Factors controlling the stability of proppant in propped fractures were identified and investigated experimentally. The results indicate that the absolute size, distribution and type of proppant may affect stability, and hence the likelihood of proppant flowback. The extent of embedment of the proppant into the rock determined by closure stress and rock hardness) was found to play a key role in stabilising the pack. in addition, channels, which may form due to proppant settling before fracture closure, were found to significantly reduce stability. These are, therefore, favourable sites for proppant flowback. A numerical study has been used to determine lengths of unpropped fractures which may be maintained open due to the rock stiffness. The implications of these results for well productivity and the understood mechanisms of proppant flowback, in the field, are discussed. Introduction The flowback of proppant from hydraulically fractured wells is of significant concern to the industry. In addition to problems of valve, line and choke erosion, there is the potential for a loss in near wellbore fracture conductivity. Furthermore, expensive separation facilities are required to filter out proppant from the hydrocarbons, and these may necessitate manning the rig. While substantial work has been carried out to identify means of reducing the amount of proppant flowed back, far less effort appears to have been directed towards understanding the mechanisms which determine whether or not a given well will back produce proppant. Hence it is difficult to decide a priori what preventative measures should be taken. The uncertainty has been concisely put by Daneshy (1) who said of techniques to reduce flowback "None of these methods works all the time, and no criteria are available to know when to do what". Various methods have been devised to reduce proppant flowback. These include resin-coating the proppant, installing mechanical screens and modifying the completion design. Resin-coated proppants have been used successfully in certain environments. P. 569^

Summary Small drilling-fluid losses provide an accurate means of detecting conductive natural fractures and pinpointing potentially productive zones. We outline improved interpretation techniques that allow losses as small as 0.5 bbl to be used reliably as fracture permeability indicators. We compare results from four case histories with conventional fracture-detection methods. Introduction Natural fracture permeability can be both a help and a hindrance to reservoir performance. For instance, a conductive fracture system may provide the necessary permeability to drain a high-porosity but low-permeability oil-saturated matrix, or it may lead to early breakthrough of water and gas within production wells. Which scenario applies to any individual reservoir depends on the characteristics of the fracture system and how that permeability is in communication with the matrix. Natural fracture permeability is usually very variable spatially and often difficult to characterize. Of critical importance during reservoir appraisal is the ability to distinguish induced fractures from natural fractures. Even more important, however, is the ability to distinguish conductive natural fractures from ones that do not contribute to rock mass permeability. However, most natural fracture detection techniques do not clearly differentiate between fractures that allow fluid flow and those that do not because they do not measure fluid-flow properties directly. Core description, for example, can characterize fracture porosity but is unable to distinguish relatively isolated porosity from porosity creating high permeabilities. Detection and characterization of fracture permeability is further complicated by its susceptibility to damage from the invasion of drilling-fluid solids. During well testing, fracture permeability often remains plugged by these solids, and thus it is often impossible to distinguish damaged conductive fractures from impermeable ones. For these reasons, naturally fractured reservoirs are frequently regarded as troublesome to appraise, and misconceptions abound about their character. One such view is that sandstones of moderate matrix permeability (10 to 30 md) do not, at depth, contain significant natural-fracture permeability, because it is thought that any fluid flow within a fracture system would be dwarfed by that through the matrix. This means that certain naturally fractured prospects may never be considered as such, and conductive fractures go unobserved because their detection is not requested. Nevertheless, production levels of thousands of barrels a day can flow from very short naturally fractured intervals.

Hydraulic fracturing is being applied to progressively higher permeability formations. In many cases, productivity improvements are controlled by the achievable fracture conductivities. An analysis of the effects of non-Darcy flow in high rate oil and gas wells shows that in gas wells, fracture conductivities are dominated by non-Darcy effects and that effective conductivities are non-linearly dependent on proppant coverage. Laboratory tests to investigate non-Darcy flow effects in proppant packs have been performed at rates much closer to field conditions than hitherto reported; both single and two-phase flows are considered. Dry gas tests have demonstrated that there is a change in pressure loss behaviour and flow regime at high rates. Three types of test involving gas and water flows have been performed. First the effects water-saturated gas flowing in a proppant pack at residual water saturation were compared with results on the same pack with dry gas. Two series of experiments were then performed to examine the effects of mobile water in the gas stream. Pressure losses are shown to be very sensitive to the water rates, and the field implications are examined. Finally field data from tests performed on high rate, hydraulically fractured, gas wells are examined. Perceived fracture conductivities are related to estimated proppant coverage, and it is shown that non-uniform vertical distributions of proppant may have a large impact on the overall fracture flow capacity. INTRODUCTION With the major fields in the North Sea and on the North Slope now being on decline, the industry is being forced to consider marginally economic developments in which the performance of individual wells is of major singnificance. As a result, hydraulic fracturing is being applied to formations with far higher permeabilities than would formerly have been considered appropriate. In higher permeability formations, the productivity increases achievable by hydraulic fracturing are strongly dependent on the effective, or flowing, fracture conductivities that can be attained1. Increased fracture lengths are of little benefit in the absence of adequate fracture conductivities.

Discussion - No abstract available.

Preliminary results of an experimental programme to determine the significance of dual phase flow in propped fractures are presented. Gas-water flows are considered. The results demonstrate that, even in cases of minimal water production, effective permeabilities may typically be reduced by a factor of 2, or more. In severe cases of water production, permeabilities may be less than 20% of their single phase values.