Lost circulation due to induced fracturing is a costly problem in the petroleum drilling industry. After a discussion of the processes of fracture initiation and propagation, this paper reports a case from the Norne field offshore Norway. A production well suffered lost circulation during drilling at a mud density significantly below what had been perceived as the local fracture gradient. Striking similarities were found between the lost circulation incident and the results of an extended leak-off test that had been performed at the same depth in an offset appraisal well just months earlier. Both datasets indicated that the formation at the relevant depth had a lower-than-prognosed minimum in situ stress, but a high fracture initiation threshold. This additional barrier to lost circulation was irretrievably lost once a fracture had initiated and penetrated a certain distance away from the borehole. Based on this understanding of the nature of the problem, and an updated set of operational constraints, the drilling operation progressed to the prognosed total depth of a well section that could otherwise have been lost. Future wells will be based on an updated set of pore pressure, collapse pressure, minimum stress and fracture pressure gradients that were established after the incident. The case illustrates that the conventional methods of determining the pore pressure and fracture gradient may carry considerable uncertainty. Introduction Lost circulation due to induced hydraulic fracturing is a costly problem in the petroleum drilling industry. Drilling fluid densities and casing setting depths are often dictated by the fracturing pressure of the penetrated formations. Particularly in offshore production wells, where cost-cutting and extended-reach ambitions are important drivers, well plans often leave a narrow operating window between borehole collapse and lost circulation. Doing such optimization in a safe way requires a thorough knowledge of the local fracture gradient. The fracture gradient, or limiting pressure gradient for avoiding lost circulation, is traditionally determined by a leak-off test (LOT). In this test the borehole immediately below the casing shoe is pressurized until fluid leak-off becomes noticeable in the pressure-volume curve, indicating that a fracture has been created [1]. The leak-off pressure (LOP) is interpreted as the first deflection from a linear pressure-volume curve. If the local fracture gradient is sufficiently well known at a certain casing shoe depth, a simpler formation integrity test (FIT) is often performed instead. In the latter test the borehole is pressurized to a predetermined value without creating a fracture, thus verifying cement integrity. A compilation of LOT's (and even FIT's) is often used to generate local and regional depth trends, and thus the predicted fracture gradient for future wells. However some pitfalls exist in using leak-off tests to generate a fracture gradient curve. Individual tests may be difficult to interpret, when no clear or unique deflection point exists. Test data is often recorded manually at a coarse sampling rate, disallowing raw data scruitiny. Some bias towards higher interpretations may even be introduced by the drilling team's eagerness to drill ahead. LOP's from a group of neighbouring wells are often scattered, giving considerable room for subjective interpretation of the local trend. Theory of Fracturing and Lost Circulation In order for lost circulation by induced fracturing to occur, two criteria will have to be fulfilled:The fracture will have to initiate at the borehole wall and grow in the near-well region.The fracture will have to propagate out of the near-well region and grow to a large surface area. In this distiction the near-well region is taken to be the volume of rock whose stress state is affected by the presence of the borehole, typically 1–2 hole diameters into the formation. Theory of Fracturing and Lost Circulation In order for lost circulation by induced fracturing to occur, two criteria will have to be fulfilled:The fracture will have to initiate at the borehole wall and grow in the near-well region.The fracture will have to propagate out of the near-well region and grow to a large surface area. In this distiction the near-well region is taken to be the volume of rock whose stress state is affected by the presence of the borehole, typically 1–2 hole diameters into the formation.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractSubsea re-injection of drilled cuttings has been applied on the Åsgard field to handle oil wet cuttings. For various reasons, oil based drilling fluids have been used in all 12 ¼" and 8 ½" sections at Åsgard. Discharge of oil wet cuttings is not permitted. Cuttings have to be re-injected to the formation or sent onshore for treatment. Onshore treatment and logistics, are expensive and involve a large emission of CO2. If reinjection can be performed, big bag handling operations are significantly reduced. Furthermore, it is difficult to have reliable logistics to bring cuttings onshore. Therefore, cuttings re-injection is the optimum solution for this operation, where the preferred injection point is below the 20" casing shoe.Re-injection is straightforward if cuttings can be reinjected into a formation underneath a large sand formation where the liquid phase of the slurry can leak off. At Åsgard there are no such large sands. Therefore, cuttings have to be re-injected into shale formations. Since the fluid does not leak off sufficiently fast in shale there is a danger that the annulus may be exposed to a too high pressure for a long time. If too large a fluid volume is injected at a time, there is a danger that the fluid may propagate as a fluid bubble to the surface.Leakage to surface has been observed four times at Åsgard. It has been recognized that the cementing results must be better than normal to hinder leakage to surface if cuttings are re-injected into shale formations without sand layers. This paper describes the leakages in detail. It focuses on the necessary improved drilling and cementing precautions to hinder leakage. The paper also describes a technique with alternating re-injection and static periods to make the liquid phase leak off into minor sands and shale, preventing the formation of too high an annulus pressure and hindering the formation of a fluid bubble propagating to the surface.

No abstract

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractSubsea re-injection of drilled cuttings has been applied on the Åsgard field to handle oil wet cuttings. For various reasons, oil based drilling fluids have been used in all 12 ¼" and 8 ½" sections at Åsgard. Discharge of oil wet cuttings is not permitted. Cuttings have to be re-injected to the formation or sent onshore for treatment. Onshore treatment and logistics, are expensive and involve a large emission of CO2. If reinjection can be performed, big bag handling operations are significantly reduced. Furthermore, it is difficult to have reliable logistics to bring cuttings onshore. Therefore, cuttings re-injection is the optimum solution for this operation, where the preferred injection point is below the 20" casing shoe.Re-injection is straightforward if cuttings can be reinjected into a formation underneath a large sand formation where the liquid phase of the slurry can leak off. At Åsgard there are no such large sands. Therefore, cuttings have to be re-injected into shale formations. Since the fluid does not leak off sufficiently fast in shale there is a danger that the annulus may be exposed to a too high pressure for a long time. If too large a fluid volume is injected at a time, there is a danger that the fluid may propagate as a fluid bubble to the surface.Leakage to surface has been observed four times at Åsgard. It has been recognized that the cementing results must be better than normal to hinder leakage to surface if cuttings are re-injected into shale formations without sand layers. This paper describes the leakages in detail. It focuses on the necessary improved drilling and cementing precautions to hinder leakage. The paper also describes a technique with alternating re-injection and static periods to make the liquid phase leak off into minor sands and shale, preventing the formation of too high an annulus pressure and hindering the formation of a fluid bubble propagating to the surface.

Several challenges are experienced with wellbores when drilling in shale above and within reservoirs. These are especially prevalent in deviated wells and wells through depleted intervals caused by some years of production. Drilling the well to total depth of the wellbore section might be easy. However, in some cases depending upon formation properties and drilling conditions it is a big challenge to come out of the hole and subsequently run casing or liner. It is of great importance to take the right actions regarding recommended downhole parameters like mud circulation rate, RPM on drillstring and pulling speed. Some of the wellbore aspects to discuss are: unstable hole, small margins between mudweight and collapse pressure, what is the right technique to use with the drillstring, flow rates and finally oil based mud versus water based mud. In case of a marginal downhole situation, it would be beneficial to test out some effects in a laboratory beforehand. Such questions were discussed and became an issue within a joint industry project on borehole stability in shales run at SINTEF Petroleum Research (SPR). A modified test cell was designed and used with the purpose of systematically testing these effects in the laboratory. This paper describes how this equipment was designed and tested in order to achieve reliable laboratory results as well as results from initial work on outcrop shale. The main objective was to a certain extent to visualize some effects in a laboratory and conclude with some good recommendations regarding drilling procedures on the rig offshore. Thus for testing some of the down hole effects, a long period of qualification tests was gone through. This was important before real tests could be done on shale at simulated downhole condition. Testing was performed on hollow cylinder samples. The initial results show that it is possible to run such laboratory simulations. It is possible to a certain extent to visualize some of the effects with varying flowrates and RPM on a drillstring with stabilizer. The laboratory tests so far suggest that high RPM on a drillstring can have destabilizing effects on a wellbore through a section with shale that has been exposed to very close to collapse conditions, calling for optimized drilling procedures in situ.