Numerical simulation of sand deposition and accumulation in a horizontal well

Horizontal wells are currently finding ever-increasing application to recover oil and gas worldwide. In Canada, one of the main applications of a horizontal well is to recover heavy oil in Saskatchewan and Alberta. These reservoirs are predominantly marginal, having thin pay and bottom water. They are, in addition, poorly consolidated; as a result, recovery operations in these reservoirs are often susceptible to sand production. Sand production, deposition, and accumulation in a horizontal wellbore is a serious problem, as it plugs and partitions the horizontal well into sections--thus negating the principal advantage of a larger reservoir-wellbore contact area for horizontal wells. Despite the significant impairment to well productivity, particularly in the case of horizontal wells, the mathematical treatment of this complex transport process has been outside the realm of conventional petroleum engineering. The present work is a systematic examination of this complex flow problem.;Of the 145 references cited in this study less than one-third are from petroleum engineering publications, many of these being experimental investigations of cuttings transport and settling velocity of fracture proppants. None of the published papers gives a satisfactory treatment of the transport and deposition of sand in a horizontal well. Nor is there any experimental data available in the literature to test a given mathematical model. Only one paper--discussed in depth in this work--touches upon the problem of suspension transport in a different context (fill-up of a fracture). This is shown to be inadequate for the problem at hand--sand deposition and accumulation in a horizontal well. Instead, a new mathematical model is developed and solved numerically. This model takes into account the buildup of the sand deposit inside a horizontal well as a function of oil viscosity, particle size, flow velocity, and other parameters. Limited (storage) memory capacity of the computer used in this study (Macintosh Quadra 800) results in the simulation of only the early stage of the deposition and solid bed expansion processes.;The numerical model is tested using field data including oil viscosity, flow rate and well diameter. The particle size is magnified, but still within a reasonable sand size range, to enable certain basic transport phenomena to be observed. Simulation runs are performed to evaluate the influence of key variables on the deposition and deposit bed expansion processes. The numerical results agree with experimental observation for the formation of sand ripples in a shallow-water environment, and make intuitive sense. Implications of the simulation results for field conditions are discussed. Scaling criteria for the transport process are also derived for possible use in future experimental studies, as well as selected simulation studies.