Energy consumption prediction of air-conditioning systems in eco-buildings using hunger games search optimization-based artificial neural network model

Abstract

The energy consumption of the heating, ventilation, and air-conditioning systems is a big concern of energy-efficient buildings in smart cities, especially for cooling and heating loads. Although many constant efforts have been adopted to improve the air-conditioning systems' performance; yet, the energy consumption for these systems is still high. This study is, therefore, to propose a novel soft computing model for accurately calculating the energy consumption of the cooling load and heating load systems in energy-efficient buildings of smart cities, namely hunger games search-based multiple layers perceptron neural network model. This model can assist energy-efficient buildings in improving the performance of the air-conditioning systems and sustainable environmental protection. Accordingly, the novel hunger games search optimization algorithm optimized a multiple layers perceptron neural network to decrease the model's error in estimating cooling and heating load. Three other benchmark intelligent models, such as Harris hawks optimization-based multiple layers perceptron neural network, particle swarm optimization-based multiple layers perceptron neural network, and grey wolf optimization-based multiple layers perceptron neural network, were also developed for similar purposes, and their accuracies were then considered as the indicators for evaluating the proposed hunger games search-based multiple layers perceptron neural network model. The results indicated that the proposed soft computing paradigm is the most robust model for estimating the air-conditioning systems' cooling load and heating load in this study. A critical report was given in this study based on the evidence regarding the unnecessaryness of the relative compactness and glazing area distribution variables in calculating the energy consumption of future buildings. They may lead to lower accuracy of predictive models for predicting the energy consumption of future buildings.