【摘要】 从供热管网终端泄漏诊断需要出发,引出求解正常供热时管网系统的终端温度。绕过传统的通过求解微分方程确定温度的方法,利用人工神经网络来预测管网系统的终端温度,从而为供热管网终端泄漏诊断提供参考依据。在应用人工神经网络时,采用基于相关分析法的节点删除法来优化网络结构提高网络性能。仿真分析表明基于这种结构优化的人工神经网络模型对供热管网终端温度预测较为准确,对泄漏诊断有一定帮助。更多还原

【Abstract】 Considered the diagnosis of the terminal leak of heat pipe network,the question of obtaining the terminal temperature was raised.The artificial neural network method was used to anticipate the terminal temperature of the heat pipe network,and the diagnosis can accord to the anticipation result.During the use of the artificial neural network some nodes were deleted to optimize the network based on the correlation analytical theory.The simulation analysis proved that the anticipation results of terminal temperature by this optimizing means were approximately exact and valuable.更多还原