Thermo-economic optimization of the hybrid geothermal-solar power system: A data-driven method based on lifetime off-design operation

The hybridization of the geothermal and solar system provides a potential solution to the problem of geothermal power plants being vulnerable to ambient temperature changes and reservoir degradation. Existing research mainly focuses on its yearly operation or layout improvement, neglecting proactive design based on the off-design performance, due partly to the complexity and elapsed time. Therefore, for the first time, this study proposes a novel data-driven method to realize the fast and accurate prediction of the system off-design performance, thereby making it possible for the lifetime design of the hybrid geothermal-solar power system considering its real-time operation. In this work, an artificial neural network (ANN) is firstly trained and validated to predict the hourly performance of the hybrid system based on an organic Rankine cycle (ORC) over a 30-year life span. Then the trained ANN is combined with a multi-objective optimization procedure to determine the best system design. Results prove the effectiveness of this ANN-based approach for the hybrid system proactive design, which could greatly reduce the calculation time while maintaining accuracy within 2%. Besides, this novel approach has proven to be more efficient in improving the lifetime electricity generation (E_tot) and net present value (NPV) by up to 17% and 14%, respectively. Moreover, compared with the stand-alone geothermal power plants, this hybrid system shows higher turbine/pump efficiency during operation and represents potential thermo-economic advantages as the price of solar collectors declines below 75 $·m⁻². The results provide a valuable reference for hybrid systems, and the ANN-based design method can also be directly applied to other power systems, thereby facilitating the in-depth development of optimization design for the renewable energy systems.