Novel optimized hybrid neuro-fuzzy approach for analysis of cold thermal storage system-assisted air conditioning system performance

Abstract

HVAC (Heating, ventilation, and air conditioning) systems can be used as a basic flexible load to provide approximately 20% of a building's energy demand and possess the capacity to transfer energy. Effective cooling load forecasting is one of the most important methods for achieving energy savings in buildings. This study proposes a Heap Optimization Based Generalized Intelligent Neural Fuzzy Control (HO-GINFC) for estimating the cooling load of an air conditioning system with cold thermal storage. It was primarily utilized to research historical meteorological data on cooling load. Using the quantitative data of a massive Saudi Arabian commercial structure, the model is validated. Using the HO algorithm, the performance of GINFC is optimized. Additionally, the cold storage tank's maximum capacity can be reached. Moreover, the results demonstrated that the HO-GINFC model accurately predicted the CV-RMSE (coefficient of variation of the root mean squared error), MSE (Mean Square Error), MAE (mean absolute error), R2 (squared correlation coefficient), and RMSE (root mean squared error) values to be 2.107%, 100%, 0.258%, 0.992%, and 1.058%, respectively. Comparative studies demonstrate that the HO-GINFC model outperforms neural networks in terms of prediction accuracy, execution speed, and resiliency when dealing with a large number of samples. The predictions of the model proposed in this work provide significant technical support for computing rising energy demands and can be implemented in actual engineering.