Thermo-economic optimization of PCM-integrated ASHP systems using machine learning–based random Forest surrogate models
Applied Thermal Engineering, cilt.303, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 303
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.applthermaleng.2026.132543
- Dergi Adı: Applied Thermal Engineering
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, DIALNET, Business Source Ultimate (EBSCO)
- Anahtar Kelimeler: Air source heat pump, Load shifting, NSGA-II, Phase change material, Random Forest, Thermo-economic optimization
- Yozgat Bozok Üniversitesi Adresli: Evet
Özet
The temporal mismatch between residential heating demand and the optimal operating conditions of Air Source Heat Pumps (ASHPs) increases both grid stress and operating costs. Thermal Energy Storage (TES) provides an effective solution by shifting heating loads to off-peak periods. This study proposes a thermo-economic optimization framework for a water-side Phase Change Material (PCM)-based TES integrated with an ASHP. A transient thermodynamic model is employed to generate training data for a Random Forest (RF) surrogate model (R2>0.98), reducing evaluation times from conventional simulations to milliseconds. The surrogate model was subsequently coupled with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to maximize the average Coefficient of Performance (COP), minimize peak compressor power consumption (Ẇcomp,peak), and maximize the daily economic savings (Ds) under Time-of-Use (TOU) electricity tariffs. The optimization results indicate diminishing economic returns with increasing PCM capacity, identifying an optimal storage range of 400–420 kg. The selected Pareto-optimal configuration decreases peak-hour compressor electricity consumption by 48.2% while achieving daily savings of 28 TRY, corresponding to a 10% reduction in operating cost relative to a conventional ASHP, with only a 4.5% reduction in the daily average coefficient of performance. The proposed EES–RF–NSGA-II framework offers a computationally efficient methodology for predictive control and demand-side management of residential heat pump systems within smart grids.