Experimental characterisation of the periodic thermal properties of walls using artificial intelligence
The energy performance of a building is affected by the periodic thermal properties of the walls, and reliable methods of characterising these are therefore required. However, the methods that are currently available involve theoretical calculations that make it difficult to assess the condition of...
| Autores: | , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2020 |
| País: | España |
| Institución: | Universidad de Sevilla (US) |
| Repositorio: | idUS. Depósito de Investigación de la Universidad de Sevilla |
| OAI Identifier: | oai:idus.us.es:11441/103384 |
| Acceso en línea: | https://hdl.handle.net/11441/103384 https://doi.org/10.1016/j.energy.2020.117871 |
| Access Level: | acceso abierto |
| Palabra clave: | Periodic thermal transmittance Energy demand ISO 13786 Multilayer perceptron Random forests In-situ |
| Sumario: | The energy performance of a building is affected by the periodic thermal properties of the walls, and reliable methods of characterising these are therefore required. However, the methods that are currently available involve theoretical calculations that make it difficult to assess the condition of existing walls. In this study, the characterisation of the periodic thermal variables of walls using experimental measurements and methods as described in ISO 13786 was assessed. Two regression algorithms (multilayer perceptron [MLP] and random forest [RF]) and input variables obtained using two experimental methods (the heat flow meter and the thermometric method) were used. The methods gave accurate estimates, and better statistical parameter values were given by the RF models than the multilayer perceptron models. For all the periodic thermal variables, the percentage differences between the actual values and the estimated values given by the RF algorithm were low. The heat flow meter and the thermometric methods can both be used to characterise accurately the periodic thermal properties of walls using the RF algorithm. The variables specific to each method, including the wall thickness and the date of construction, affected the accuracies of the models most strongly. |
|---|