技术中心

摘要 ：针对供热过程耦合特性和节能控制的需要，提出了一种基于径向基函数（RBF）递归神经网络的供热解耦控制方法。通过典型信号响应与最小二乘结合的方法得到供热耦合系统模型，利用RBF递归神经网络进行解耦控制，消除了质调节、量调节通道间的非线性强耦合作用。仿真结果证明该方法具有良好的解耦控制特性，满足供热系统多回路控制的要求。 关键词 ：供热过程，神经网络解耦，RBF递归神经网络，嵌入维数预估 Application of RBF recursion neural network to heat supply decoupling control By Chen Lie★, Zhu Xueli, Qi Weigui and Fang Xiumu Abstract ：According to the coupling characteristics in heat supply process and the demands of energy saving control, proposes a novel heat supply decoupling method based on radial basis function (RBF) recursion neural network. By establishing the heating coupling system model with typical signal response and least?square method, applies the RBF current neural network to eliminating the strong influence between quality?adjust and quantity?adjust channels. The simulation result shows that this method has a good decoupling performance and can meet the demands of multi?loop control of heating system. Keywords ：heat supply process, neural network decoupling, RBF recursion neural network, embedding dimension estimation ★Harbin Institute of Technology, Harbin, China