径向基函数网络在虚拟加工对象重构中的应用

Application of radial basis function networks in reconstruction of virtual machining object

针对虚拟制造中无加工对象数据的加工过程仿真,提出适用于离散无规则边界测量数据的径向基函数网络方法,以重构加工对象曲面,阐述了径向基函数网络曲面重构的基本理论,分析了径向基函数网络宽度参数的确定方法;给出了径向基函数网络曲面重构函数的具体实现过程,计算了重构的误差。以实例验证了该方法的可行性,并与传统曲面拟合方法相比较,得出该方法的重构精度高,改进了传统曲面拟合在处理非均匀截面数据点时曲面形状的失真和运算的不稳定现象,从而显示了径向基函数网络重构曲面的优越性。

To deal with the non-uniform discrete measured data points for surface reconstruction in process simulation, a method based on Radial Basis Function Network （RBFN） was proposed. The basic theory of RBFN for 3D surface reconstruction was presented, and the process for calculating the surface function was realized. The method of defining parameter factor Q was discussed and the implemental process of RBFN for 3D surface reconstruction was provided. Reconstruction error was calculated. Feasibility of the method was validated by an example. The comparison between this method and the traditional method of surface reconstruction was introduced. Result showed this method could greatly improve reconstruction efficiency, and overcome the shortcomings of surface-shape losing and calculation unsteadiness.