호텔 객실 온도에 대한 모델 기반 제어로직 개발 및 비교

Building Performance Simulation (BPS) tools and data driven models can be applied to Model Predictive Control (MPC). For the development of the simulation model by BPS tools (e.g. EnergyPlus), a modeler’s subjective assumptions and unknown inputs are involved. In addition, a model calibration is essential to reduce the gap between model prediction and the measurement. In contrast, the data-driven model is simply based the correlation between inputs and outputs, and it requires far less information, time and effort. The authors developed two simulation models for a hotel room. One is an EnergyPlus simulation model and the other is an Artificial Neural Network(ANN) model. Then, using two simulation models, the authors developed two control algorithms: (1) Model-predictive control, (2) rule-based controls. In the paper, the performance of the aforementioned two controls are addressed.