During production of hydrocarbons, the formation is subjected to increasing levels of effective stress, resulting from the reduction in pore pressure. In weak, but consolidated sandstones, this can lead to shear failure in the rock surrounding the perforations and the borehole. Sand production in weakly consolidated formations is generally assumed to be a two-step process, with the shear failure being the first step and the transport of the sand out of the perforations and up to the surface being the second step. Existing sand production prediction models have focused on predicting the onset of sanding by predicting the drawdown at which failure of the formation will start. A further development are models, which try to predict the total volume, which can be expected to be produced, by assessing the geometrical extend of the failed zone. These volumetric estimates of sand production are often based on rock mechanical models, which predict the extent of a yielded zone, using various constitutive material laws. This assumes that all material from a yielded zone will be produced as sand. Combining a model for a volumetric estimate of sand failure with a model for transport of solids in a fluid, might eventually lead to a realistic sanding rate prediction. Existing sand production prediction models tend to assume that the formation disintegrates into sand grains the moment yield or failure occurs and that these grains are taken up by the fluid flow at the time when failure occurs. However, in reality the failure-characteristics of sandstone and the pick-up of the failed material by fluid flow are a very complex process and very little research effort has been dedicated to the morphological description of rock failure.
Using an endoscope camera, it was possible to observe the failure behavior and sand transport during sanding experiments in different types of sandstones. Two different failure behaviors of perforation-sized cavities were identified in these experiments. These failure patterns resulted in very different shapes of the failed zones and during the experiments very different amounts of sand were produced and at different rates. It can be assumed that this behavior can also be found downhole and we concluded that the different failure behaviors of sandstone could be a geological reason for the different severity of sanding problems in various oilfields. Using image analysis of thin sections the different failure behavior could be linked to petrographical rock properties and a predictive model of sandstone failure characteristics was developed. This paper describes the development of the model and the application of this model to formations from an oilfield in the Norwegian Sea.
Sand production is a common and very damaging problem in hydrocarbon production from clastic reservoir rocks. Over the last decades a lot of research by service as well as production companies has been devoted to developing models for sand production prediction. In terms of the sanding problem clastic reservoirs can be divided into three groups, according to the strength of the reservoir rocks, which is directly related to their risk for formation failure and sanding. The first class is reservoirs with very low strength rocks, either unconsolidated or very weakly consolidated. In these reservoirs it is generally assumed that sand free production can only be achieved by installing sand exclusion completions like sand screens and / or gravel packs. At the upper end of the strength scale are reservoir rocks with such a high strength that formation failure is not induced even by large drawdown and /or depletion. In between is a range of reservoirs with intermediate rock strengths, which can be produced sand free if certain precautions are applied, but where formation failure can occur and sand production can become a problem. This is the group of reservoir rocks where sand production prediction models are most useful and mostly aimed at. These types of reservoirs are mostly completed with perforated wells without further sand control measures, if the sanding risk allows to do this. In this paper we will focus on this category of weakly consolidated rocks, with a UCS range of about 500 to 3000 psi (3 - 20 MPa).