神经网络预测控制在中央空调水系统控制中的应用

摘要

In accordance with the features of aqueous system of central air condition, e. g. nonlinearity, large time delay and large inertia, we analyze the existing control methods which exist in the central air conditioning aqueous system control. This paper integrates the neural network and the predictive control directly in the non-linear big detention central air conditioning control system. The disadvantages of pure PID for objects with large time delay and large inertia, i. e. poor control quality, weak capability of anti _ interference are overcome by this method. The experiments indicated that this algorithm which has the fine control performance is simple and advantageous to the project realization.%针对中央空调水系统的非线性、大迟延、大惯性等特点,分析了现有的控制方法在中央空调水系统控制中存在的问题;文章将神经网络与预测控制相结合直接应用于非线性大延迟的中央空调控制系统.该控制方法克服了单纯PID控制对大迟延大惯性对象调节品质差、抗干扰性弱的缺点,神经网络预测控制还有效地补偿了传统预测控制基于线性模型的局限性;工程实例表明,该算法具有优良的控制性能,有利于工程实现.