Document Type : Original Article
Authors
1
Department of Civil Engineering, Tabriz University, Tabriz, Iran
2
Department of civil engineering, Mohaghegh Ardabili university, Ardabil, Iran
3
Department of civil engineering, Islamic Azad university of Ardabil branch, Ardabil, Iran
4
Departement of Mechanical Engineering, University of Mohaghegh Ardabili, Ardabil, Iran
Abstract
This study tackles the imperative of energy-efficient building management by marrying advanced optimization algorithms with heating load (HL) prediction within the realm of heating, ventilation, and air conditioning (HVAC) systems. Highlighting the pivotal role of HL prediction in optimizing HVAC operations, fostering energy efficiency, and realizing cost savings, this research pioneers innovative strategies. It introduces a fusion of the African Vultures Optimization Algorithm (AVOA) and the Sand Cat Swarm Optimization (SCSO) with the Naïve Bayes (NB) model, aiming to elevate heating load prediction accuracy and streamline HVAC system optimization. These algorithms are employed to improve HVAC system control, equipment sizing, energy management, and cost reduction. The significance of accurate HL prediction in achieving energy efficiency, cost-effectiveness, and environmental sustainability in building operations is showcased. To gauge the predictive efficacy of the models, an array of performance metrics, including R2, RMSE, MSE, WAPE, and the NSE, were employed for assessment. These evaluations demonstrate that the NBSC model stands out as the most exceptional predictor in terms of real-world applicability and accuracy. It achieves an outstanding maximum R_train^2 value of 0.987, showcasing a high degree of explanatory power and exhibiting an impressively low 〖RMSE〗_train value of 1.166, signifying minimal prediction errors in comparison to other models. Additionally, the NBAV obtained a valuable result based on an R2 value of 0.978 and an RMSE value of 1.510, indicating the model's reliable results. This study did not only produce an accurate model.
Keywords
)