基于改进型Elman神经网络预测模糊神经网络控制的变风量空调设计

由于空调控制系统具有非线性、大滞后、时变性等特点，提出了一种基于改进型Elman神经网络的模糊神经网络控制算法，其预测输出与实际输出的差值作为模糊神经网络控制器的输入，使空调控制系统具有较高的控制精度和良好的动态特性和鲁棒性。仿真结果表明：与传统PID控制相比，基于Elman神经网络的模糊神经网络控制具有较强的鲁棒性，学习能力强，控制精度高，控制效果好。并具有自适应能力，应用前景十分广泛。

Research of Prediction Fuzzy Neural Network Control Based on Improved Elman Neural Network for VAV System Design

As the air-conditioning control system has the nonlinearity, large time delay and time variation and other characteristics, a method of fuzzy neural network control based on the modified elman neural network was put forward, and the input of the controller includes error of the output of prediction and the actual output, so that it has the high controlling precision, good dynamic characteristic and robustness. The simulation results showed that when comparing with the traditional PID control system, fuzzy neural network control based on the modified elman neural network has stronger robustness and adaptive ability, higher control precision, better control effect, stronger skills of learning and wider application prospect.