A Predictive Exergy-Adaptive Genetic Algorithm for the Optimization of the DMR Liquefaction Process

Area of Science:

• Chemical Engineering
• Thermodynamics
• Artificial Intelligence

Background:

• Growing global energy demands necessitate improved efficiency in industrial processes.
• The dual mixed-refrigerant (DMR) natural gas liquefaction process presents complex optimization challenges due to coupled parameters.
• Conventional optimization methods struggle to achieve global optima for DMR processes.

Purpose of the Study:

• To develop an intelligent optimization strategy for the DMR process.
• To integrate exergy analysis with machine learning for enhanced process optimization.
• To investigate the impact of key operating parameters on energy consumption and exergy efficiency.

Main Methods:

• Development of a detailed Aspen HYSYS simulation for the DMR process.
• Integration of exergy analysis with a Predictive Exergy-Adaptive Genetic Algorithm (PEA-GA).
• Utilization of a radial basis function (RBF) neural network for accelerated optimization search.

Main Results:

• Optimized scheme reduced specific energy consumption (SEC) from 0.281 to 0.252 kWh/kg LNG.
• Exergy efficiency was increased from 31.62% to 35.45%.
• Significant reduction in minimum temperature differences in heat exchangers, improving composite-curve matching.

Conclusions:

• The PEA-GA strategy effectively optimizes DMR processes for energy savings.
• The findings offer theoretical guidance for energy-saving retrofits in LNG facilities.
• Intelligent operation strategies can significantly enhance the performance of liquefaction plants.