铸件曲面及其模型曲面逆向工程并行设计技术

研究目的是通过逆向工程方法实现铸件曲面及其模型曲面的并行设计。已知与铸件安装配合的曲面数据及铸件各截面厚度,确定铸件的加工变形和加工余量后,分别求出铸件曲面和模型曲面的离散有序数据作为神经网络的样本数据。建立RBF神经网络拓扑结构,利用神经网络拟合铸件曲面和模型曲面,实现两种曲面的并行设计,曲面精度高,计算速度快。通过实例证明该方法可以实现曲面的并行设计,缩短研制和设计周期,提高效率。

Parallel Surface Design of Casting and Mould Based on Reverse Engineering

The aim of this research is to realize parallel surface design of casting and its mould based on reverse engineering.At first a surface with which the casting is assembled and matched is known,the thickness of the casting section is provided and the machining deformation and allowance of the casting are decided;the serial discrete data of the surfaces of the casting and mould can be obtained and treated as sample data.Then,Radial Basis Function(RBF) neural network topology is constructed and employed to compute the numeral points of casting and mould surfaces.The reconstructed surfaces of casting and mould can be gained.The surfaces have high precision and the training rate of RBF network is fast.The actual example showed that this method can realize parallel design.The development cycle of research and design of new part is significantly reduced and efficiency is improved.