Machine learning-driven multi-objective optimization of fin geometry for enhanced charging and discharging performance in LTES system


İZGİ B.

Journal of Energy Storage, cilt.98, 2024 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 98
  • Basım Tarihi: 2024
  • Doi Numarası: 10.1016/j.est.2024.113210
  • Dergi Adı: Journal of Energy Storage
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC
  • Anahtar Kelimeler: Fin, Machine learning, Melting, Optimization, Phase change materials, Solidification
  • Yozgat Bozok Üniversitesi Adresli: Evet

Özet

This study investigates the impact of an innovative fin design on a vertical cylindrical Latent Thermal Energy Storage (LTES) unit and optimizes the fin geometry to enhance stored energy per mass (Est), power during charge (Pch), and power during discharge (Pdch). Utilizing a Computational Fluid Dynamics (CFD)-based numerical model, various fin geometries are explored via the Central Composite Design (CCD) method. Objective functions for optimization are established using machine learning algorithms. Multi-objective genetic optimization yields optimal solutions maximizing all three objectives, which are further compared against a no-fin scenario. The combined effects of the fin parameters on all three objectives are examined and the effect of fins on heat transfer and flow characteristics is discussed. Results demonstrate that while the addition of fins reduces stored energy per mass, it substantially increases charging and discharging power. Under equal importance weights for all three objectives, optimizing fin geometry leads to a 6.57% decrease in Est alongside significant increases in charging power (75.08 %) and discharging power (28.91 %) in comparison to the finless case. Moreover, fins have a greater influence on enhancing power than reducing energy per mass. Even when stored energy per mass is given 80 % importance weight, optimizing fin geometry leads to only a 3.25 % decrease in Est, while significantly increasing charging power (63.19 %) and discharging power (7.69 %) compared to the no-fin scenario.