Sand production is a serious problem in many petroleum basins worldwide. It costs the oil and gas industry tens of billions of dollars annually. Additionally, uncontrolled sand production results in costly workover operations and, in some cases, can result in environment hazards due to the release of hydrocarbons associated with catastrophic wellhead and/or production equipment failure. On the other hand, if an unnecessary sand exclusion completion scheme is adopted in one well it will cause oil/gas production impairment and loss of oilfield revenue. Rock strength plays a main role in the analysis of sanding onset prediction. Regarding that subject, the in situ stress and the reservoir depletion are also important players. This article presents a sanding onset probabilistic analysis that takes into account the uncertainties associated to the determination of rock strength and in situ stress over the predicted bottom hole flow pressure at sanding onset.
In order to determine the sanding onset at a sandstone reservoir, the rock strength will be estimated in three distinct scales: small (uniaxial tests), middle (scratch test) and large (wellbore logs and correlations). According to this procedure, it is possible to scale up the rock unconfined compressive strength from lab to field. In parallel, the in situ stress was estimated considering leakoff tests, minifracs and wellbore logs. Finally, an analytical model is used to estimate the stress around the perforated tunnel. As design criterion, it is not permitted that the maximum hoop stress surpasses the formation effective strength (U). This mechanical rock property is related to the cavity strength registered in a hollow cylinder test.
Using this kind of methodology, in a pore pressure vs. bottom hole flow pressure diagram, failure probability contour maps for sandstones were generated for two of Petrobras' ultra-deep water fields. For the first analyzed field, previously to the analysis itself, wells have been completed with fracpack technology. After performing an analytical and experimental study using the proposed methodology, two wells (gas and oil) were completed in a conventional way. Up to this date both wells are producing without sanding. Throughout the article these wells will be called "Well A". The second field is still at the appraisal phase and wells have been drilled in order to estimate rock mechanical properties and in situ stress state with the purpose of minimizing the proposed methodology uncertainty. Based on this analysis, a long term well test will be performed without any sand exclusion equipment inside the well. Henceforth this well will be named "Well B".
The proposed methodology is an assemblage of an analytical simulation of stress around cavities and an experimental effort to estimate the unconfined compressive strength based on lab tests and wellbore logs in conjunction with empirical correlations.