基于径向基函数网络的热物性参数辨识

Pyro-physical parameter identification based on radial basis function neural networks

将径向基函数网络应用于热物性参数辨识 ,提出了完整的数学模型 ;数值模拟结果表明 ,此法具有相当的精确性 ,成功地克服了反问题中误差累积放大的弱点 ;结果很容易推广到多维或多热物性参数辨识的情形。

Radial basis function (RBF) neural networks method was applied to the pyro-physical parameter identification and a valid mathematic model was presented. The numerical simulation results show that the problems of error accumulating and amplifying in inverse problem can be overcome in this method and very accurate results can be obtained. The method can be extended to other parameter identification problems.