物铣削能耗预测  

Energy Consumption Prediction for Carbon Fiber Reinforced Polymer Milling Based on PSO-LSTM

作　　者：张华[1,2] 张福童 鄢威[1,3] 江志刚 ZHANG Hua;ZHANG Futong;YAN Wei;JIANG Zhigang(Academy of Green Manufacturing Engineering,Wuhan University of Science and Technology,Wuhan 430081,China;Hubei Key Laboratory of Mechanical Transmission and Manufacturing Engineering,Wuhan University of Science and Technology,Wuhan 430081,China;School of Automatic and Traffic Engineering,Wuhan University of Science and Technology,Wuhan 430081,China)

机构地区：[1]武汉科技大学绿色制造工程研究院,武汉430081 [2]武汉科技大学机械传动与制造工程湖北省重点实验室,武汉430081 [3]武汉科技大学汽车与交通工程学院,武汉430081

出　　处：《组合机床与自动化加工技术》2025年第4期52-57,65,共7页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目(52375508,51975432,52075396);武汉科技大学“十四五”湖北省优势特色学科(群)项目(2023B0405)。

摘　　要：碳纤维增强聚合物(CFRP)应用日益广泛,但由于材料的各向异性、高刚度、高比强度等特性使得其二次加工能耗影响因素较多,难以实现准确的预测。针对上述问题,提出了一种基于PSO-LSTM的CFRP铣削能耗预测方法。首先,基于正交试验方法设计了在不同进给量和切削速度下纤维取向为0°、30°、45°、75°的CFRP铣削实验,分析了不同加工参数对能耗的影响;构建了基于粒子群优化长短期记忆神经网络(PSO-LSTM)的预测模型,实现了CFRP铣削能耗的准确预测;最后,对所提模型和方法进行验证,相比于其他算法,所提方法通过性能评价指标MAE为1.972 82、R^(2)为0.998 72,RMSE=2.594 12,验证了该模型的有效性和优越性。Carbon fiber reinforced polymer(CFRP)is increasingly widely used.However,due to the anisotropy,high stiffness,high specific strength and other characteristics of the material,there are many factors that affect its secondary processing energy consumption,making it difficult to accurately predict.In response to the above problems,a CFRP milling energy consumption prediction method based on PSO-LSTM was proposed.First,based on the orthogonal experimental method,CFRP milling experiments with fiber orientations of 0°,30°,45°,and 75°under different feed amounts and cutting speeds were designed,and the impact of different processing parameters on energy consumption was analyzed.A prediction model based on particle swarm optimization long short-term memory neural network(PSO-LSTM)was constructed to achieve accurate prediction of CFRP milling energy consumption.Finally,the proposed model and method were verified.Compared with other algorithms,the proposed method passed the performance evaluation index MAE of 1.97282,R^(2) of 0.99872,and RMSE=2.59412,which verified the effectiveness and superiority of the model.

关 键 词：CFRP 铣削 PS0-LSTM 能耗预测 

分 类 号：TH161[机械工程—机械制造及自动化] TG54[金属学及工艺—金属切削加工及机床]