Deep learning optimization of a biomass and biofuel-driven energy system with energy storage option for electricity, cooling, and desalinated water. (15th February 2023)

• Record Type: Journal Article
• Authors: Hai, Tao; Alsharif, Sameer; Aziz, Kosar Hikmat Hama; Dhahad, Hayder A.; Kumar Singh, Pradeep
• Is Part Of: Fuel. Volume 334(2023)Part 1
• Publication Date: 2023-02-15
• Subjects: Biomass -- Gasification -- Energy Storage -- Decision Making Parameters -- Machine Learning -- Optimization -- CAES -- Desalination; Fuel -- Periodicals; Coal -- Periodicals; Coal; Fuel; Periodicals; 662.6
• DOI: 10.1016/j.fuel.2022.126024
• Languages: English
• ISSNs: 0016-2361

Abstract

Highlights: Simulation of a biomass and gasigication based energy system. Using CAES as an efficient energy storage option. Techno-economic assessment of the system for the feasibility study. Multi-objective genetic algorithm optimization for better design. Deep learning based optimization of the proposed system. Abstract: To improve the turnover of thermodynamic cycles, combined cycles have gained a great deal of interest today. The primary objective of these systems is to maximize the utilization of wasted energy from power cycles to initiate cooling, heating, and desalination cycles. In the context of this project, the general cycle comprises a primary portion of power generation, the generation of freshwater, and cooling along with the essential heating of water. Additionally, compressed air energy storage was utilized to lower the expense of the complete cycle. Because of this, we should switch to using compressed air during the off-peak hours of the day and night when the power demand is at its highest. This article also includes a simulation of the gasification process, in which the higher temperature of the generated products is utilized to pre-heat the air. Considering each set of decision variables, the duration of each simulation ranges from 10 to 15 s. It is vital to utilize machine learning techniques to decrease the time needed for optimization to discover the ideal points. In conclusion, the genetic algorithm demonstrated that the exergy turnover and economic cost of the optimal point of the newly introduced cycle are equivalent to 36.21% and 6.56 $/h, respectively.