Geothermal energy-driven systems with integrated waste heat recovery units such as the use of fuel cells and thermoelectric module can help to improve the renewable energy contribution in the energy mix. Data-driven optimization can improve their economic and environmental performance and their macro-projection can help in the achievement of net-zero plans. This article extends the use of a framework containing the usage of data modeling and artificial intelligence to conduct different optimization scenarios of the geothermal-driven energy system. It includes the improvement of the economic, exergetic, energetic, and environmental performance through the development of various optimization scenarios. This is done through the development of an extensive thermodynamic model and validation based upon energy, exergy, economic, and environmental evaluations. Different machine learning techniques are adapted for digital twinning of the six performance indicators as a function of nine design variables including operational, source, and economic variables. It is shown that the artificial neural network offers the best statistical fit as compared to the other machine learning techniques including RMSE: 0.1768, R2:0.9999, MSE:0.0312, and MAE:0.1107 for the total work output. Energyefficient design has yielded a total work output of 1044.86 kW, with a first law efficiency of 0.3322. The economic design offers the lowest cost of electricity at only 34.004 $/hr. The sensitivity analysis has shown that the following parameters are the most sensitivity: turbine inlet temperature (18.19%) and pressure (18.23%), geothermal inlet temperature (16.34%) and pressure (18.00%), and the ammonia water concentration at the inlet of separator (15.96%).
