Abstract:
Sand production from oil-producing wells is a severe problem in the petroleum industry, mainly associated with weak and ultra-weak unconsolidated shallow formations and is an undesirable process of producing sand along with crude oil. Its better understanding and prediction lowers the operational costs and reduces the ecological impact that sand production causes when producing oil. Since the problem has been known for several decades, there are some analytical models available for steady-state flow, which fail to predict time-dependent flow, such as well shut-ins in this study.. The well shut-in is the process when a well is intentionally stopped producing for a certain period due to some needs: changing downhole equipment, workovers, well testing, cleaning sand screens, etc, and restart at a similar flowrate after being brought back to production. Therefore, the main objective of the current work was to rebuild well production in laboratory conditions to experimentally investigate sand production patterns during multiple well shut-ins. Several sand production experiments on large sandstone specimens were conducted using the High Pressure Consolidation System (HPCS) at different overburden stresses (3000 kPa-5000 kPa) and pressure drawdowns (500 kPa-1850 kPa). It was observed that, additional sanding is expected after every well shut-in in a diminishing mannerat a certain ratio between the strength of the material and the combination of the applied stresses. The preceding investigation on the mechanical behavior of the material was conducted by the means of laboratory experiments, such as triaxial shearing tests, one-dimensional consolidation tests, and a supplementary DEM study. The obtained parameters such as inherent shear strength, failure angle, stress-strain curves were also used as inputs for the prediction model’s adaptation for multiple well shut-ins and revealed the most possible material failure mode around wellbore, which reasoned the adaptation of the viscogranular plastic flow model for multiple well shut-ins using Fast Fourier Transform. It was found that the adapted model reasonably predicts the magnitudes of sanding rates and their damping behavior both for experimental and well data.