径向基函数神经网络地质曲面重建及应用

RBF Network Based Model for Reconstructing Geological Surface and Its Application

建立了一种适用于地质曲面重建的径向基函数(RBF)神经网络模型，并给出了具体算法.和一般曲面插值方法相比，该模型对原始数据的分布和边界条件无特别的要求，因而适合于遭到破坏或勘探数据较少的地质曲面的重建.RBF网络比传统的BP算法有较快的收敛速度，使得该模型有较大的实用价值.将该模型应用于某煤田，结果表明，该模型的拟合效果较好，能较好地反映煤层的实际分布情况.

A radius basis function(RBF) networkbased model su itable for reconstructing the geological surface was established, and its algorithm prese nted in detail. Compared with the common methods of interpolation, the RBF network mode l has no special requirements for original data and boundary con ditions. Therefore, it is es pecially suitable for reconstructing the geological surface which explora tion data are not enough or destroyed in some degree. Compared with the traditional BP network, the RBF network model is fast in convergence, thus is valuable in practice. The results of applicatio n of this model to the reconstructon of coalbed surface agree well with the actual situation.