Optimisation de la position des soupapes de refoulement intermédiaire dans un compresseur à spirale avec apprentissage par renforcement approfondi. 

Optimization of intermediate discharge valve positions in a scroll compressor with deep reinforcement learning.

Numéro : 1299

Auteurs : YOO J., JEONG T., KIM D.

Résumé

As seasonal climate changes, modulating the pressure ratio of the scroll compressor is required to operate heating and cooling cycles. Over-compression of a scroll compressor caused by a wide range of operating conditions reduces energy efficiency and gives structural damage to the scroll. Intermediate discharge valves (IDVs) applied in the scroll compressor can avoid over-compression by discharging compressed refrigerant. However, designing the positions of IDVs for proper pressure control without geometric conflict requires time-consuming works due to manual design. This research aims to optimize IDV positions located on a fixed scroll with deep reinforcement learning to maximize the energy efficiency for given scroll design, number of the IDV, and range of operating conditions. An analytical pressure simulation for the scroll compressor, which is based on the energy and mass balance equations, is established as an environment of the reinforcement learning. A proximal policy optimization agent is trained to maximize the summation of adiabatic efficiency for various operating conditions. The geometric conflict of the IDV is considered by using penalty on a reward function, and the reward evolution of the training is presented to show the optimization results. The optimization technique proposed in this study is found to reduce the time cost of developing scroll and IDV design by the automatic design process.

Documents disponibles

Format PDF

Pages : 10 p.

Disponible

Gratuit

Détails

  • Titre original : Optimization of intermediate discharge valve positions in a scroll compressor with deep reinforcement learning.
  • Identifiant de la fiche : 30030305
  • Langues : Anglais
  • Sujet : Technologie
  • Source : 2022 Purdue Conferences. 26th International Compressor Engineering Conference at Purdue.
  • Date d'édition : 15/07/2022

Liens


Voir d'autres communications du même compte rendu (83)
Voir le compte rendu de la conférence