Sand production is a complex phenomenon in oil wells where sand grains or chunks are detached from the formation during loading processes and transported to the wellbore by the flowing fluids, either oil, water or gas. This article presents a complete set of different laboratory tests using 3D printed rocks for sand production evaluation and phenomenological understanding. The use of the 3D printed samples reduces sample-to-sample variability, and increases the testing capabilities when performing destructive tests. This small variation in properties maximizes the level of confidence and repeatability in the results, which is required for the sand production assessment program presented in this article. The performed tests include UCS (Unconfined Compressive Strength) and TXT (Triaxial) tests which allow sample's mechanical and hydraulic characterization as well as an assessment of the evolution of properties with plastic strains. Additionally, TWC (Thick-Walled Cylinder) tests are performed over hollow cylinder samples to determine the TWC parameter. This parameter is useful for sand production modeling and validation of samples' mechanical behavior to anticipate the expected sand production results. Finally, complementing the aforementioned tests, a novel geomechanically advanced laboratory test for sanding quantification in hollow cylinder samples is presented. This laboratory test includes the application of axial and radial stresses while performing radial fluid flow through the sample and the measurement of produced sand. The laboratory testing program proposed helps to evaluate the behavior of printed samples and similar natural samples in a sand production scenario.
Sand production commonly occurs in poor or uncemented formations. Some sand-related problems include the reduction in oil production and the flow equipment damage (down-hole or at surface conditions). However, when there is high risk (including economic impacts) of producing sand, methods for sand control and mitigation are required.
Several ways to infer sand production potential have already been explored such as dimensionless relationships, total compressibility to shear modulus ratio (Khamehchi & Reisi, 2015), neural network applications (Kanj & Abousleiman, 1999) and analytical solutions. Unfortunately, these methods do not predict the sand production levels required to properly design the development plan for a wellbore or reservoir.