A number of robust predictive methods for establishing sanding thresholds have been developed over the past decade. Having identified when the onset of sanding occurs, recent research efforts have focused on determining the rate at which sand will be produced once these thresholds are exceeded. In this paper a new analytic model for predicting the rate of continuous (steady-state) sand production is described. This sanding rate model is consistent with the threshold prediction model, and utilizes as its basis the non-dimensionalized concepts of loading factor (near-wellbore formation stress normalized by strength) and Reynold's number (a function of permeability, viscosity, density and flow velocity). Interpreted this way, the results of laboratory sand production experiments are used to derive an empirical relationship between loading factor, Reynold's number and the rate of sand production. A second empirical sand production ‘boost factor’ incorporates the effects of water production. The derived model is compared with field data from a total of six wells from two fields, for a wide range of flowing conditions. The predictions are a good match to the field data, typically overestimating the field-measured data by a factor of less than four. However, as the model is for continuous sanding only, this degree of overprediction is considered acceptable for field application, as it provides some compensation for short-lived transient sand production at rates higher than steady-state values.

You can access this article if you purchase or spend a download.