Compact Thermal Modeling And Dynamic Thermal Management Of Smart Building

Today,building energy consumption accounts for a large proportion of total energy consumption,while building smart energy control methods can lead to high energy efficiency and significant energy savings.On the other hand,the requirements of people on the building automatic management,living environment and temperature comfort are becoming higher and higher.To solve these problems,it is important to enhance the current level of smart building control systems by establishing accurate building thermal models and dynamic,efficient energy consumption-temperature control systems conveniently.Most of the existing building performance modeling methods require the physical structure of the building and cannot be explicitly expressed.Therefore,the focus of this paper is to build a compact thermal model of buildings under the premise that the priori structure of the building is unknown,and then carry out dynamic thermal management based on the thermal model.First,an accurate method to model the building thermal behavior is proposed in this paper,which can be applied to general buildings and complicated large buildings without having to obtain the physical information such as building heating,cooling and ventilation.This new modeling approach is based on the recently proposed subspace identification method to get a controllable state space model for establishing thermal performance.The compact thermal model of the building is established by obtaining historical data such as the ambient temperature of the building,the energy consumption of the heating/cooling equipment,and the energy consumption of the other load devices in the room.To make full use of the data and avoid the underfitting and overfitting problems that system identification method generally exists,an improved numerical algorithm for subspace system identification is proposed.This approach generates multiple thermal models based on the training samples to find the optimal model order,and then builds the final model by performing conventional system identification using the generated optimal order.This paper begins with the real Building Automation System(BAS),and establishes some practical building(such as office buildings,data centers)thermal models.The experimental results show that the proposed method can find the optimal model order,and create a compact thermal model with the unknown building structure.The thermal model established by the proposed method can achieve 0.2-3% average error.Then,the dynamic management of indoor temperature is realized in this paper.Based on the previously established thermal model,the model predictive control algorithm is improved,and the indoor temperature of the building is maintained by setting the target temperature of each zone and predicting the expected value of power consumption of heating/cooling and ventilation equipment.By feedback adjustment and rolling optimization to ensure the efficient use of energy,the indoor temperatures are maintained at the constant values with high precision,at the same time,ensuring a better efficiency of energy consumption.The simulation results show that the temperature of the room is very close to the target value after one single time step of thermal management.After no more than 5 time steps,the zone temperature is basically stable at the target value.This new building dynamic thermal management method can help us achieve the indoor temperature constant and improve building comfort.