基于MPC方法的供热系统一次侧流量实时预测

Real-Time Prediction of Primary Flow by MPC Method in Heating System

基于模型预测控制方法,使用离散的受控自回归模型建立二级网动态热传输滞后模型与热力站模型,结合机器学习算法中的多项式拟合方法对二级网模型和热力站模型中的参数进行辨识校准,并基于模型结果对未来工况条件下的热力站一次侧流量进行预测,为供热系统质量调节提供依据。使用实测数据对模型进行了验证,实际偏差在5%以下,为供热系统流量调节的工程实践提供了良好的指导。

Based on the model predictive control method, this paper uses the discrete controlled autoregressive model to establish the dynamic heat transfer delay model of the secondary network and the thermal station model. The polynomial fitting method of machine learning algorithm is applied to identify and calibrate the parameters of the secondary network model and the thermal station model. The primary flow rate of the heating station under future operating conditions is predicted based on the model results, which provides a basis for the quality-based regulation of heating system. The model is verified by measured data, and the actual deviation is less than 5%, which provides a good guide for the engineering practice of heating system flow regulation.