Using Machine Learning to Predict Performance of Two Cogeneration Plants from Energy, Economic, and Environmental Perspectives.

This study deals with multi-objective energy systems' performance analysis and optimization, including power generation and chilling. The system studied comprises a fuel cell, a gas turbine, an absorption chiller, and a steam recovery generator. This way, the cycle is thermodynamically modeled to allow catching optimum design points using the genetic algorithm. The study compares the optimum points of this cycle with that in a hybrid fuel cell (FC) - gas turbine (GT) cycle. The study uses machine learning methods for optimization to reduce calculation time and costs. This energy system can generate 500–1000 kW of output power. The cooling load varies from 10 to 65 kW, depending on the decision-making parameters. According to the optimization results, the energy efficiency can be improved by up to 65%, while the total cost rate can be diminished by up to $16 per hour in the improved cycle. Environmentally, the exergoenvironmental index of 0.4803 and the sustainability index of 2.443 were obtained for the hybrid gas turbine-fuel cell cycle.