Réduction de la dépendance aux données dans la prédiction des performances d’un conditionneur d’air grâce à la modélisation de type boîte grise.

Reduced data dependency in air conditioner performance prediction through gray-box modeling.

Numéro : 2284

Auteurs : YOUSAF S., BRADSHAW C., KAMALAPURKAR R., SAN O.

Résumé

This paper presents a gray-box model developed for unitary air conditioning equipment. The proposed gray-box component-based model uses lumped heat exchanger models with a new formulation for UA of each heat exchanger in terms of the system model inputs which are outdoor and indoor temperatures along with indoor supply air flowrate. Key input parameters crucial for UA prediction are identified through analysis of the Pearson correlation matrix. Symbolic regression, a genetic algorithm, method is then utilized to derive the functional form of the UA correlation. Superheat and subcooling sub models are also formulated for both heat exchanger models. HEXs along with compressor, expansion valve and fan models are integrated into system model using energy conversation principles. The model accurately predicts cooling capacity, COP, and SHR for three modern air conditioning systems, with a MAPE of less than 3.4%. These findings indicate the potential utility of the gray-box model for future predictions, particularly with Building Energy Models (BEM) like EnergyPlus.

Documents disponibles

Format PDF

Pages : 10 p.

Disponible

Gratuit

Détails

  • Titre original : Reduced data dependency in air conditioner performance prediction through gray-box modeling.
  • Identifiant de la fiche : 30033191
  • Langues : Anglais
  • Sujet : Technologie
  • Source : 2024 Purdue Conferences. 19th International Refrigeration and Air-Conditioning Conference at Purdue.
  • Date d'édition : 17/07/2024

Liens


Voir d'autres communications du même compte rendu (51)
Voir le compte rendu de la conférence