Research On Dynamic Room Temperature Prediction Methods Of Building Equivalent RC Models Based On System Identification

The energy background in our country is that total energy demand rises dramatically but energy supply is in short.There are still some energy waste phenomenons such as discarding wind power.By using the "virtual energy storage" technology of buildings,the load-side resources can be incorporated into the normal operation and scheduling of the power system,which can effectively alleviate the pressure of intermittent energy sources such as wind power generation and solar power generation being connected to the power grid.When buildings store energy during off-peak hours and release energy at the right time,the thermal comfort of indoor environment should be guaranteed.Average building room temperature is an important index reflecting both building energy level and environmental thermal comfort,as well as an important content of building demand response.Therefore,it is of great significance to predict the dynamic building room temperature efficiently.In this paper,the equivalent RC models are built based on the prior knowledge.An innovative parameter identification method of step by step is put forward.With the historical data driven,genetic algorithm is used to identify model parameters by stages.The first step is to identify the false parameters about thermal resistance and thermal capacity.The second step is to identify the true and false parameters about solar radiation.The third step is to identify the implied influence factor.The fourth step is to calculate equivalent real parameters of a building.In addition,a program module of building dynamic temperature prediction is developed.Problems of unstable model parameters and a single forecast form are solved in this paper.The method requires less conditions,but it is convenient and its result has high prediction accuracy.In the demand response of buildings,the method can not only calculate the response time of room temperature based on a well-setting temperature,but also guide the building energy supply strategy.It has positive applicable value.In the condition of intermittent heating,the first-order,second-order and third-order equivalent RC models are compared with the one-day dynamic room temperature prediction results which are compared with the simulation result of TRNSYS software respectively.It is found that:(1)The mean absolute percentage in the predicted result of short-term dynamic room temperature is from 2% to 6%.(2)In the case of large variation of room temperature,the third-order equivalent RC modelhas the highest accuracy of dynamic room temperature,followed by the second-order model,and the first-order model has the lowest prediction accuracy.