Understanding fractures and fracture networks is absolutely vital for determining the best places to drill and for pinpointing sweet spots. Fracture characterization is also important for optimizing completion and production, both in vertical and horizontal wells. In mudrocks and shale plays, understanding fractures as well as the geomechanical properties is an important part of the well design.
Welcome to an interview with Stephen Sturm, whose work with fractures and fractured reservoirs spans many of the world's shale plays.
--What is your name and your experience in oil and gas?
Stephen Sturm, I have close to 36 years of oil and gas related experience. My background is petrology and stratigraphy. Until recently, I was the Technical Team Lead for Schlumberger’s Borehole Interpretation Group in Denver that focused on integrated fracture and geomechanical characterization leading to completion design and efficiencies in unconventional reservoirs. I think we processed and interpreted well over 1,500,000 ft of image data from almost every reservoir in the United States during the last 14 years.
I began my career with RPI as a research geologist working on multi-subscriber regional studies of hydrocarbon play trends within western basins and basin-wide databases. Following the acquisition of RPI by Intera Information Technologies, I had the opportunity to work on projects in Kuwait, Norway, Chile, Venezuela, Mexico, China, Indonesia, and Siberia. In 1996, Intera was acquired by GeoQuest (a division of Schlumberger). From 1996 – 2003, I was a technical lead geologist and project manager on projects ranging from GRI Tight Gas Assessments in Wyoming to proprietary studies identifying restimulation and recompletion candidates in the Greater Green River Basin, Wyoming. During the first 20 years of my career, I had the great fortune to be able to describe well over 250,000 feet of core from reservoirs all over the world, as well as spend quite some time out in the field mapping exposed reservoir rocks.
--What is your experience with fractures in reservoirs?
My knowledge of fractured reservoirs is largely a result of my exposure to a vast library of image data coupled with acoustic data, as well as observations from core and outcrop data. The volume of data that has been captured in the subsurface tells a very compelling story that contradicts many of the industry and academia philosophies of ‘naturally-fractured reservoirs’.
The data largely indicates that there is an inverse relationship between natural fracture density and production. Fractures need to be recognized as ‘failures or flaws’ that compromise seal integrity ranging from the laminae to formation scale.
Most of the natural fractures encountered in the subsurface are relatively small, discrete features that are poorly connected laterally and vertically. In order for a fracture to increase permeability, it must behave geomechanical ‘laws’, which are that it needs to be oriented close to S_Hmax, and cannot have a fracture dip of less than 70°. Otherwise, horizontal and vertical stress will result in collapse. Rarely, are conjugate fracture sets encountered that are from the same event.
--What are some of the lessons learned in general?
I think the most important question to ask is what is the origin of the fracture/fracture system. Second, what does the morphology (height, shape, orientation, density, and occurrence within the sequence) of the fracture(s) tell us. Lastly, the relationship between stress and failure is critical. Stress considerations regarding drilling and coring, far field stress, and the effects of uplift on basin margins, all tell a compelling story and are commonly misinterpreted.
What I have learned is that almost all fractures in the subsurface fall into three categories;
1. Natural fractures originating from internal pore-fluid expansion in kerogenous sequences, or ‘expulsion fractures’. These are almost always oriented to S_Hmax at the time of ‘expulsion’
2. Natural fractures or faults related to external stress, or tectonics. These are localized close to the failure point, often related to rapid increase in dip magnitude. Chaotic fracturing close to a fault is common.
3. Fracturing related to overbalanced drilling, or drilling-induced fractures. These tell the most compelling story related to geomechanics, mud loss and predictability of well stimulation.
Another lesson learned is that the joints that are observed on basin margins and uplifts typically have little in common with the natural fracture population in the subsurface. Most of these are de-stressing features from rocks being uplifted several thousands of feet. Unfortunately, quite a bit of effort has been expended on these features that has resulted in poor exploitation strategies.
--What are some of the main misconceptions about fractures, both induced and natural?
There are several misconceptions that are profuse in the industry. The first being “the rocks are too tight and can’t produce without open-natural fractures”. This is usually encountered when operators have promoted a play utilizing regional linear features and or seismic attributes, and not evaluating matrix data or understanding perm-to-gas. I don’t know how many times I have heard ‘we drilled the well between fracture swarms’ after telling a client there were no fractures in the image data. Unfortunately, the core warehouses in the US are filled with similar wells.
The second misconception is that fractures are ‘open’ and filled with hydrocarbon. This would require a few conditions that are extremely difficult to satisfy. Primarily, that the internal fluid pressure within the fracture would have to exceed minimum horizontal stress as well as the adjacent reservoir (matrix) pressure. Secondly, because natural fractures are calcite-lined or filled ‘envelopes’ segregated by high-stress clay layers, there is limited means to charge them from an external source to create differential pore-fluid stress. Incomplete calcite and/or silica cement within the fracture may result in propping open the limited space, but it also reduces the fracture permeability.
The last misconception is that conductive fractures are ‘open-natural’ fractures. Drilling mud is conductive, hydrocarbons are resistive. Conductive fractures are a result of mud pressure exceeding formation pressure and are usually a result of overbalanced drilling or encountering weak rocks. There are excellent correlations between large aperture drilling induced fractures and extensive mud losses. Additionally, acoustic data confirms these features extend an appreciable distance from the wellbore to justify mud loss volumes.
--What are some of the main “need to know” facts about fracture networks?
The important takeaway is that fractures are more likely to negatively affect reservoir performance than enhance it. I think of fractures as ‘failures’ or ‘flaws’ in the reservoir. Before these fractures were mineralized they provided pathways for hydrocarbons to migrate, or escape their encapsulated pore-pressure cell. Where high-density fracture networks are encountered, reservoir pressures are generally lower – in some cases normal to below normal.
Next, there is quite a bit of discussion in literature that stimulating natural fractures will result in ‘complexity’ and greater SRV (Stimulated Rock Volume). This may be true, especially with natural fractures oriented normal to S_Hmax. However, dense calcite-filled fractured intervals have lower pore volume and post-frac closure may not have beneficial results.
In contrast, induced (conductive) fractures at the wellbore usually are very good predictors of hydraulic fracture geometry, frac conductivity and height growth. Occurrence of large aperture, drilling induced fractures are a characteristic of both low pressure zones and mud loss flags. These will also be flagged on acoustic logs as having higher DT Shear or have higher anisotropy.
Lastly, targeting fracture networks or fault zones with a drill bit usually has negative consequences. These rock sequences typically are lower stress and subject to failure. With almost certainty, drilling challenges and trying to run pipe will occur.