Exergy analysis and optimization of the Rankine cycle in steam power plants using the firefly algorithm

The analysis of the exergy efficiency has always been considered as a fundamental criterion to study the behavior of the thermodynamic cycles. In this research, the exergy analysis of a steam power plant for generating electricity with Rankine thermodynamic cycle is carried out. Zarand steam power plant, which is located in the Kerman province, is considered as a case study. In order to optimize these thermodynamic processes and to achieve the highest exergy efficiency value, some primary parameters were considered as the decision variables. By changing the values of these parameters, an attempt was made to enhance the exergy efficiency by using a novel approach. The six decision variables, which are, output temperature and pressure values of the boiler, as well as the output pressure values of the four stages of the turbine, were chosen on the basis of probability of variations in a certain range of electricity generation parameters for the studied power plant. The exergy efficiency was considered as the objective function. Afterwards, optimization of the power plant by employing the firefly algorithm, which is one of the relatively latest invented algorithms for solving the optimization problems, was carried out. The firefly model performs the optimization process inspired by the behavior and action of fireflies to attract mates and reject enemies. For the purpose of analysis of the exergy efficiency, at the first stage, the optimization of exergy efficiency function was performed for the studied steam power plant, and then the results were compared with the solutions obtained using the genetic and particle swarm optimization algorithms. Final results are indicative of the fact that by appropriate changes in the decision variables and employing the firefly algorithm, the exergy efficiency of the thermal power plant increased from 30.1 to 30.7037 percent. This increase was equivalent to 0.6037 for the cycle, and compared to the results obtained from the genetic and swarm particle optimization algorithms, it was 0.04% and 0.0398% higher, respectively.