Quantifying and integrating decision-making for geothermal well life-cycle risk under geological uncertainty: A review

Abstract

As global energy systems shift toward low-carbon models, deep geothermal energy is an important clean resource with stable output and high potential. Commercial development, however, remains constrained by heavy upfront capital costs and geological uncertainty. Well construction and surface facilities typically account for 50 to 75 percent of total project investment, and extreme downhole conditions, such as high temperatures, abrasive formations, and complex fracture networks, readily trigger incidents causing substantial non-productive time (NPT) and cost overruns. Managing risk while improving cost efficiency across the full well life cycle is therefore a central industry challenge. This paper reviews recent progress and frameworks for integrated decision-making and risk control in geothermal wells. In siting, target-area selection has evolved from qualitative expert judgment toward spatial decision systems combining multi-criteria evaluation (MCE) with machine learning; with well-field co-optimization, these reduce blind early siting and sunk costs. In construction, where mechanistic and numerical models are hard to parameterize and computationally costly, data-driven algorithms now support real-time rate-of-penetration (ROP) prediction, transient bottom-hole thermal management, and incident early warning. In investment appraisal, probabilistic simulation, value-of-information (VOI) theory, and life-cycle assessment (LCA) have moved evaluation beyond deterministic single-well cost estimates toward multi-objective optimization of levelized cost, long-term thermal revenue, and carbon footprint.