Avoiding Proppant Flowback in Tight-Gas Completions with Improved Fracture Design

Abstract

Abstract In this paper we describe the possibility and benefits of incorporating the stability analysis of the proppant pack during the design stage of hydraulic fracturing treatments. After the well is treated and cleaned up, the flowback of proppant from a fracture-treated formation is highly undesirable for several reasons, including possible damage to the wellhead and flowlines, operational complications, and last but not least, decrease in well-productivity. The issue has been studied on an empirical basis and the most important factors have been determined. Nevertheless, qualitative models suggested so far seem to work only under limited conditions and currently there is no clear methodology for predicting the occurrence of proppant flowback while designing the treatment. In this work we review previously suggested prediction methods and analyze the proppant flowback patterns experienced in 24 South-Texas tight-gas completions. As a result of the study we conclude that the predictive power of the available models is not satisfactory and the implications for "proppant stability control agents" are not based on convincing evidence. A new semi-mechanistic model is proposed that shows reasonable agreement with both laboratory and field data. A methodology is suggested to incorporate the proppant flowback prediction at the fracture design stage. The suggested methodology is based on the concept of "minimum necessary departure from optimality to satisfy technical constraints", in this case the constraint being to keep the likelihood of proppant flowback under a certain threshold. Introduction Since the late 1940's, there have been more than a million fracturing treatments performed in the United States1. In a regular fracturing operation, some flowback of proppant occurs right after the treatment is completed. This stage is referred as the "clean up" phase and is unavoidable since under-displaced proppant will remain in the wellbore after the operation. In this stage, personnel and equipment are usually still in place and any produced solids can be easily handled. In contrast, flowback during the production phase of a fractured well can undermine the potential benefits of the stimulation treatment and represent operational complications. Along this paper we will refer to this latter situation as "proppant flowback". The problems associated with proppant flowback are the following:Local loss of fracture conductivity, which generates a reduction in the potential benefits of a hydraulic fracturing treatment.Damage to the equipment. (Consisting in abrasion to valves, tubing, surface pipelines and other equipment). The first and most obvious approach in dealing with this problem consisted in applying operational rules of thumb. The most common of these techniques focused on maintaining the production rates below a critical value that was determined as critical to initiate the flowback of proppant in a specific well. The evolution of forced closure technique, that is closing the fracture rapidly in order to trap the proppant grains in a "uniform distribution" generating a more stable fracture has been proven usually necessary but not sufficient. It has been reported2 that these methods do not work all the time, partly because their justification ignores the actual mechanisms that create instability in the proppant pack. Numerous flowback control additives have been offered by various service companies3. Unfortunately, there is lack of evidence that they work under conditions of high formation closure stresses. The use of resin coating in particular, has been reported unsufficient in some cases4. On the other hand, some experimental studies5 have helped deliniate the mechanisms behind proppant flowback. However, the available predicting models rely on empirical correlations6,7,8 and may be less reliable as the application conditions deviate from the conditions of laboratory experiments on which the correlations are based. The goal of this work is to summarize available information and improve the predictive power of the models.