Novel Neural Network Optimized by Electrostatic Discharge Algorithm for Modification of Buildings Energy Performance

Proper analysis of building energy performance requires selecting appropriate models for handling complicated calculations. Machine learning has recently emerged as a promising effective solution solving this problem. The present study proposes novel integrative machine model predicting two parameters residential buildings, namely annual thermal demand (DThE) and weighted average discomfort degree-hours (HDD). is feed-forward neural network (FFNN) that optimized via the electrostatic discharge algorithm (ESDA) analyzing characteristics finding their optimal contribution to DThE HDD. According results, proposed an double-target can predict required with superior accuracy. Moreover, further verify efficiency ESDA, was compared three similar optimization techniques, atom search (ASO), future (FSA), satin bowerbird (SBO). Considering Pearson correlation indices 0.995 0.997 (for HDD, respectively) obtained ESDA-FFNN versus 0.992 0.938 ASO-FFNN, 0.926 0.895 FSA-FFNN, 0.994 SBO-FFNN, ESDA provided higher accuracy training. Subsequently, by collecting weights biases FFNN, formulas were developed easier computation HDD in new cases. It posited engineers experts could consider use along investigating buildings.