摘要
提出了基于对数变换的数据预处理改进算法,测试表明效果较好。以加工面积、电极损耗比、表面粗糙度为输入参数,脉冲电流、脉冲宽度、脉冲间隙、放电间隙、伺服基准、伺服速度、加工速度为输出参数,提出了基于改进BP神经网络的电火花加工工艺选择模型。经过与实验数据的比较,该模型能真实反映机床的加工工艺规律,能实现在给定加工条件下进行电加工参数的自动选择。
An improved algorithm in neural network (NN) data pretreatment was presented. By testing, the data pretreatment method is better than before. A process selection model in electrical discharge machining (EDM) based on improved back-propagation (BP) NN is introduced. Input parameters of the model are the discharge area, the relative wear ratio and the surface roughness. Output parameters of the model are the discharge current, the pulse width, the pulse interval, the discharge interval, the servo standard, the servo velocity and the machining velocity. Based upon the experimental results, the model may really simulate the process laws in the machine, and can be applied to auto-selection of the processing parameters under a given machining demand.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2005年第18期1617-1621,共5页
China Mechanical Engineering
基金
中国工程物理研究院自然科学基金资助项目(20040322)