Modeling and optimization of hybrid geothermal-solar energy plant using coupled artificial neural network and genetic algorithm

Avoiding solar thermal energy storage reduces efficiency in hybrid solar-geothermal energy systems, making them impractical. To address this challenge, a synergistic approach involves the integration of these resources in the construction of hybrid power plants. An experiment is conducted with five independent variables: evaporator temperature, separator pressure, inlet pressure of turbine 2, effectiveness of vapor generator 1, and desalination mass ratio. Unlike most energy systems that rely on a single optimization pattern, this study utilizes response surface methodology (RSM) to design and gather data through simulation. Additionally, an artificial neural network (ANN) is employed alongside RSM to establish mappings from independent variables to response variables, including thermal efficiency and levelized cost of product. The selection of objective functions derived from ANN is predicated on their commendable performance, denoted by an R-squared value of 1. Furthermore, a cost function is formulated with the dual aims of maximizing thermal efficiency and minimizing the levelized cost of product. This function is subsequently optimized through the application of genetic algorithms (GAs). The findings elucidate that specific parameter values-namely, a desalination mass ratio of 2.43, separator pressure of 455.77 kPa, effectiveness of vapor generator 1 of 0.82, inlet pressure of turbine 2 of 12000 kPa, and evaporator temperature of -11.51 CC-conducive to the optimal condition are identified, yielding a thermal efficiency of 30.47% and a levelized cost of product of 13.04 $/GJ. This endeavor is anticipated to furnish an algorithmic framework not only for modeling hybrid plants but also for optimizing electrical power generation processes.