人工神经网络在电解加工阴极设计中的应用研究

Application of Artificial Neural Networks to Tool Design in Electrochemical Machining

将人工神经网络分析方法应用于电解加工阴极设计中。将工件曲线和ｃｏｓθ法求得的初始阴极曲线构成的加工间隙用简单三角形单元进行网格剖分，然后以神经单元取代有限元单元和节点，用人工神经网络求解描述极间电位分布的拉普拉斯方程。采用逐层求解电位分布和将电位误差转换为位置误差的方法，实现由工件形状直接设计电解加工阴极。

Presents a method to apply the artificial neural networks(ANN) to the tool design in electrochemical machining(ECM).To design directly the tool electrode for a given workpiece contour,the domain between the given workpiece and a preliminary cathode constructed with the cos θ method is divided into simple triangles and the elements and nodes in finite element method have been replaced by neural units, and the ANN has been utilized to solve the Lapalaces equation which governs the electric field between the two electrodes in ECM.The tool electrode is obtained from converting deviation of layer voltage into contour deviation, while the layer voltage is solved one by one.