Shale Rock Mechanical Characteristics: Integration and Validation Using Crossplots

"Proper shale characterization is crucial in shale gas drilling, exploration, completion and production. Unlike porous sandstone and carbonate reservoirs, the challenge with shales comes from their micro- to nano-darcy permeability ranges and the anisotropy of their mechanical and hydraulic properties. Most porosity-dependent logs are not sensitive enough to capture the true behavior of shales. Consequently, mineralogy and other strength and stiffness-related properties must be determined to characterize magnitudes and variances in shale mechanical and flow properties. In addition, these properties must pass quality assurance/quality control (QA/QC) practices to give confidence in their use in subsequent log calibrations, during geomechanical modeling, or in fracture design models. This paper presents a technique to integrate shale properties that are measured by various methods, and a methodology to perform QA/QC on the data before application.Shale properties originate from various sources, including downhole and laboratory measurements. Physical samples for laboratory testing are usually limited to a few depths. Lab tests are often costly, and replicating in-situ conditions can be difficult or impossible. In addition, results frequently show wide scatter, reflecting inherent depth variability and lateral heterogeneity. Wireline logs provide nearcontinuous data, but rarely yield direct measurements of rock strength properties. Although geomechanical and fracture models use log data, calibrating it with reliable and previously verified lab-measured properties is important. This paper presents a process to perform QA/QC on lab-measured data using various crossplots.Ten mechanical, microstructural and physical properties are graphically displayed to visually characterize shale strength behavior in ambient and in-situ conditions, whether measured or modeled statically or dynamically. These properties are often sourced from different departments. The properties are integrated but are individually subject to the previously mentioned uncertainties. Pairs of property values are crossplotted and compared with correlations from the literature or previously established with qualified data from related formations. Conformance with these correlations gives confidence in the data, whereas outliers are flagged as questionable. Repeating the process with questionable data crossplotted against other parameters for the same depth highlights the property with the greatest uncertainty. The determination of its value is investigated to locate the source of the inconsistency and establish a more reliable value. Where this is not possible, the user has a rational basis for rejecting that property value."