Self-heating optimization of integrated system of supercritical water gasification of biomass for power generation using artificial neural network combined with process simulation. (1st June 2023)

• Record Type: Journal Article
• Authors: Liu, Shanke; Yang, Yan; Yu, Lijun; Cao, Yu; Liu, Xinyi; Yao, Anqi; Cao, Yaping
• Is Part Of: Energy. Volume 272(2023)
• Publication Date: 2023-06-01
• Subjects: Supercritical water gasification -- Biomass -- Power generation -- Artificial neural network -- Process simulation -- Self-heating; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2023.127134
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: Integrated supercritical water gasification of biomass for power generation (ISSCWBPG) is a promising energy conversion technology. ISSCWBPG is a complex system affected by multiple factors that determine whether the system can self-heat and whether it needs an external heat source. So it is a meaningful job to establish a model that can predict its power generation and heat difference. A process model with 86 types of biomass as raw materials was established, and 4709 samples of power generation indicators were obtained. The artificial neural network (ANN) was constructed based on these samples. Its coefficients of determination (R 2 ) on the test set are above 0.999 in the cases of power generation and heat difference, showing good generalization ability. Self-heating optimization of ISSCWBPG using the ANN model was carried out by taking wheat stalk as an example. Under the determined design flow and different temperatures, the matching operating parameters (DMC, FR) and the corresponding power generation under self-heating conditions were obtained. Compared with the calculation of the thermodynamic equilibrium model of Aspen plus, the errors were all within 10%. This work will provide theoretical guidance for the process design and optimization of the ISSCWBPG. Graphical abstract: Image 1 Highlights: ANN combined with process simulation was built to self-heating optimize for ISSCWBGP. The determination coefficients of ANN model on the test set were all above 0.999. The impact of features to targets were studied by SHAP method based on game theory. The optimal corresponding operating parameters of the autothermal system was obtained.