Evaluation of Structural Driven Geothermal Systems: A Comprehensive Geophysical Analysis in Thar Platform, Southern Indus Basin, Pakistan

Abstract

Geothermal energy resources are a renewable energy source that is an emerging field worldwide. These resources are economically viable and environmentally sustainable. Geothermal energy potential exists in Pakistan; however, these resources have not yet been fully tapped due to a lack of research interest and proper methodology. The current study aims to determine the potential of geothermal energy in the subsurface by utilizing 2D and 3D seismic and well data to explore the geothermal potential of the sandstone reservoir in the Southern Indus Basin of Pakistan. The study area, Sanghar Block, lies in the Thar Platform. The major subsurface structures present in this area are horst and graben, which are considered perfect structure for the trapping of geothermal energy. The detailed studies are performed on the Lower Goru Formation, which is considered as major reservoir in the study area. Effective evaluation of geothermal reservoir characteristics from well and seismic data plays a fundamental role in harnessing subsurface geothermal resources. Facies identification was achieved using borehole data through artificial intelligence techniques, indicating that the key facies present in the Lower Goru Formation are shale and sandstone. As the shale containing naturally radioactive lithology have a radioactive element like U, Th and K, which are also considered a source of heat. It has been considered based on present studies that the existence of radioactive elements such as U, K, and Th in these facies is a dynamic source of heat in the subsurface. Geothermal reservoir properties such as average porosity, the volume of shale, heat production, radiogenic heat production, and permeability were computed from well logs and seismic data. A DFFNN was utilized to demonstrate the variation of geothermal reservoir characteristics along the seismic transect. In the DFFNN data is split into 70 % for training and 30% for testing purposes. The models are optimizing by using multiple hidden layers which control the over and underfitting of geothermal and petrophysical model parameters. The major function of DFFNN is to boost the incorporation of well and seismic data for geothermal reservoir characterization by estimating rock characteristics gained from model based seismic inversion. DFFNN technique achieved excellent correlation values from 85-98% for geothermal and petrophysics properties by utilizing multiple attributes while in traditional techniques which often suffer from poor resolution and high ambiguities when estimating these properties. The current research is innovative because of its amalgamation of machine learning and statistical methods, which permits the evaluation of geothermal properties (average porosity, the volume of shale, heat production, radiogenic heat production, and permeability) on seismic sections that are typically insights in the studied interval. The results of subsurface geothermal reservoir characteristics average values derived from logs curve data: average porosity (15.90%), volume of shale (33.80%), heat production (0.933 µW/m3), radiogenic heat production (1.20 µW/m3), and permeability (16.37 mD) are relatively promising which signifies that the present study zone is promising for geothermal potential. The quality control of seismic to well ties is confirmed by generating the relationship between synthetic seismogram, real seismic and time to depth chart, while the validation of facies prediction is confirmed by using well reports. The geothermal analysis is compared with published and international values for the validation of results Based on the current results, it has been determined that the innovative methods have enhanced prediction accuracy and minimized the ambiguity in geothermal characteristics, and this study has a positive impact on Pakistan renewable policy as it provides an alternate source of fossil fuel and coal and promote renewable energy target. Key words: Geothermal Energy, Heat Production, Radiogenic Heat Production, Permeability, Machine Learning, Deep Feed Forward Neural Network.