径向基函数神经网络逼近空调管路温漂趋势的研究

Research on Temperature Trends of Air Conditioner Pipes Approximated by the Radial Basis Function Neural Network

针对空调管路的应变测试数据会随环境温度而产生漂移的现象,根据应变电测法,提出了基于径向基函数(RBF)神经网络拟合逼近管路真实应变值的方法,即结合正则化正交最小二乘法,逼近现实中由温度引起的漂移,间接求出空调管路较为真实的交变应变值.实验结果表明该方法是可靠的.

In light of test data of air conditioning pipes drifting with environmental temperature, based on the strain measuring method, an approach based on the RBF neural network is proposed to fit the true strain values of pipes. The method adopts the regularization orthogonal least squares method, approaches the reality drift caused by temperature change, and derives indirectly more real alternating strain value. Experimental results show that the method is reliable.