Optimizing novel thermal energy storage systems: enhancing melting efficiency with tubes, stands, and advanced machine learning techniques

The transition to renewable energy is crucial to meet growing global energy demands while minimizing environmental impact. Solar energy, a leading renewable source, faces limitations due to its intermittent nature, making energy storage systems essential for continuous supply. Thermal energy storage systems (TESSs) provide a compelling solution, especially by utilizing latent heat storage with phase change materials (PCMs), which efficiently store large amounts of energy. However, PCM-based systems suffer from low thermal conductivity and weak heat transfer performance. Numerous techniques to amend the thermal conductivity of PCMs, whether applied individually or together, show promise for improving the efficiency of TESSs. However, challenges like increased viscosity, decreased latent heat capacity, high costs, and complex manufacturing often limit their practical use. Thus, it is crucial to investigate more straightforward and more cost-effective solutions, like incorporating extended surfaces into the PCM container. In this study, a novel approach is taken to ameliorate the heat transfer surface area in TESSs by incorporating tubes and stands inside the PCM container rather than relying on conventional fins. A new compact TESS design was introduced, featuring fixed tubes supported by aluminum-made stands. Two artificial neural networks (ANN) models were trained with input parameters, including the length of the vertical stands, horizontal stands, and tilted stands, to anticipate the half-melting time and complete melting time of the PCM. In the end, two optimal designs of O-TESS1 and O-TESS2 were proposed by a single-objective optimization, focusing on minimizing the half and complete melting times of the PCM, respectively. A multi-objective optimization was also performed to balance the mentioned objectives. Using the generated data, a Pareto front and TOPSIS-selected designs were developed to visualize the trade-offs between these competing objectives. It took approximately 8515 s for the stand-free system to melt 50 % of the PCM, and it required more than 18,000 s to melt the material thoroughly. In contrast, O-TESS1, which had a half-melting time of 3192 s, reduced the melting time by an impressive 62.51 % compared to the stand-free TESS. O-TESS2 completed melting in 8401 s, during which the stand-free system reached only a 0.495 melt fraction, highlighting a remarkable 102 % improvement for O-TESS2 in achieving complete melting compared to the stand-free configuration. Besides, over the 18,000 s, stand-free TESS could only absorb 5277 kJ, which was much lower than the optimized samples of O-TESS1 and O-TESS2 with 8079 kJ and 8144 kJ, respectively. The TOPSIS-selected configuration demonstrated an intermediate performance, effectively balancing the features of O-TESS1 and O-TESS2. This design reached 50 % melting in just 3249 s, and achieved complete melting in 8942 s. In the end, to demonstrate that the proposed optimal configurations are not only efficient during melting but also perform well during solidification, the solidification behavior of the optimal designs were examined.