A Novel Air-Conditioning Load Prediction Based on ARIMA and BPNN Model

Abstract

Accurate air-conditioning load forecasting is the precondition for the optimal control and energy saving operation of HVAC systems. Many forecasting techniques such as support vector machine (SVM), artificial neural network (ANN), autoregressive integrated moving average (ARIMA) and grey model, have been proposed in the field of air-conditioning load prediction. However, none of them has enough accuracy to satisfy the practical demand. Therefore, a novel method integrating ARIMA and Artificial Neural Network (ANN) is presented to forecast an air-conditioning load. ARIMA is suitable for linear prediction and ANN is suitable for nonlinear prediction. This paper also investigates the issue on how to effectively model short term air conditioning load time series with a new algorithm, which estimates the weights of the ANN and the parameters of ARMA model. Experimental results demonstrate that the hybrid air conditioning load forecasting model can be an effective way to improve forecasting accuracy achieved by either of the models used separately.

Prediction of building energy consumption using an improved real coded genetic algorithm based least squares support vector machine approach

“…All of the initial individuals ( and ) were generated randomly from the defined search space. The lower and upper bounds of the search space were selected to allow and to be in the range [2,8] and [1,3], respectively. The crossover and mutation probabilities were 100% and 35% [38], respectively.…”

“…Because of the strong nonlinear mapping behavior of energy consumption data, some researchers have begun using artificial intelligent techniques to focus on the prediction of building energy consumption. In particular, the artificial neural network (ANN) theory is widely used in nonlinear time series prediction, which has been applied generally to predict building energy consumption [1][2][3][4][5][6][7][8][9]. However, conventional ANN has several drawbacks, such as the need for a large number of controlling parameters, difficulty in obtaining stable solutions, and thus lack of generalizability.…”

A novel building cooling load prediction based on SVR and SAPSO

“…In present, because of its strong non-linear mapping ability, artificial neural networks (ANN) are widely accepted as a technology offering an alternative way to tackle complex and ill defined problems, which have been popularly applied to predict the building cooling load [2] and building energy consumption. ANNs-based models seem to obtain improved and acceptable performance in cooling load forecasting issue, however, the conventional ANNs still suffer from several weaknesses such as the need for a large number of controlling parameters, the difficulty in obtaining stable solutions, the danger of overfitting and thus the lack of generalization capability.…”

Prediction Model on Room Temperature Side Effect due to FastADR Aggregation for a Cluster of Building Air‐Conditioning Facilities

“…On the other hand, AR models distinguished in transient changes are considered suitable for slowly changing room temperature. While there are a few studies combining these two predictive models [16][17][18][19][20], we here adopt AR-NN combined model proposed in our previous studies [21][22][23].…”