Inspection of the Identification of Thermal Modeling Process in Buildings: A Proposed Methodology for a Case Study in Tropical Climate

Abstract

A methodology to identify thermal models based on RC networks is proposed and verified in two case studies located in Panama City. Data for identification is obtained through simulation via a whitebox model for a whole year. The dispersion of the parameter estimation is studied by training the models with different datasets. The models based on the 1R1C and 2R1C networks are the only ones with consistent estimates, hence, identifiable parameters. The 2R1C model that incorporates the mean radiant temperature as an input is notably better at describing the thermal dynamics, obtaining an RMSE of 0.5386 • C during validation for a training set of six months.

A Simplified Model Parameter Identification Methodology for Buildings Indoor Thermal Behavior Control: A Case Study in a Tropical Climate of Panama

“…In [24] a 6R5C network is simplified to a 2R1C network and able to successfully model a residence equipped with air conditioner, with an absolute error of 0.72°C when the model is trained with 20 days of data. In [25] a methodology was developed to identify thermal models based on RC networks for a passive residence located in Panama. As a hybrid modelling technique, grey-box models are also able to couple different model structures to improve or complement model estimation.…”

The Thermal Network Approach to Model Occupants’ Heat and CO2 Generation Interactions: A Case Study in an Office Building in Panama

“…Subsequently, these differential equations were represented in the state of space to be processed by the system identification developed in MATLAB, following the proposed methodology in [16].…”