Energy-saving Optimization Strategy Of Building Cooling Load Based On Model Predictive Control

The central air-conditioning system of large public buildings accounts for more than 40% of the total energy consumption of buildings.The energy-saving optimization control of central air-conditioning systems is the focus of building energy conservation.Although the central airconditioning energy-saving operation in China is concerned by the majority of scholars at present,the control of building cooling load is mostly based on experience adjustment,lacking corresponding theoretical support,which easily leads to indoor thermal comfort is difficult to meet the demand and energy consumption is wasted.Therefore,improving the control strategy of building cooling load is of great significance for guiding the energy-saving optimization operation of central air-conditioning.This paper proposes a model predictive control strategy for building cooling load.By analyzing the thermal characteristics of buildings,a building thermal model based on equivalent circuit method is established.By considering the air conditioning energy consumption and building thermal comfort,the model predictive control of cooling load is designed.The effectiveness of the control strategy was verified by a large office building in Guangdong Province.In addition,the “Building Cooling Load Energy Saving Optimization Software Platform” was developed to provide a corresponding software platform for building cooling load model prediction control..The main research work of this paper includes:(1)Due to the complexity of the envelope structure of large public buildings,it is difficult to model the thermal characteristics from the mechanism.This paper simplifies the unsteady heat transfer process of the building into an equivalent RC circuit and establishes a building thermal model in the form of a third-order state equation.Then,based on the measured data,the model parameters are identified online by Recursive Least Squares method(RLS),pseudomeasurements are constructed,and the room temperature prediction values are corrected by Kalman Filtering(KF).The results show that the RLS-KF algorithm has higher prediction accuracy and stability than the single RLS algorithm.When the sampling period is 15 min,the EEP is reduced by 80.3%.(2)Large public buildings are unsteady thermal systems with large inertia,nonlinearity and time-varying.The conventional control strategy is difficult to meet the control requirements of real-time and stability.This paper proposes model predictive control of building cooling load,and tuning controller performance based on prediction step number and weight matrix.Firstly,using the state equation and weather forecast information,the multi-step prediction equations at room temperature are recursively transformed;then the cost function of thermal comfort and energy consumption is constructed,and the rolling optimization of air conditioning cooling load is realized by solving the quadratic programming problem with constraints;At the same time,parameter identification and state observation are introduced as feedback correction links.The simulation results show that compared with PID control,the model predictive control has higher temperature control accuracy and can reduce energy consumption by 8.5%.(3)In addition,in the process of acquisition,transmission and storage of the relevant operational data of the central air-conditioning system,abnormal data is inevitably generated due to sensor failure,communication interference,and network delay.In order to avoid interference to the control system,this paper proposes an online identification and repair method for abnormal data based on autoregressive model.The average recognition rate of the abnormal data is 97.2%,and the mean square error of the repair is 0.413.(4)Combining with engineering practice,a software platform is developed,which encapsulates data online preprocessing method,building thermal characteristics modeling method and model predictive control method of air-conditioning cooling load.The software provides effective data analysis and model predictive control platform for central airconditioning operator.