Automated Calibration Of Greenhouse Energy Model Using Hybrid Evolutionary Programming (EP)-Energy Plus

Abstract

This paper presents an optimization approach of calibrating a tomato greenhouse energy model using hybrid Evolutionary Programming (EP)-EnergyPlus. The proposed methodology applies automated simulation-based approach by coupling Matlab and EnergyPlus to perform building energy simulation and obtain the best variables configuration with minimal error between the simulated and measured energy of the greenhouse. The proposed method is tested using a tomato greenhouse system located in Universiti Putra Malaysia (UPM). The greenhouse envelope is built using 0.15mm thick Transparency Plastic Film. Meanwhile, the electrical loads in the greenhouse consists of 6 exhaust fans, 2 axial fans, 5 fluorescent lamps and 1 irrigation pump. An Evolutionary Programming (EP) algorithm is chosen and programmed in Matlab to find the best configurations for optimum calibration of the greenhouse energy model. Three variables were chosen to find the best configuration which are the operating hours of Exhaust Fan, Axial Fan and Water Pump. The EP optimization algorithm in Matlab is coupled with building energy simulator, EnergyPlus using BCVTB as the coupling tool. Result shows that the EnergyPlus-EP model can provide NMBE and CV(RMSE) within the range recommended by the IPMVP protocol. The proposed method is not only requiring less computation time but also effective in searching for the best variables configuration with minimal error.