Cooling Load Forecasting Based On Hybrid Machine-Learning Application With Integration Of Meta-heuristic Algorithm

The imperative of accurately assessing the Cooling Load, which denotes the requisite energy to regulate temperatures within a given space, underscores its fundamental role in energy conservation, proficient management, and strategic planning. Precise prognostications of energy consumption are pivotal for optimizing resource allocation and fostering sustainability. The continuous refinement of predictive models is indispensable for bolstering the efficacy of energy systems in tandem with technological advancements. This research presents hybrid machine learning models integrated with advanced optimization techniques tailored for accurately predicting Cooling Load in buildings. By synergizing machine learning and optimization, it strives to pioneer predictive and managerial methods for cooling energy requirements, thus enhancing overall sustainability in the built environment. To attain this objective, the research employs the Gaussian Process Regression model in conjunction with both the Zebra Optimization Algorithm and the Gold Rush Optimizer. A comprehensive comparative analysis was conducted to meticulously scrutinize the predictive capabilities of the proposed models. As evidenced by the results, the GPZO (GPR+ZOA) model emerged as the frontrunner, attaining an outstanding R2 value of 99.6 percent. Furthermore, it showcased the lowest REVISE value, an impressive 0.596. These compelling findings unequivocally highlight the superior predictive accuracy and optimization proficiency of the GPZO model in accurately forecasting cooling load.