Lessons Learned from the Habanero EGS Project (Hogarth & Holl, GRC Transactions 2017)

Provenance: Operator-side post-mortem (Hogarth Energy Resources; University of Queensland) in GRC Transactions 2017. The best-documented commercial EGS failure: technically working, economically dead. Retrieved via research agent; full text extracted with pdftotext.

Project outline

• 2000–2016, Geodynamics Ltd., Innamincka Granite, Cooper Basin, South Australia. Six deep wells; rock 230–264 °C; gradient 31 °C/km. Hundreds of millions of AUD spent.
• 1 MWe-class binary pilot plant ran 160 days in 2013, generating only enough to power the reinjection pump, camp and ancillaries.
• Verdict: “despite the technical success of the project, it failed to satisfy key economic measures and the entire project is now in the process of being abandoned.”

Lessons

1. “Hot rocks are not dry.” The granite fault was overpressured by ~34.5 MPa (5,000 psi). Habanero-1 kicked, lost 250 m³ of heavy mud, needed a 69 MPa BOP and 15.2–15.6 ppg mud; every later well had similar problems.
2. One fault, not a fracture network. Almost all flow in all four wells came from the Habanero Fault (~6 m intensely fractured zone, 1 m cavern, dipping 10° WSW). Most natural fractures are closed (Rittershoffen analogue: <3% of observed fractures permeable). The planned “3-D volume of interconnected joint sets” did not exist.
3. Stimulation needs pre-existing open, critically stressed structures. Jolokia-1: 591 m of open granite held at 69 MPa surface / 116–122 MPa downhole (above minimum principal stress) for 14 days, yet injection averaged 1 L/s with 73 microseismic events — “there were no open fractures with high slip likelihood in the entire 591 metres.”
4. Seismicity is intrinsic. First events at Habanero-4 occurred only 2 MPa above baseline pressure; the fault is critically stressed. Small-volume stimulations (2.2–2.5 ML water) raised productivity 5 → 25 kg/s/MPa. “Induced seismicity is related to long-term hydraulic enhancement and is unavoidable in a successful stimulation.”
5. Circulation stayed pilot-scale. 15.7 kg/s max at 212 °C (2008–09, 560 m spacing, 61,000 t circulated); 19 kg/s max at 215 °C (2013, 690 m spacing, 165,000 t). Commercial EGS targets are 50–100+ kg/s per well. Output was thermally stable during tests, but a single-fault reservoir meant recoverable energy far below plan.

Key quotes

• “However, despite the technical success of the project, it failed to satisfy key economic measures and the entire project is now in the process of being abandoned.”
• “Nevertheless, injection rates averaged only 1 L/s (0.4 bbl/min) and only 73 seismic events were detected during the stimulation.”

Relevance to the cooling wiki

Habanero is the counter-case to Fervo: same “engineer a reservoir in hot granite” pitch, but relying on natural fracture networks and open-hole bullhead stimulation. Fervo’s later success with cemented casing and propped multi-stage fractures is best read as a response to exactly these failure modes. It also shows that a 230 °C resource is worthless as drive heat if it cannot be circulated at scale.