Optimization of Solar-Geothermal Hybrid Power Plant System Through Deep Learning

In 2017, the Kingdom of Saudi Arabia (KSA) began to construct NEOM, a city powered entirely on renewable energy. Geothermal and solar are two of the most abundant renewable resources in NEOM, with great potential for electricity generation. What's more, the hybridization of two resources can significantly enhance the performances of the binary geothermal power plant. However, most of the existing simulation models fail to integrate the geothermal reservoir, the solar field and the power plant, leading to inaccurate evaluations of the electricity capacity of both energy resources. Therefore, in this work, we present a novel optimization framework in which the geothermal reservoir model, the solar field model as well as the hybrid power plant model are fully coupled. The geothermal production temperature from the reservoir is predicted through decline curve analysis (DCA) and deep neural network (DNN), which is then coupled with the in-house solar field model and power plant model. Finally, the three parts in the integrated model are optimized simultaneously by a multi-objective optimizer for the best thermodynamic and economic performances. Results show that the DNN model can accurately and efficiently predict the parameters of the decline model with R² scores of 0.973, 0.961, 0.953, 0.998 and 0.996, with an error of 0.56 ± 0.42% for the bottom hole temperature and 0.55 ± 0.42% for the surface production temperature. The average CPU time is 0.0026 s per case. Both the solar field model and the power plant models are validated with data from the literature with tolerable deviations. Optimization results indicate that the supercritical configuration is the optimal configuration for both the geothermal stand-alone and hybrid power plants. The hybridization of the solar and geothermal can generate more electricity and reduce the levelized cost of electricity. Our optimization work can provide guidance to the implementation and operations of the geothermal reservoir, the solar field and the power plant under optimal conditions.