Optimisation study on solid–gas sorption based thermal energy storage using artificial intelligence approaches

作者：D. Li, Cebrail Turkeri, Yulong Ding, Y. Li · 发表于：Energy Conversion and Management · 年份：2026 · DOI：10.1016/j.enconman.2025.121022 · 研究领域：Adsorption and Cooling Systems、Carbon Dioxide Capture Technologies、Chemical Looping and Thermochemical Processes

• Optimisation of energy storage based on solid–gas sorption using AI approaches. • Data-driven framework based on ANNs for system modelling and performance prediction. • Intelligent framework based on MGA to enhance the system economic viability. • A storage system, using activated carbon and CO 2 , was developed and tested. • A critical enthalpy threshold identified to determine the system profitability path. Thermal energy storage based on solid–gas sorption is a promising solution for efficient renewable energy utilisation. However, the complex nonlinear dynamics of system operation and interdependent economic factors make performance prediction and profit exploration challenging. To address this, an optimisation platform based on Artificial Intelligence was developed. The platform consisted of a data-driven framework based on Artificial Neural Networks (ANNs) for predicting and monitoring system performance, and an intelligent framework based on Multi-population Genetic Algorithm (MGA) to enhance the system economic viability. An experimental storage system was developed and tested, and experimental data collected under varying adsorbate input conditions were used to train and validate the ANN models and develop economic models for optimisation. By accounting for temporal dependencies and delayed physical responses, the developed ANN models achieved high predictive accuracy, with absolute relative errors below 1 %. By establishing logic algorithms that demonstrated inter...