Improving Closed-Loop Geothermal Predictions with CMG STARS

Closed-loop geothermal systems offer a pathway to unlock heat from non-hydrothermal (“dry”) formations, but their feasibility depends heavily on how heat transfer is modeled.

A comprehensive study conducted by JOGMEC (Japan Oil, Gas and Metals Corporation) evaluated multiple closed-loop configurations using four simulation approaches:

• CMG STARS 
• COMSOL Multiphysics 
• GEOSIM (CRIEPI geothermal simulator) 
• Analytical models

Outcome: CMG STARS predicted production temperatures up to 20-40°C higher than other tools by capturing thermally-driven natural convection in the surrounding formation. This highlighted how modeling assumptions can significantly influence performance predictions across closed-loop geothermal systems.

Significance of Closed Loop Geothermal in Japan

Japan has significant geothermal potential but faces constraints:

• Limited hydrothermal reservoirs 
• Regulatory and environmental restrictions 
• Large volumes of untapped hot dry rock 

Closed-loop geothermal systems address this by:

• Circulating fluid in a sealed well system 
• Extracting heat via conduction and convection

This offers significant advantages:

• Geographic flexibility: applicable wherever subsurface temperatures are sufficient, regardless of rock permeability or hydrothermal fluid presence
• Reduced scaling risk: no contact between working fluid and formation water eliminates the mineral precipitation and flow decline issues endemic to open hydrothermal systems
• Regulatory advantages: lower environmental impact and reduced conflict with hot spring resources and national park regulations
• Lower induced seismicity risk: no hydraulic fracturing required (unlike Enhanced Geothermal Systems / Hot Dry Rock approaches)

Operational Context

The study evaluated multiple closed-loop configurations to understand how different system architectures respond to subsurface conditions and modeling approaches.

• CMG STARS → Full reservoir + wellbore + thermal physics 
• COMSOL → Pipe-flow + conductive heat transfer 
• GEOSIM → Axisymmetric geothermal model 
• Analytical → Simplified heat transfer

The study classified closed-loop and related configurations into two primary architectural families, each with multiple sub-variants.

U-Loop Systems

• Multilateral wells connected at depth 
• Large heat exchange area

Coaxial Systems

• Single vertical well with inner/outer flow paths

Table 1. Full taxonomy of closed-loop configurations evaluated. 

Model Overview

All tools were run with a shared base case representing a hypothetical Japanese geothermal site. Rock properties were defined for two representative lithologies encountered in Japanese geothermal drilling:

Table 2. Rock property inputs for base case simulations. 

Key Results
1. For configurations where heat transfer is primarily conduction-dominated, all simulation tools showed strong agreement.

• Production temperatures: 
  • ~145-175°C depending on flow rate 
• Cross-tool variation: 
  • Typically within ±10°C 

2. For configurations where formation-fluid interactions become significant, simulation results showed greater sensitivity to the underlying physical assumptions.

• STARS: ~130-150°C 
• COMSOL / GEOSIM: ~100-110°C 
• Difference: 20-40°C

Insight: CMG STARS predicted production temperatures 20-40°C higher than those of other tools by capturing thermally driven natural convection. 

3. LCOE (Levelized Cost of Electricity) Analysis
LCOE was calculated over a 30-year project life using standard discounted cash flow methodology, with operating costs, fixed asset taxes, and decommissioning costs included. 

Table 3. LCOE base case estimates.

Insight: Table 3 indicates that drilling complexity, not thermal performance, is the dominant cost driver for U-loop systems, whereas coaxial systems are more sensitive to thermal modeling accuracy. Accurate temperature prediction, such as that provided by STARS, directly influences these economics.

Key Takeaways

• CMG STARS provides a more comprehensive representation of subsurface physics, enabling engineers to evaluate closed-loop geothermal systems with greater confidence across a range of configurations.
• U-loop systems: 
  • Conduction-dominated → tools agree 
• Coaxial systems: 
  • Convection-sensitive → tools diverge 
• Ignoring formation convection leads to: 
  • Lower predicted temperatures 
  • Misleading economic conclusions 

Best Practices

• Select simulation approaches that represent both wellbore and formation-scale heat transfer 
• Ensure models capture natural convection effects where formation permeability is non-negligible 
• Use tools capable of predicting realistic production temperatures under long-term operation 
• Align economic evaluations with physics-based simulation outputs, not simplified assumptions

Conclusion

This study demonstrates that:

Simulation methodology plays a critical role in evaluating closed-loop geothermal systems. Approaches that incorporate full subsurface physics, including formation-fluid interactions, can provide a more complete understanding of heat recovery potential and support more informed engineering and investment decisions.