KEM-24b Effect of pressure maintenance by fluid injection on seismic risks Groningen, continued

In 2022, the KEM-24a study was published, which investigated the effect of fluid injection on seismicity, using the Groningen reservoir as an example. Building upon previous studies and available data on the Groningen reservoir, a model was developed to investigate the effect of fluid injection on seismicity. The results of this study were, however, inconclusive. The model was incapable to simultaneously mimic historic changes in the reservoir pressure and observed seismicity. As such, the forecasting potential of the model could not be validated. One reliable finding of this study was that injection of N2 is more efficient in affecting reservoir pressures compared to CO2. Therefore, in this follow-up only N2 injection is considered. Studies into geothermic systems have shown that the injection of relatively cool fluids into relatively warm reservoir may trigger seismicity. Injection of a cold fluid causes a portion of the reservoir to cool down. This temperature reduction is accompanied by thermal contraction of the reservoir rock, leading to a stress drop and potentially activating nearby faults (e.g. De Simone et al., 2013; Jeanne et al., 2017; Pandey et al., 2018). So, it is important to consider this temperature effect, in addition to pressure effects, when evaluating the potential for fluid injection into depleted reservoirs. Despite all efforts, there is no definite answer on the net effect of fluid injection in depleted reservoirs on seismic risk. Therefore, the KEM-panel has recommended to conduct a follow-up research. The State Secretary of Economic Affairs and Climate Policy decided to follow this recommendation and therefore KEM-24b is initiated.

Are there examples of fluid injection in reservoirs similar to the Groningen reservoir? Worldwide fluid injection has been applied in several fields. Are any of those similar to Groningen in size and/or reservoir and overburden characteristics? What can be learned from these cases in terms of important criteria, factors, etc.?

 What are possible injection scenarios for mitigating seismicity once production has stopped? What scenarios could be applied to mitigate seismicity after production? Would those scenario’s focus on minimising the pressure difference across the field, overall increase of pressure everywhere in the field, or are there other options? How much fluid, taking N2 as an example, should be injected to have an effect on reservoir pressures and the seismicity? Is that technically feasible?

How can the SHRA model be adapted to include fluid injection? The current version of the SHRA model is not capable to include pressure and temperature effects of fluid injection. In what way should the different parts of the model be adapted to be able to study the pressure and temperature effect of fluid injection?

 What is the effect of fluid injection on the overall seismic risk? What is the effect of the injection scenarios defined in research question 2 on the seismic hazard and risk profile of the Groningen field compared to a scenario without injection. How do the results.

RESEARCH QUESTION

This project was set up as a follow up to KEM-24 project (which was entitled “Effect of pressure maintenance by fluid injection on seismic risk”). At the end of that project, a conclusive judgement on the value of fluid injection for reservoir pressure maintenance of and reducing seismic hazard could not be made. However, indications for the potential (only) positive effects of gas injection, especially nitrogen, were found. So, the objective of this follow-up project KEM-24b was to establish whether a large-scale injection of nitrogen gas into the depleted Groningen gas field can decrease the number of earthquakes, and consequently reduce the seismic hazard and risk. In other words, at this stage, only potential beneficial effects of nitrogen injection were studied and reported.

RESEARCH PROJECT REPORT

There were four research tasks in this project. Those tasks and the results reported can be summarized as follows.

A literature study, in which the available knowledge on fluid-injection-induced seismicity and risk mitigation strategies was compiled, with particular attention on lessons that can be learned from existing cases (e.g., wastewater disposal, gas storage, geothermal heat production) and are relevant to the Groningen case. The study showed that the injection of nitrogen into a depleted gas reservoir with the aim of reducing seismicity has not been done anywhere in the world. Nevertheless, the magnitudes of seismic events observed in cases of fluid injection that are relatively analogous to the Groningen case (in terms of expected possible geomechanical failure mechanisms, injection depth, and reservoir properties) have been relatively small (magnitudes ≤ 1.7).  

Identification of scenarios for nitrogen injection. In total, eight scenarios were considered. The unit of gas injection rate was selected to be 1 ZB = 1,58 billion m3/year (equivalent to the production capacity of Zuidbroek nitrogen factory). In three scenarios, the injection rates were set to 1 ZB. In five other scenarios, injection rates of 10 ZB were considered. The latter scenarios, although unfeasible based on current nitrogen production capacities, were considered to determine the upper bound of nitrogen injection effect and to illustrate some mechanisms associated with re-pressurization and their impact on seismicity. Results show that nitrogen injection at rates currently possible will result in a noticeable pressure increase (up to 90 psi over 30 years of injection) compared to no injection, and will lead to a reduced pressure gradient across the field. With 10 times larger injection rates, the pressure increase will be much faster and much higher (up to 700 psi over 30 years of injection).

Seismic hazard and risk analysis, which considered the potential (only) beneficial effects of gas injection and the resulting pressure changes on the overall seismicity using the SHRA Groningen. Results show that the injection of nitrogen can lead to a reduction of predicted seismicity compared to no injection case. The reduction is significantly larger for higher injection rates. Correspondingly, there is a significant reduction in the seismic risk and the number of buildings that do not conform to the safety norm decreases significantly. The extent of possible reduction depends on the injection rates, spatial and temporal injection patterns, and the duration of injection. However, negative effects should be added for conclusive results on this matter.

Exploring necessary adjustments to the seismic source model. The need for improving the current seismic source model was investigated, with particular attention on the ability to investigate potential negative effects of nitrogen injection. It was concluded that the source model should be updated by adding a non-linear rate-type compaction branch, so that potential consequences of temperature reduction due to injection can be investigated.

In summary, the injection of nitrogen seems to have a potential for reduction of overall seismicity. However, seismic hazard and seismic risk cannot be established on this project alone, as the potential adverse effects of injection were not considered. Determining the feasibility of risk reduction requires a comprehensive study and cost-benefit analysis, using suitable models that account for nonlinear deformation, thermal and compositional effects, so that potential negative consequences can be fully quantified. Also, aspects related to the required volumes of nitrogen and the potential for its production and technological requirements need to be investigated.

The project has been evaluated by the KEM scientific research panel.

KEM RESEARCH PROJECT EVALUATION

It is concluded that, while this study has been limited in scope, all research questions have been effectively addressed.

The study has investigated beneficial effects of nitrogen injection only. It has worked with simple injection scenarios with very simple spatial distribution of injection wells. The model that is used here had been developed for the simulation of gas production, and not gas injection into a depleted gas reservoir. Therefore, some potentially significant effects, such as compositional and thermal effects, have not been included.

Determining the feasibility of risk reduction requires a comprehensive study and cost-benefit analysis, using suitable models that account for nonlinear deformation, thermal and compositional effects, so that potential negative consequences can be fully quantified. Also, aspects related to the required volumes of nitrogen and the potential for its production and technological requirements need to be investigated.