Proactive Hybrid Risk Assessment Framework for Geothermal Drilling: Insights from the Utah FORGE Field

Abstract Persistent inefficiencies in geothermal drilling, marked by frequent tool failures, directional control challenges, and extended non-productive time (NPT), continue to inflate project costs and delay well delivery. In some developments, drilling alone consumes up to 75% of total project expenditure, far exceeding typical oil and gas benchmarks. To address these inefficiencies, this paper introduces a comprehensive risk assessment framework capable of rapidly identifying, quantifying, and mitigating drilling-related risks in unconventional geothermal wells, thereby improving both economic and operational performance. The proposed methodology integrates and analyzes over 550 validated NPT events from seven wells drilled at the Utah Frontier Observatory for Research in Geothermal Energy (FORGE) site, leveraging more than 300 Daily Drilling Reports (DDRs), directional survey data, mud logs, and end-of-well reports (EOWRs). Events were categorized into standardized classes and examined to determine dominant failure modes, root causes, and potential mitigation strategies. A probabilistic Monte Carlo simulation approach was then applied to quantify risk and uncertainty, capturing the stochastic nature of drilling operations. Results indicate that five primary problem areas, including rig equipment failures, directional tool malfunctions, logging tool issues, poor hole cleaning and wellbore instability, and operational/personnel errors, account for over 60% of total NPT at the FORGE site. Well 16A(78)-32 recorded the highest cumulative NPT (420 hours), driven by frequent rig equipment breakdowns, repetitive tool malfunctions, and avoidable operational oversights, such as improper rig-up procedures and premature tool deployment. To enhance predictive capability, a Markov Chain Monte Carlo (MCMC) model was developed to generate probabilistic forecasts of future NPT distributions, preserving 90% of total NPT variance while reducing computational cost by focusing on the most critical risk categories. Additionally, a depth-indexed risk score heat map was generated to visualize the expected contribution of each risk category, enabling more informed decisions in well construction planning, tool selection, and resource allocation. This study represents the first technical application of an MCMC-based probabilistic risk model for NPT analysis in the Utah FORGE unconventional geothermal field. The framework provides high-resolution uncertainty quantification, prioritizes actionable risks, and offers a scalable, transferable solution for both geothermal and hydrocarbon drilling operations. By integrating this approach into well planning and execution, operators can reduce NPT, control drilling costs, and make data-driven operational decisions with quantified uncertainty.