Comparative Analysis of Predictive Models for Temperature Decay in Air Conditioned Space Influenced by Human Occupancy and Humidity

Abstract

Accurate prediction of indoor temperature is crucial for optimizing energy use and ensuring thermal comfort in air-conditioned environments. The study presents an empirical approach to model the cooling behaviour of a controlled room under varying conditions of air conditioner (AC) setpoint, occupancy, and humidity. Three predictive models linear, exponential (Newtonian cooling), and empirical were developed from the experimental data collected for the time taken for every 0.5 deg. C drop in room temperature. The empirical model, which incorporates humidity, occupancy, and room volume, demonstrated superior accuracy over traditional linear and exponential models, as evidenced by lower mean error and root mean square error (RMSE), and a higher coefficient of determination (R²). The empirical model showed excellent agreement with actual observations, with a mean percentage deviation of just 10.3%, a root mean square error (RMSE) of 0.15 deg. C, and a high R² value of 0.97. It successfully predicted the cooling time within – 20 to +20 seconds and accurately captured the cooling coefficient trends with respect to temperature setpoint and occupancy. The study establishes a reliable framework for predictive climate control based on real-world thermal interactions.