Pressure Maintenance in the Groningen Gas Field (Part III)

Pressure Maintenance in Bureaucratic Purgatory

An evaluation by NAM to identify possible mitigations against production-induced seismicity through pressure maintenance was released in 2016, about 25 years after the first earthquake. This technical report from the operator provides a negative advice regarding Groningen Pressure Maintenance (GPM), primarily due to the high cost of nitrogen injection. 

While the GPM report is thorough, it is clear there is a technical preference for nitrogen, while then conveniently blaming the high cost of nitrogen gas compression to render the project infeasible. The 118-page report discusses water injection for less than one page.

Below I give a few examples of Dutch gas fields where operating the field in a narrow reservoir pressure window has helped minimize earthquakes. Then, I explain why water is the best practical and economic contender for pressure maintenance in Groningen, and it should be given more consideration by the experts.

Pressure Maintenance in Dutch Gas Fields

Over the last three decades, NAM has changed out some producing gas fields to gas storage. These smaller gas fields around the Netherlands also target the Rotliegendes sandstone. While the goal of these projects was not to maintain (reservoir) pressure, cyclic injection / production operations in these storage fields – to accommodate peak winter demand – has caused the reservoir pressure to fluctuate within a relatively tight window.

For example, pressure maintenance in the Norg gas storage field with yearly pressure cycles provides some confidence that small injection pressure changes do not instantly trigger earthquakes. As shown in Figure 7, Norg earthquakes stopped after keeping the reservoir pressure within a tight pressure window between 250 and 325 bar during yearly injection-production cycles.

Figure 7: Reservoir pressure history in the nearby Norg Gas Storage Field. While injection triggered an earthquake when reservoir pressure increased by 70 bar, no repeat earthquake has occurred while cycling reservoir pressure within a ~75 bar operating window.

The Bergermeer gas storage field experienced earthquakes with a similar magnitude distribution as the Groningen earthquakes during its production phase. Now, as the field operates within a narrower pressure window, earthquake magnitudes have receded to much lower levels.

In a recent overview of earthquakes in Holland by Muntendam-Bos et al., it is stated that in cyclic underground gas storage, “Magnitudes of these prospective earthquakes are expected to remain below the level observed during depletion.”

A Water Injection Proposal for Groningen

Water injection is not the go-to methodology for secondary recovery of a gas field, and as such the suggestion here is unconventional. It has, however, been evaluated and used before in other gas fields. Water injection into the Groningen Gas Field has some of the following benefits:

• Fresh water is available in enormous quantities in a country with 80 cm (31 inches) of rainfall per year;
• Groningen has several pumping stations like Gemaal Rozema that are already pumping vast quantities of fresh water into the sea to keep the land, that is mostly below sea level, from flooding. As such, fresh water is mostly a waste product in the Netherlands;
• The existing infrastructure of canals provides easy access to fresh water at all Groningen Field well sites;
• Water is already injected in the field near Termunten, so NAM already conducts this on a small scale;
• Water is already present within the Rotliegendes sandstone, and as such, no new materials are introduced downhole. This is important because new materials may impact shear plane slippage and earthquake generation differently;
• A major benefit of water over any compressible gas is that it provides hydrostatic pressure along the depth of the well, which eliminates the need for much horsepower to pump the material into a well. This gives water a massive economic advantage.
• As a hybrid solution, water can also be mixed with nitrogen to create a foam through use of a surfactant. The combination of both still provides the benefit for the hydrostatic head in the well through use of water and avoids expensive gas compression.

The water injection strategy that I propose for GPM has the following specifics:

• Water injection from ~100 wells along the reservoir’s flanks, below the Gas-Water Contact (GWC; see Figure 8). It is key to inject only into the water-bearing zone of the reservoir for each injector to minimize gas entrapment. Injection is from the flank, aiming to gradually push gas production from the periphery to the elevated central areas of the field over the years. The well design will have to be modified to ensure injection strictly below the GWC, and bottomhole pressure gauges will have to be installed to constantly measure the differential pressure generated from the hydrostatic column as compared to the reservoir pressure;
• Fresh water from canals will be treated in a water treatment plant at every injection site with bleach and iron inhibitor. We assume use of 0.35 gpt (0.035%) is sufficient to clean the water from any bacteria;
• Injection rates per well should be limited to about 1.3 cubic meter per minute (~8 bpm) to stay below frac initiation pressure for about a 30-meter (100-ft) depth interval along each vertical well. A simple radial flow equation shows that these rates keep pressure at the injection well within about 10 bar (145 psi) above the reservoir pressure. Rates may have to be further limited to remain within this narrow pressure threshold. Acidizing sweeps may have to be scheduled regularly to remove skin to help maintain injectivity;
• Average reservoir pressure should remain about the same as today, around 70 bar (1,000 psi) by ensuring that yearly production volumes are matched with yearly injection volumes. Some local variability is inevitable, and at each injection well we expect at least a 10-bar (145-psi) increase. Experts should evaluate the tightness of this operating window. With the suggested number of injection wells and the limited water injection rate per well, this limits yearly gas production from the field to about 5 Bcm per year. Although that may not sound like a lot, it would provide a sustainable rate that could be increased as the GWC in each well rises and as more wells make the switch from producer wells to injector. At this rate, the field could produce for another 50 – 100 years.

Figure 8: SE-NW cross-section of the Groningen Gas Field, showing the Gas-Water Contact at a depth of about 3,000 m. The author proposes water injection from wells along the reservoir’s flanks below the GWC.

In the GPM report, water was quickly dismissed as an injectant. The following problems were anticipated with the injection of water, ultimately deeming it technically unfeasible (text in italics is directly from the report):

Firstly, a voidage replacement scheme with water would require a water injection rate of 1-2 mln m3 /d, which if injected under reservoir fracture conditions would need 100-250 wells. From a geomechanical point of view, fracturing is not desirable. Injection at pressures below fracture conditions would need 650-1,300 injection wells, which is about two to four times the number of wells drilled in the Groningen field to date (350 wells) and therefore impractical in view of the surface requirements for drilling locations, rigs required, and water pipelines. Sourcing this amount of fresh water is also not possible and, to prevent souring of the reservoir, utilisation of seawater would require world-scale water treatment facilities. Water that is introduced to the reservoir would be back-produced, which would likely require well interventions to install velocity strings or gas lift to bring the well fluid to surface. Furthermore, as the produced water will be saline, a significant extension of water handling and disposal capacity would be needed on the Groningen clusters. 

My response: I propose to use only existing wells at the periphery and no new wells. Rates are to be kept below frac rates. This limits the injection volume significantly and will limit the production of the field. That’s one of the keys – lower injection rates lead to lower production rates (~5 Bcm/year), which would allow the field to be produced for 50+ years depending on the achievable injection rate and number of wells available. Instead of the suggestion to inject 1-2 million m3 in the field daily, my suggested plan only assumes 0.2 million m3 daily water injection volumes, and does not require new wells or frac’ing of existing wells.

Secondly, water injection at large scale requires water to be injected close to faults, which very likely increases the risk of earthquakes.

My response: I propose to inject water below GWC, not introducing new materials, but elevating the GWC over time.

Thirdly, given that a water front moving through the reservoir would trap gas at saturations of about 26% (trapped gas saturation), introducing a water injection scheme would be at the cost of at least 26% of the remaining gas, assuming a perfect water flooding of the reservoir (i.e. without bypassing of free gas). 

My response: The looming shutdown of the field represents 100% gas entrapment. However, I expect gas entrapment to be significantly lower if you inject below the GWC and if it is sourced from the flanks at low rates, avoiding the creation of a vertical "water front". 

A Simple Cost Analysis

In the analysis below my goal is to only provide ballpark cost estimates for the big-ticket cost items in the table to help us see the big picture.   

Rambaran et al. (2018), find at least a factor of 10 cost increase between water injection and N2 injection, primarily due to the much higher cost of compressors vs pumps, and the much higher fuel cost to operate compressors vs pumps. This is mostly due to Mother Nature providing “zero-cost” hydrostatic pressure through water.

For reference, at €240/MWh, the remaining value of all recoverable gas (~500 Bcm) in Groningen is about €1 trillion (€1,000 billion), setting yearly production revenue for 5 Bcm at about €10 billion. Last Friday, Dutch TTF gas futures were €188/MWh, still valuing the resource at €900 billion. How do cost of these secondary recovery techniques compare?

In our simplistic cost model for water injection in the table below, understand that most prices listed are based on US prices, and they may be different in the Netherlands. Startup costs are expected at about €130 million, mostly for well design changes (blocking injectors wells to intervals below GWC, installing gauges), centrifugal pumps with some backups and a water treatment plant for storing and filtering water, and treating it with UV light to kill bacteria. In rounded numbers, the yearly cost for water injection would be about €50 million, or less than 1% of the yearly production revenue. This yearly cost includes basic chemicals to treat the fresh water and diesel to run the centrifugal pumps.

The 4-stage compressors needed for nitrogen compression are about 100x (!) more expensive than the centrifugal pumps needed for water injection, adding significantly to the initial investment. But it does not stop there: the yearly cost of nitrogen compression is expected to be closer to €1 billion, mostly due to much higher diesel consumption.

As an alternative to nitrogen, CO2 could also be considered. However, it has even more drawbacks than nitrogen, especially since it is not available at scale. Both N2 and CO2 could be added to some sites where it is available in conjunction with water injection, adding surfactants to the water to inject a foam. This would have the benefit of a significant hydrostatic head, eliminating the need for compression.

Below is a quick comparison of various injectants, assuming 5 Bcm of yearly Groningen production, catering to about 15% of the Netherlands’ yearly consumption:

A Way Forward?

My fear is that the state and operator’s recent support for structural upgrading of homes and businesses provides them with a false sense that further production of the field is now on the table. The state and operator’s successful strategy to do very little to address the root cause of earthquakes while continuing production, albeit at some lower level, may have instilled the belief that this strategy can continue while addressing the current energy crisis. This would be a mistake. 

It is likely that there will always be some “emergency” to arm-twist Groningers to accept production of the gas under their feet. Today it is Russia and the Ukraine invasion. Tomorrow, maybe it is an ultra-cold winter that requires Groningen gas to help prevent energy-poor folks from freezing.

Even without a crisis, consider that fossil fuels, including gas, will be around for a lot longer than most people think today – certainly longer than the unrealistic mandated phase-out of natural gas in 2050. If the world "transitioned" from 87% fossil fuels to 83% fossil fuels for its primary energy consumption in the last 30 years, how can we somehow think it is reasonable to assume that it will go from 83% to 0% in the next 30 years? That's not a transition. Why can’t we be realistic about how long we need natural gas? 

If there is a need for production from Groningen, and we believe there will be, then what number of earthquakes and cumulative earthquake magnitude is “acceptable”? If you agree with me that that answer should be as close to zero as possible, then some pressure maintenance is inevitable. 

The Dutch State and the NAM owe it to Groningen to devise a realistic long-term production plan. Any continuation of production should include a profit-sharing clause to compensate Groningers for their troubles. Transparency about long-term plans by the operator and government is a requirement to regain their social license to operate.

Water injection is a technical and economical answer for earthquake minimization while producing the remaining gas from the Groningen Field.

I hope experts can evaluate these suggestions, tweak them, suggest a water injection field trial, and learn about the further benefits and limitations of this technique as it applies to Groningen. 

Thank you for technical guidance and cost estimates, Thomas Doiron, Eric Bengtson, Joel Siegel and Kyle George.