Estimation of Chiller Electric Power of Central Air Conditioner using Feedforward Neural Network

Abstract

With the development of technology, the proportion of renewable energy with stochastic and intermittent characteristic, like wind power, photovoltaic power, is higher and higher in generation side, which results in the controllability of power generation decreases. And accordingly, the potential of demand response receives attention. Central air conditioning loads are suitable for demand response for their high proportion in load and large heat inertia. And accordingly, the power estimation of chiller of central air conditioner is essential. However, the control technique of chiller is trade secret for central air conditioner company. In this paper, a novel power estimation method of chiller using feedforward neural network is proposed. Furthermore, central air conditioning loads are explored to participate in AGC control of power system according to the estimated chiller power. Experiment and simulation results verify the effectiveness of the proposed method.