DEVELOPMENT OF A COOLING LOAD PREDICTION MODEL FOR OFFICE BUILDINGS OF BAYERO UNIVERSITY KANO

Abstract

Overestimation of building cooling load results into an oversized cooling system with consequent
waste of energy. Conversely, thermal comfort would be sacrificed if the building cooling load is
underestimated because of the use of undersized cooling system. The computer-based methods for
cooling load calculation are costly while most of the inverse models are complex and require
enough time to train the models. In this paper, a simplified multiple linear regression model was
developed using the concept of Taguchi orthogonal array. The model developed has high
performance based on the computed ????2 and variance inflation factor. The prediction model
developed was validated using resampling technique and the consistency in the root mean squared
error (RMSE) in all the holdout samples indicates the high accuracy of the model developed. The
proposed model has high inference and prediction powers and could be used for predicting cooling
load of office buildings of Bayero University Kano (BUK) or any building with similar
characteristics.