Injection in Shale: Experience on the Norwegian Continental Shelf
Waste injection in shale has been performed on the Norwegian Continental Shelf (NCS) for more than 15 years. Techniques have been developed that allow wells to dispose of several million barrels into individual shale domains with matrix permeability in the nanodarcy range and no permeable layers.
Waste injection in shale, with matrix permeability in the nanodarcy range and without the presence of any permeable layers, has been performed on the Norwegian Continental Shelf (NCS) for more than 15 years. To avoid leakages to the seafloor using this method, techniques have been developed that allow wells to dispose of several million barrels into individual shale domains, with vertical propagation of the disposal domain less than 1,000 ft above the injection point. Recently, use of frequent 4D interpretations of seismic surveys shot over a permanent sensor array allowed detailed domain mapping and independent dynamic monitoring.
Changes of North Sea regulations in the mid- to late 1990s made the seabed disposal of oily cuttings and other waste from drilling and production impossible without the use of significant topside cleaning systems. Meanwhile, the development of fields required increasingly complex wells, resulting in the almost systematic use of oil-based mud for efficient drilling. This led to a dual challenge for the industry, with an increased amount of waste generated and the difficulty of discharging it. Options were then reviewed between shipping onshore and treatment or reinjection into the subsurface with cuttings-reinjection (CRI) wells (a typical design of a CRI well in the Ekofisk Field region is outlined in the complete paper). Most operators concluded that the CRI option was much safer and more economically favorable than shipping to shore and onshore treating and disposal.
A survey of all CRI wells on the NCS noted that 110 had been or were operating, with approximately 30 of them still active at the time of the survey. This confirms the attractiveness of CRI wells as a means of waste disposal, at least as far as the NCS is concerned. Unfortunately, the same survey also revealed that 14 of these 110 wells had suffered from a leakage to the seabed, thus illustrating the risks associated with CRI operations.
An outcome of the investigation was the discovery that a vast majority of the leakage-to-seafloor cases occurred in wells where shallow sand bodies were lacking or where their quality was very poor. This justifies the general strategy traditionally adopted by Norwegian operators (i.e., avoiding the upward migration of hydraulic fractures with high-permeability sand bodies); this emphasizes the risks associated with the lack of good-quality “blanket” sands above the injection point.
Challenges and Opportunities for Waste Injection in the Ekofisk Field Region
The Ekofisk field is situated in the southern region of the Norwegian North Sea, close to the British and Danish sectors in an area where the Utsira sand is not present. The main challenges are the amount of waste generated every year at a regular rate (i.e., approximately 850,000 bbl/yr to be injected) and the absence of any significant permeable layer in the overburden of the field that can act as a barrier for the vertical propagation of the hydraulic fractures created during injection. One certainty is that the injection of such high volumes cannot be performed through a single discrete fracture, because it would necessarily extend all the way to the seafloor.
The main opportunities, though, involve the quantity and quality of monitoring systems used in the Ekofisk field, which are almost unique in the world. The systems include frequent seafloor bathymetry, regular logging of all wells, permanent downhole gauges, use of fiber optics along the tubings, and permanent seismic arrays on the seafloor, allowing both passive seismic monitoring and active seismic shooting every 6 months. A side benefit of this exceptional array of techniques is the possibility of detecting any anomaly caused by waste reinjection and the possible migration of fractures toward the seafloor.
The decision by the operator to inject the waste into the Ekofisk overburden resulted from a balanced approach using the opportunities offered by the monitoring systems to detect any abnormal vertical propagation of the waste-disposal domain and possible leakage that could have been triggered by the lack of a good-quality sand layer in the Ekofisk overburden formation. The environmental and economic benefit of this balanced approach has been demonstrated clearly by the more-than-15-year history of cuttings injection into the overburden in six wells in the greater Ekofisk area with no abnormal upward migration of the waste-injection domain, with no leakage to the seafloor, and with the capacity of some wells to accommodate up to 5 million bbl of waste fluids.
Round-the-Clock Monitoring and Real-Time Analysis
The analysis of past leakage cases shows that data acquisition alone is insufficient to ensure injection confinement and the avoidance of surface leakages. Consequently, a joint-industry project (JIP) on injection safety was undertaken to develop a software system capable of performing automatic round-the-clock monitoring and real-time analysis of injection data for any given well. The architecture of this system is shown in Fig. 1.
The system has two key functions. One is to send an alarm or “flag” in case of anomalies in the injection process, and the other is to extract as much information as possible from the injection records. Both functions are performed simultaneously and benefit greatly from each other (e.g., the real-time extraction of the reservoir pressure in the injection domain allows round-the-clock calculation of the injectivity index and the detection of abnormal increases, which could be indicative of upward migration and eventual leakage).
The round-the-clock monitoring-and-automatic-alarm system fulfills three main roles, corresponding to three levels of alarms.
The first level simply ensures that all sensors are active when the injection process occurs and that adequate measurements are recovered effectively. The second level ensures that the procedures carefully established by the operator are followed effectively by the personnel performing the job. The third level detects and reports anomalies in the performance of the CRI well, which could correspond to a loss of injection efficiency or could mean an upward migration of the disposal domain, leading to leakage.
It has been possible to dispose of millions of barrels of waste fluids in a limited volume of shale through a 10-ft-long perforation interval without any undue vertical extension of the disposal domain. The systematic extraction of information from all injection cycles has allowed quantification of three important features.
1. Injection of volumes of significant magnitude has been achieved by creating a multitude of fractures in the disposal domain instead of having a discrete number of fractures accommodating the injected fluid.
2. The secondary permeability in the disposal domain was repeatedly measured to be in the 0.1-darcy range, while the matrix permeability of the shale under realistic downhole conditions was measured to be between 10–9 and 10–8 darcies.
3. Another important observation is that the pressure inside the fracture network connected to the well is far in excess of the overburden stress and that the fracture network does not show any sign of extension after the well has been shut in, despite this very large net pressure.
From a physical perspective, two mechanisms acting simultaneously can be identified: One impairs the propagation of the primary fracture, and the other favors the opening of secondary fractures.
In the case of very-low-permeability shale such as that above the Ekofisk chalk hydrocarbon reservoir, the important issue is what is happening inside the fracture ahead of the fluid front. The matrix permeability value makes it obvious that no fluid from the rock matrix can be displaced to fill the volume created by the newly formed fracture ahead of the injected fluid. Consequently, this fracture volume ahead of the fluid becomes filled with vapor. This phenomenon is known as cavitation. It corresponds to the partial dehydration of the shale and leads to mechanical swelling when the dehydrated shale is contacted by water.
Here, it must be emphasized that mechanical-swelling stresses can often depend on the level of dehydration of the shale and are mainly independent of shale mineralogy. In turn, this swelling stress applies not only at the tip of the fracture but also over its entire length from previous sudden propagation events and can be viewed as a stress barrier added to the minimum in-situ stress. This added swelling stress explains why the pressure required to propagate a hydraulic fracture in shale can be much higher than the minimum in-situ stress and why no fracture propagation was ever observed during the shut-in periods of the CRI wells. The swelling stresses developed during the fracture-propagation process contribute to the mechanical equilibrium of the system, and the absence of propagation after well shut-in is further proof that the system is at a mechanical equilibrium during shut-in periods, despite the presence of a high net pressure in the fracture network.
Another mechanism governing the creation of the multiple-fracture network observed around the CRI wells in the Ekofisk field is the thermal effect. The fluid injected during the CRI operations in Ekofisk is colder than the formations where the injection takes place. Under typical conditions, the fluid is injected at 70 to 85°F for a virgin formation temperature of 175°F, corresponding to a cooling of approximately 100°F. This means that the formations in contact with the injected fluid will be cooled.
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 170851, “Injection in Shale: Review of 15 Years of Experience on the Norwegian Continental Shelf and Implications for the Stimulation of Unconventional Reservoirs,” by F.J. Santarelli and F. Sanfilippo, Geomec, and R.W. James, H.H. Nielsen, M. Fidan, and G. Aamodt, ConocoPhillips, prepared for the 2014 SPE Annual Technical Conference and Exhibition, Amsterdam, 27–29 October. The paper has not been peer reviewed.