호텔 객실 시뮬레이션 모델에 대한 세 가지 보정 방법의 비교

The dynamic simulation model requires detailed inputs such as heat conductivity of each material, operating schedule of lights/people/ equipment/HVAC, performance curves of mechanical equipment, control logic, etc. In addition, there are uncertain parameters that cause the gap between simulation prediction and measurement. Therefore, calibration of the simulation model is required. In this paper, three calibration approaches are discussed to develop a simulation model for a hotel room. [1] a simple manual calibration based on an expert’s intuition and knowledge. [2] Genetic Algorithm-based calibration. [3] expertise + GA based calibration. The first approach delivers an accurate calibrated model with far less time and effort compared to the second and third approaches. The second approach requires considerable computation time, but fails to produce a reliable simulation model. For the second approach to be successful, the lower and upper bounds of uncertain inputs as well as kinds of inputs to be estimated must be carefully determined. The model out of the third approach performs best since it combines the optimization technique with expertise and knowledge of an simulation expert.