Dual-channel modular phase change material thermal energy storage inspired by plate heat exchangers: ANN-assisted optimization using a multilayer perceptron
International Communications in Heat and Mass Transfer, cilt.180, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 180
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.icheatmasstransfer.2026.112409
- Dergi Adı: International Communications in Heat and Mass Transfer
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Academic Search Ultimate (EBSCO), Engineering Source (EBSCO)
- Anahtar Kelimeler: Heat transfer fluid channel, Multilayer perceptron, Phase change material, Prediction model, Solar-driven resources, Thermal energy storage
- İstanbul Gelişim Üniversitesi Adresli: Evet
Özet
The growing integration of renewable energy systems has heightened the demand for efficient thermal energy storage technologies to balance energy supply and demand. Phase change material (PCM)-based latent heat storage systems are attractive owing to their significant energy storage density and nearly isothermal operation; nevertheless, conventional configurations often suffer from low heat transfer rates, non-uniform melting, and limited scalability. In many heat-exchanger-inspired designs, heat transfer fluid (HTF) channels are located only along the outer boundaries of the PCM enclosure, leaving the central region thermally underutilized. To mitigate these drawbacks, the current work develops a novel modular energy storage system inspired by plate heat exchanger architecture. The system consists of 60 independent units that allow flexible assembly and scalable storage capacity. A key innovation is the incorporation of an additional HTF channel inside the PCM enclosure, forming two parallel heat transfer paths that enhance the effective heat transfer surface and decrease thermal resistance. To accelerate the design process, a multilayer perceptron surrogate model was developed to anticipate stored energy during charging. A two-stage optimization framework based on a genetic algorithm and the TOPSIS method was applied to identify the optimal configurations. The optimized designs demonstrated remarkably better performance than the reference configuration. At 9000 s, the half-time optimized design stored the largest amount of energy, reaching 9402 kJ, followed closely by the balanced design, which stored 9287 kJ, while the full-time optimized design stored 8654 kJ. In comparison, the core design stored only 3818 kJ at the same time. Compared with the core design, the stored energy improved by approximately 146% for the half-time optimized design, 126.6% for the full-time optimized design, and 143.2% for the balanced design. At the end of the melting process (18,000 s), the liquid fractions were 0.9649 for the half-time optimized design, 1 for both the full-time optimized design and the balanced design, and 0.3745 for the reference configuration. Correspondingly, the stored energy increased from 6457 kJ in the core design to 12,585 kJ, 13544 kJ, and 13,113 kJ for the half-time optimized design, full-time optimized design, and balanced design, respectively, representing improvements of up to 109.8%.