Use of Artificial Neural Networks for the Evaluation of Thermal Comfort Based on the PMV Index

Abstract

This study aims to develop an artificial neural network (ANN) model to predict the predicted mean vote (PMV) index, a key Indicator of thermal comfort. Based on the ASHRAE II dataset, our approach uses the six PMV variables: air temperature, relative humidity, air velocity, radiative mean temperature, clothing insulation, and metabolic rate. However, accurately calculating PMV to determine the thermal comfort of a space can be complex due to the non-linear relationships between these different parameters. Sensitivity analysis of these parameters, performed by the Spearman rank method, identifies the most influential parameters on thermal comfort. The ANN model is trained and tested on 26,805 datasets. The results demonstrate a strong predictive capacity of the ANN, attested by a coefficient of determination R2 of 0.99 and a low root mean square error RMSE.