Techno-Economic feasibility of enhanced geothermal systems (EGS) with partially bridging Multi-Stage fractures for district heating applications

The problem

Hot rock a few kilometres underground could heat homes in winter and displace coal. Where the permeability of this rock, the ease with which fluid flows through it, is too low, water cannot circulate. An enhanced geothermal system (EGS) creates flow paths by hydraulic fracturing, the injection of water at high pressure to crack the rock. Drilling and stimulation (fracturing) exceed 75% of the cost of medium- and low-grade EGS, those with poorer geothermal resources (Lukawski et al., 2014; Tester, 2006).

A common design joins two parallel horizontal wells with vertical fractures that cross from one well to the other (Gringarten et al., 1975). Most of the injected water then flows only through the fractures. This limits the volume of rock that is cooled and causes early thermal breakthrough, the time when cold water reaches the production well (Li et al., 2016; Yu et al., 2021). We previously proposed a partially bridging design (Yu et al., 2021). Each fracture connects to only one well, and injection and production fractures alternate along the wells. Water must then cross the rock between fractures, the stimulated reservoir volume (SRV). Our goal in this study was to test whether this design could supply district heating at a profit.

What we did

We modelled the reservoir and wells with the heat and mass transfer simulator FEHM. Two horizontal wells, 300 m apart, at depths of 1 to 5 km were joined by 4 to 20 fractures. Each year had a 120-day heating season and 240 days of shut-in, with the wells closed, over a 20-year life. A binary search found the highest constant heat output, in megawatts (MW), that the reservoir could sustain for 20 years. Our economic model costed exploration, drilling, fracturing (USD 0.75 million per fracture), surface plant, and operation and maintenance, paid for with a loan at 5.48% interest. Income came from the floor area heated, under a Chinese design code for district heating. We converted the heat to the coal and CO2 it would displace.

Because each simulation took several days, we ran 25 simulations in a Box-Behnken design, which varies four parameters together: temperature (100 to 300 °C), depth (1 to 5 km), SRV permeability (0.1 to 10 millidarcies, mD) and fracture number (4 to 20). We fitted a polynomial surrogate model to the results (R² of 0.997) and checked it against 12 further simulations.

What we found

We find that SRV permeability and reservoir temperature are the main controls on profitability in our model. With other parameters at benchmark values (250 °C, 3000 m, 3 mD and 10 fractures), heat output rose roughly linearly with temperature and profit increased. Below 120 °C, or below 0.4 mD, the project did not repay its investment within 20 years. The levelised cost of heat (LCOH) ranged from USD 30 to 250 per megawatt-hour (MWh) across the permeabilities tested.

Deeper reservoirs gave more heat output and more profit, with diminishing returns. Deeper rock can take a higher injection pressure without the fractures growing. Higher pressure drives more flow through the rock. However, drilling costs rise quadratically with depth. The shortest time to profit, about 8.5 years, came at 4000 m.

Adding fractures also gave diminishing returns. This is because more fractures along a fixed well length add entry and exit points without adding rock. Water then takes a shorter path through the rock and is heated less. At 16 to 20 fractures, each added fracture gave almost no extra heat output while stimulation costs continued to increase.

At the Qiabuqia field in China, drill core gave permeabilities of 0.26 to 0.7 mD (Lei et al., 2020). We assumed 218 °C at 3600 m (Lei et al., 2020). In the surrogate model, 14 fractures gave the highest profit. Because the SRV permeability is unknown, we sampled it 20,000 times from a distribution with a mean of 1 mD, assuming that stimulation raises permeability by up to a factor of ten above the core values. The median heat output was 52.4 MW (95% interval 33.6 to 73.2 MW), the median LCOH was USD 46.7/MWh (33.0 to 84.6) and the median profit was USD 110 million (24 to 185 million).

In the most favourable permeability case, the 14-fracture design could heat up to 2.8 km² of floor space, about 28,000 dwellings of 100 m². Over 20 years it would displace nearly 600,000 tonnes of coal and more than 1.5 million tonnes (Mt) of CO2. However, the SRV permeability at Qiabuqia is unknown. In our model, the viability of the project depends on it.

With the paper's conversion factors, the most favourable figures imply a constant output of 83 to 85 MW, and the median output of 52.4 MW would heat about 1.8 km² and displace about 0.97 Mt of CO2. These values were calculated for this page and are not reported in the paper.

Limitations

Our fracture design is idealised and ignores interactions with natural fractures. We treated the SRV as a uniform porous medium with rock and water at the same temperature. Fast flow paths that let water bypass the rock therefore cannot be represented. Our cost models are uncertain and are likely to need updating as more experience is gained.

We considered only seasonal heating. Year-round industrial uses of EGS heat could improve the economics. At Qiabuqia, more reservoir data could reduce the uncertainty in our inputs, and a better choice of simulations could reduce the surrogate model's uncertainty.

Glossary

A binary search finds a target value by repeatedly halving the interval that contains it.

A Box-Behnken design is a set of simulations that varies several parameters together at low, middle and high values, using fewer runs than every combination would need.

District heating supplies heat to many buildings from a central source through a pipe network.

An enhanced geothermal system (EGS) is an artificial geothermal reservoir created by fracturing hot rock whose permeability is too low for water to circulate.

The levelised cost of heat (LCOH) is the lifetime cost of a project divided by the heat it delivers.

A megawatt (MW) is a million watts of power, and a megawatt-hour (MWh) is the energy delivered by one megawatt for one hour.

A millidarcy (mD) is a unit of permeability, the ease with which fluid flows through rock.

A Monte Carlo analysis repeats a calculation many times with randomly sampled inputs to estimate the range of outcomes.

R² is the fraction of the variation in the simulated outputs that the surrogate model reproduces.

The stimulated reservoir volume (SRV) is the rock between injection and production fractures, through which water must flow in our design.

A surrogate model is a fast formula fitted to the results of a slow simulator, used in its place.

Thermal breakthrough is the time when cold injected water reaches the production well and the produced water starts to cool.