The Under-Appreciated and Often Neglected Geomechanical Model

by Charles Hager, stimulation and reservoir engineer

Today I’d like to talk about the necessity for using geomechanical models in the upstream oil & gas industry.?? I’m going to go on a limb to suggest that it’s probably the most valuable pursuit for most every subsurface endeavor we pursue, yet the most neglected.

The geomechanical model is the skeleton that is the framework for so many of our most sophisticated technical endeavors.?? Allow me to explain…

?One approach to generate a geomechanical model is to use a set of bulk density, compressional sonic, and shear sonic logs.?? We typically integrate, or sum up, the bulk density log to surface to determine the overburden or vertical stress.? Next, we incorporate pore pressure. This is usually constrained by using a combination of direct pore pressure measurements and the application of normal compaction trendlines. In a final step, we constrain the horizontal stresses (minimum and maximum), by incorporating poisson ratio (stress transference), and Young’s Modulus (stiffness) and adjusting the tectonic strain components to calibrate the geomechanical model. ?It is somewhat straightforward in homogeneous isotropic systems, throw in anisotropy and that’s a completely different talk. Armed with this framework, there are a plethora of uses for this model:

• Geology: ?It has been my experience that a geomechanical model will often be your best friend for picking geologic layers!? Yes, I just said that. Often times, Poisson Ratio and Young’s Modulus values will quickly show you geologic transitions more readily than mineralogy, gamma ray, or other log responses.?
• Petrophysics: ??How often do we run lab tests on cores?? Whether we are looking for natural fracturing descriptions, stress dependent permeability and porosity, or even tri-axial tests for rock properties, the petrophysicist will often use the geomechanical model for calibration and framework purposes.
• Geophysics: ?When we perform seismic surveys, we are bouncing sound waves off of reflective surfaces, or passing them through layers of various impedance values.? Pre-Stack Migration is your best example where a detailed model is critical.? It is a sort of geomechanical model that allows us to build these synthetic examples to predict and match measured surveys.?
• Drilling: ??We use this critical model for designing mud weights and casing string setting depths.? We also use geomechanics to control the hole from collapsing in as well as from breaking down.? It is a critical tool for well design.? Furthermore, it is even more critical during horizontal well drilling where hoop stresses and overburden stresses may cause collapse of your borehole.??
• Hydraulic Fracturing:? ?The geomechanical model will show you where your injected fluids and proppant are most likely to go. ?As with the path of least resistance, the lowest stressed intervals will prefer to take your pumped fluids, and the contrast of bounding layers will dictate how well a stimulation stays “in zone” or grows height.? Couple the model with fluid injection rates and viscosities and we have the basics for fracturing simulation.
• Microseismic Surveys:? Lastly, the microseismic survey is exceptionally critical to the best survey possible.? For each seismic event, the sound wave must travel through different layers of varying travel speed.? It is the best detailed geomechanical profile, with the most detailed impedance layering that will lead to the best interpreted results.?? Each event recorded at the listening receiver must be traced back to where the source should have occurred.? Each recorded event requires a lot of calculations with a lot of sources for error.?

While this is a quick 6 areas which are all dependent on the geomechanical model, I’m sure that there are more that could be discussed.? With this, how often is a full sonic and density log suite captured and used as the basis of design?

Below is an example of a chalk reservoir.? The geomechanical model makes it much easier to generate detailed and reliable layering than with the initial log data.