Distributed Model Predictive Control Of Indoor Temperature In Central Heating Systems

In northern urban areas of China,heating energy consumption accounts for 30% to60% of the total building operation energy consumption.The traditional heating system has been in the stage of rough management for a long time,which is prone to the problems of indoor overheating and high energy consumption.The development of new-generation information technology has promoted a new trend of intelligent heating with energy saving,comfort and high efficiency as the main features,which aims at user needs.By finely controlling the whole network of “source,network,station and household”,the heating system can operate greenly,safely,economically and efficiently.In the context of the new era of extensive intelligent and informative upgrading of heating system,this paper delves into the method of indoor temperature control for end-users of centralized heating system and dynamic hydraulic balance adjustment of heating pipe network inside buildings,so as to meet the heat demand of various users.The details are as follows:(1)In order to accurately predict building indoor temperature to formulate the optimal control strategy for room temperature of each user,a simplified grey box model of muti-zone building thermal response is established,and the parameter identification method of the grey box model is proposed.Firstly,thermal dynamic characteristics of the multi-zone building are analyzed,and the RC thermal model and the state space equations characterizing the model are established based on the equivalent circuit method.Secondly,using Particle Swarm Optimization(PSO)algorithm to identify the unknown parameters for the grey box model,and introducing Artificial Bee Colony(ABC)algorithm to improve the defect that PSO algorithm is easy to fall into local optimal,so as to improve the identification accuracy.Finally,the effectiveness and accuracy of the modeling and parameter identification methods are verified based on the relevant heating data of the 6#residential building in a district in Xi’an.The experimental results show that the indoor temperature prediction model obtained from the identification can stably track the actual room temperature trend under different weather conditions or when the adjacent room temperature changes,and the error is kept within ±0.5℃,showing good prediction accuracy and strong robustness.(2)Establish the hydraulic condition calculation model for the building internal heating network,providing the basis for the subsequent study of room temperature regulation and dynamic hydraulic balance optimization of heating pipe network.Develop mathematical models for the main components of the heating network inside the building,such as valves and pipes.Based on the network topology of the system,the hydraulic condition calculation model of the internal heating network of the building is constructed by combining the knowledge of network graph theory,Kirchhoff’s voltage law and current law,and the basic loop matrix method is used to solve it.The case study shows that the model makes it possible to calculate key hydraulic parameters such as flow rate and pressure drop in the heat supply network according to the relevant controllable parameters under various hydraulic conditions.(3)In order to meet the different heating demands of the end-users in the heating system and to guarantee the hydraulic stability of the system,a room temperature distributed model predictive control method is proposed.Firstly,by combining the previously established indoor temperature prediction model and the hydraulic condition calculation model,the indoor temperature distributed model prediction control system architecture and regulation strategy are designed for the central heating system with both hydraulic and thermal coupling.Then,an improved Sparrow Search Algorithm(ISSA)is proposed for rolling optimization of indoor temperature distributed model predictive control,with the objective of minimizing the total impedance of the heating system pipe network.The experimental results show that the proposed distributed model predictive control method can obtain a better control strategy than the traditional PID control method,and can respond to system changes at a faster rate with good predictability,stability and robustness.In this paper,focusing on the end-users of heating system and the heating system inside the building,the optimal control of room temperature and dynamic hydraulic balance of heating pipe network under distributed architecture is researched.Solving the problems of lag,thermal unevenness and deterioration of hydraulic stability of the heating system cased independent adjustment of users.It can provide guidance and reference for the promotion of intelligent optimal regulation and algorithm application for end-users of heating system.