Modeling and Optimization of HVAC System Based on GA-ANFIS Algorithm

Abstract

Building heating, ventilation and air conditioning (HVAC) systems consume more than half of the building's energy consumption, and the efficient operation of the system is a key goal to save energy and costs of the building, as well as to ensure a comfortable living and working environment. This paper proposes a data-driven method based on the HVAC system air handling unit (AHU) energy consumption prediction and optimization method and selected adaptive network fuzzy inference system (ANFIS) to predict the system energy state and consumption. By comparing and analyzing ANFIS with other data-driven models, it is found that the modeling ability of ANFIS surpasses the most other data mining algorithms for nonlinear systems, and it can be used to establish the nonlinear relationship model between the system energy consumption and optimal control point parameters including controllable and uncontrollable parameters. Our method integrates genetic algorithm (GA) to optimize the prediction and optimized the air supply temperature and static pressure of the AHU, and to ensure the balance between energy consumption and indoor comfort. Experiments on real data show that GA-ANFIS modeling and optimization algorithms can save about 16% energy consumption of HVAC.