Estimation of RC thermal model parameters using genetic algorithm
Abstract
This article presents the study result of using genetic algorithms to estimate parameters for the thermal dynamic models built based on networks of thermal resistors and capacitors. The proposed RC model structure includes 5 thermal resistors and 2 thermal capacitors, also called 5R2C thermal model. This model is an improved model based on the standard 5R1C thermal model. Parameters needed to be estimated are thermal capacitors and resistors. The genetic algorithm is used for the estimation of parameters. The simulation which is based on real collected data from a building shows that the model obtained gives a relatively high accuracy. In addition, the effectiveness of the genetic algorithm is also compared to that of the parameter-scanning algorithm based on correlation coefficient.
Moraref