Concentrating solar power (CSP) is a high-potential renewable energy source that can leverage various thermal applications. CSP plant development has therefore become a global trend. However, the designing of a CSP plant for a given solar resource condition and financial situation is still a work in progress. This study aims to develop a mathematical model to analyze the levelized cost of electricity (LCOE) of Thermal Energy Storage (TES)-integrated CSP plants in such circumstances. The developed model presents an LCOE variation for 18 different CSP configurations with TES incorporated for Rankine, Brayton, and combined power generation cycles, under regular TES materials and nano-enhanced TES materials. The model then recommends the most economical CSP plant arrangement. Within the scope of this study, it was found that the best configuration for electricity generation is a solar power tower with nano-enhanced phase change materials as the latent heat thermal energy storage medium that runs on the combined cycle. This returns an LCOE of 7.63 ct/kWh with a 22.70% CSP plant efficiency. The most favorable option in 50 MW plants is the combined cycle with a regular TES medium, which has an LCOE of 7.72 ct/kWh with a 22.14% CSP plant efficiency.
Efficient hydration and cooking of grains and legumes are essential for reducing energy consumption and improving the quality of cooked products. This study presents an experimental and machine learning framework to model and optimize the hydration cooking relationship in long grain and short grain rice, chickpeas, and mung beans. All experiments were conducted in triplicate (n = 3), and results are reported as mean ± standard deviation. Soaking experiments were performed at 25-55°C for up to 8 h, and moisture uptake was quantified using gravimetric analysis. The Peleg model successfully described hydration kinetics across all grain types and soaking temperatures, yielding coefficients of determination (R2) ranging from 0.96 to 0.99. Cooking trials quantified cooking time and energy consumption, while external datasets were incorporated to improve model generalization. A multilayer feedforward neural network was implemented and trained to predict cooking time and energy requirements from soaking temperature, duration, and moisture fraction. The model achieved strong predictive performance on external validation datasets (R2 = 0.94, MAE = 1.8 min), while internal test results yielded R2 values of 0.98 for cooking time and 0.97 for energy consumption. Cooking time was reduced by up to 40 ± 3%, and energy consumption decreased by approximately 18-22%, depending on grain type. Statistical analysis confirmed that the effects of soaking temperature on hydration, cooking time, and energy consumption were significant (p < 0.05). Optimized hydration also promoted uniform moisture distribution and minimized grain fracturing and splitting. The integration of experimental kinetics and machine learning prediction provides a robust foundation for intelligent cooking process design and demonstrates a pathway toward AI enabled smart cookers with adaptive temperature and energy control.