Sand Production Prediction: A New Set of Criteria for Modeling Based on Large-Scale Transient Experiments and Numerical Investigation

摘要

Using novel physical model experiments and numerical analyses, a set of criteria are proposed that can be used to determine the onset of sanding and its severity in terms of rate and duration.Experiments on large-scale hollow cylinder specimens were conducted involving real time sand production measurement under various conditions. Synthetic sandstone was used in the experiments whose strength properties were comprehensively determined. A numerical approach was used for simulating the proposed experiments. The material behaviour was simulated using an elasto-plastic stress-strain relationship. The model simulates interaction between fluid flow and mechanical deformation of the medium in predicting sand production. The criteria considered for sanding are more comprehensive than conventionally used and included modeling strain softening of the material accompanied with shear bands formation as well as tensile failure. In the post disaggregation phase, additional features are considered including allowing for the removal of the disaggregated material and making the necessary adjustments to the size and properties of the domain under consideration. Hence, the model is considered to be suitable for time-dependent analysis of the rock as it undergoes disaggregation and production induced by depletion, drawdown, and water-cut. By relating the sand rate to production over the life of a field, the tool can be used in assisting with the completion and operation design of the wells.The model shows good agreement with experimental results in terms of rock deformation and sand rate. It predicted the onset of shear failure and the subsequent strain softening initiated from the cavity face and propagated inside the medium. Strain softening proved to be the main mechanism for material disaggregation. The good agreement between thenumerical and experimental results under comprehensive and complex conditions bodes well for its field applications.