Gasification process modelling and optimization using Gaussian process regression and hybrid population-based algorithms

Gasification holds a central role in the thermochemical conversion of diverse carbon-rich feedstocks into valuable syngas, making a substantial contribution to the advancement of environmentally sustainable clean energy generation. The methodical modelling and optimization of gasification processes hold paramount significance in augmenting their overall operational efficiency. As part of the ongoing research endeavour, a novel approach is introduced that combines Gaussian Process Regression (GPR) modelling with the Population-Based Vortex Search Algorithm (PVSA) and the Dingo Optimization Algorithm (DOA). The core aim of this methodology is to enhance and optimize gasification processes. GPR serves as a surrogate model used to proficiently capture the intricate relationships between input variables and gasification performance metrics. The implementation of GPR ensures predictive accuracy, facilitating a more streamlined exploration of the design space while concurrently reducing the demands on computational resources. The integration of GPR modelling in conjunction with the hybrid approach, incorporating PVSA and DOA, markedly augments both the efficiency and precision in the design and control of gasification processes. The GPPV hybrid model has achieved the most optimal result with the highest R² value of 0.989 and 0.987 for the CH₄ and C₂H_n and the lowest RMSE of 0.476 and 0.164 for CH₄ and C₂H_n, indicating the reliability of the PVSA in optimizing the GPR model in predicting the syngas of gasification process. The framework expounded upon in this investigation provides a sturdy foundation for the progression of gasification technology, encompassing a diverse array of applications in the domains of clean energy production and sustainability endeavours.