基于变论域的分散模糊推理方法求解传热反问题

建立了基于变论域的分散模糊推理方法。首先，设计了一组分散的模糊推理单元，以各测点处温度的计算值和测量值之间的误差分别作为模糊推理单元的输入变量，同时根据误差的大小调整输入变量论域的值，进行模糊推理得到推理结果；然后，对各推理结果进行加权综合，产生各个待反演温度的补偿值并对各个待反演温度的猜测值进行修正；最后，实现导热反问题的求解。选用不同的初始猜测值、温度测点数目以及温度测量误差进行了数值试验，证实了基于变论域DFI算法的有效性。

Decentralized Fuzzy Inference Method Based on Variable Yniverse for Solving Inverse Heat Conduction Problem

A Decentralized Fuzzy Inference （DFI） method based on variable universe is proposed. Firstly, a group of fuzzy inference units are designed. The calculated temperature value of each measuring point and the error between the measured ones serve as the input variables of the fuzzy inference unit respectively. In addition, according to the size of error, and combining the above operations, adjustment for the value of input variable universe is implemented for fuzzy inference to obtain the inference results. Secondly,weighted synthesis for each inferenee re- sult is made so as to produce the compensation value of each wait inversion temperature, and its guess value is corrected. Eventually, the heat conduction inverse problem is solved. A numerical experiment, based on selection of various initial guess values, temperature measuring points and temperature measuring errors,is conducted for validation test, which shows an effectiveness of the DFI method based on variable universe.