基于BP神经网络L－M优化算法的喷液式线切割工艺参数优化

摘要

In order to solve high energy cutting often because of discharge electrode for insufficient liquid and dry cutting and wire breaking phenomenon , based in the working fluid pressure casting for liquid , adding a tracking injection auxiliary system was put for-ward.The loss of liquid jet between the nozzle and the cutting surface could be reduced or eliminated by the system , and the average flow rate of the working fluid was increased , moreover discharging condition was improved .In the spray system to build the orthogonal optimization experiment, influence of power amplifier, pulse interval, cutting thickness, pulse width, wire speed and other factors on the cutting efficiency was researched , the optimization of wire cutting technology parameters was determined , and the 200 mm2/min above cutting efficiency was obtained .By L-M optimization algorithm of BP neural network to build the wire cutting technology for net-work prediction model , prediction accuracy was high .Reliable technology parameters for the prediction model are provided for cutting of tracking type liquid spraying high speed ( HS) wire cutting machine so as to meet the needs of actual processing .%为了解决大能量切割时常因放电极间供液不足而出现干切和断丝的现象，提出在工作液高压浇注供液方式的基础上，增添跟踪喷液辅助系统。该系统可减少或消除供液喷嘴与切割表面间的喷液流失，提高工作液的极间进入量和极间平均流动速度，进而改善极间放电条件。在搭建的喷液系统上进行了正交优化试验，研究了功放管数、脉冲间隔、切割厚度、脉冲宽度、运丝速度等因素对切割效率的影响，确定了线切割优化工艺参数，获得了200 mm2／min以上切割效率。借助L－M优化算法的BP神经网络搭立了线切割加工工艺网络预测模型，预测精度较高，为跟踪喷液式高速走丝线切割机的高效切割提供了可靠的工艺参数预测模型，满足实际加工需要。