Abstract The paper presents a field case based upon a reservoir operated by Saga Petroleum in the North Sea. The reservoir stress path was determined from a series of minifrac and fall off analyses and step rate tests on various appraisal tests and injection wells. The results show a strong heterogeneity of the stress path with a large difference between the heavily faulted Southern part of the field and the more quiet Central part of it. The stress path was further confirmed by a series of mud loss incidents during drilling that are briefly summarised. Furthermore, in the case of two injectors, the step rate tests were performed after a significant re-pressurisation of the reservoir -i.e. 30 and 70 bars respectively. Such tests gave extremely low fracturation pressures compared to what was the reservoir stress path upon depressurisation. In fact, the fracturation pressures measured are consistent with the maximum depletion but not the reservoir pressure at the time of the test. In other words, the total minimum stress in the reservoir seems to be governed by the maximum depletion of the zone and not by the present level of pore pressure. Two wells from another North Sea reservoir will also be presented showing a similar behaviour upon repressurisation. Note that such an observation is consistent with one measurement performed on Ekofisk and published earlier on. These observations of heterogeneous and irreversible stress paths will be analysed and compared to other observations such as the behaviour of surface subsidence during aquifer depletion and repressurisation. In particular, it is shown that the reservoir stress path is compatible with a rigid plastic behaviour of the rock mass that has already been more or less observed in the case of the Venice subsidence. However, core measurements do not seem to obey such behaviour. The difference can then be attributed to scale effects as the stress measurements via fracturing tests affect large rock volumes. As a consequence, the results reported here give an insight about what could be the rheological behaviour of a reservoir rock mass at a very large scale. Possible applications to large scale rock mechanics modeling such as compaction, subsidence, hydraulic fracturing, are then highlighted. Finally, the practical consequences of such a state of affairs in terms of reservoir management are drawn.

The paper presents a series of well tests performed on various North Sea weak reservoirs and will show how the skin factors measured can be interpreted in terms of sand production. In particular, the results presented show that sand production results in abnormally low skins independently of what the perforation conditions were in terms of under or over balance. First, a series of tests were performed on injectors after a very limited amount of back production for clean up. Such tests were injection tests followed by clean up and a very good correlation existed between skin factor and under balance conditions. However, after important sand production, the wells were tested again under injection and all showed very low skin factors. Second a series of four production tests were performed on a well shot over balance and the evolution is clearly one of decreasing skin factor indicating sand production. Third, for a given field, all tests performed with a total draw down - i.e. draw down plus depletion - superior to 70 bars show an extremely low skin as the other show high skins. The balance conditions for such wells were very diverse. Fourth, it will be shown how on wells producing sand, the stimulation of the levels that produce sand leads to dramatic flow concentrations from very limited intervals and various PLT results will be presented. The practical consequences of such results will then be presented and three essential aspects will be dealt with. The perforation balance policy will be discussed in the light of the results presented above. The use of well test as a systematic mean to predict sand production will be outlined. Finally, it will be shown that test interpretation techniques ought to be refined to achieve a bigger understanding of the results in terms of description of the near wellbore area. P. 375

Abstract Stress trajectories around faults have been simulated with a 3-D finite element model, based on 8-model brick elements in five layers. Preliminary simulations imply that stress trajectory deflections in the neighbourhood of fault zones can be expected if there is a large geomechanical contrast between the fault zone and the adjacent rocks and if far field stress anisotropy is small. On the other hand, if the horizontal stress anisotropy is large, significant stress deflections appear unlikely to occur near a fault. Within a softer' fault zone H will be aligned approximately parallel to the trend. Magnitudes of maximum and minimum stress are modelled for soft and hard fault simulations. Significant modifications in stress magnitudes developed in the vicinity of the fault zones. P. 33

Abstract This paper shows how using the Drilling Gptimization Simulator (DROPS) has been applied in the planning phase of two North Sea wells. Both the 12.25" and the 8.5" sections were optimized on Well #1 and the 12.25" section was optimized on Well #2. Geological Drilling Logs (GDLs) were generated from offset drilling data in the same field for both the two wells. Both wells with the different hole sections were simulated with the available bit types and designs. The optimum bit selection and the drilling parameters every meter for the hole-sections were optimized. The hydraulic limitations of the pump and the planned mud weight program were used to optimize the bit hydraulics in conjunction with the bit operating parameters, different pull depths and bit designs. The 12.25" sections showed savings potential of 72.4 percent on Well #1 compared to the last well drilled in the field. This included optimizing bit type, hydraulics, operating parameters and pull depths. Well #2 showed savings potential of 15 percent using oil based mud and 20 percent using water based mud compared to the two other wells drilled previously in the field with oil and water based mud respectively. It was concluded that Well #1 could save $ 1.3 million, while Well #2 was originally drilled close to optimum and only showed a potential savings of $ 0.3 million when using optimum drilling parameters. The optimization of the 8.5" section was more complex because of several coring intervals. The five full hole sections between the coring intervals were optimized. The five subsections were drilled through formations of different length and lithology. The bit designs, pull depths, number of bits, hydraulics and operating conditions were optimized for each section. In addition to optimization of the drilling parameters, simulations for reuse of bits were conducted to further cut the drilling costs. The cost savings on the 8.5" section for well#1 showed a maximum potential for saving varying from 23.4 to 78.8 percent for the different sections. From the simulations it can be concluded that the optimum drilling scenario of the 12.25" and 8.5" fitll hole sections for well #1 would reduce the drilling cost with more than $ 4.5 million. P. 161

Summary This paper presents a new composite drillpipe design that allows for an increase in the drilling reach of high-angle, directional wells. Horizontal reach is greatly limited by resisting friction forces that are created by the interaction between the drillstring and the lower side of the borehole. As the drillstring is advanced further into a highly directional borehole, the resisting friction forces give rise to ever-increasing applied-torque and axial (push)-force requirements. The gravitational forces (i.e., drillstring weight) dominate the buoyancy forces in nearly all modern drillstring designs that are currently used in highly directional drilling operations. These dominating gravitational forces create reaction forces between the drillpipe and the borehole wall. The reaction forces, when multiplied by the coefficient of friction between the drillstring and the borehole, yield tangential surface forces on the drillstring (normal to, and parallel with, the pipe axis). These forces resist the applied torque that rotates the drillstring and the applied axial (push) force that advances the drillstring. The new composite drillpipe design discussed in this paper reduces the specific weight of the drillpipe so that it will have nearly neutral buoyancy in 15.4-lbm/gal drilling muds. With this reduction in drillpipe specific weight, a drillstring with this new design will have greatly reduced gravitational forces and, in turn, greatly reduced resisting friction forces. This design feature allows the applied torque and axial force to be more effective, thus allowing greater extended-reach capabilities. It is estimated that the current extended horizontal-reach approximate limit of 10 km could be increased to 15 km or more with the application of this new composite drillpipe design. Such an increase in reach would have an important impact on the development of oil and gas fields that are economically dependent on advanced directional drilling technologies. This new composite drillpipe technology has been extensively bench and field tested and has been successfully used in short-radius drilling operations. The design features of the proposed light-weight composite drillpipe are discussed and compared to current horizontal drilling operational requirements (using a steel drillpipe). The new light-weight designs can be used with either steel or titanium tool joints (with a filament-wound composite-tube body structure between the tool joints). The drillpipe may be fabricated in either 30- or 42-ft lengths. Engineering feasibility studies have shown that this new light-weight composite drillpipe will be economical to fabricate and should provide significant economic advantages in the development of oil and gas fields requiring extensive directional drilling operations.

This paper gives effective procedures to drill wells in a depleted reservoir. From the fact that depletion result in decreased horizontal stresses and fracture gradients, the working hypothesis is that a depleted reservoir should be drilled by keeping the Equivalent Circulating Density (ECD) as low as possible. This have been achieved by using a low mudweight and applying the downhole pressure sub actively; which imply focus on the combination of hole cleaning, lower pump rates, avoiding backreaming, and careful breaking of circulation before pumps are started. The pressure sub and theoretical calculations show that bottom hole pressure give peak values under the following circumstances: Too quick start of pumps/breaking circulation, too high mudweight, too high mud viscosity, too high tripping velocity, poor hole cleaning, too high pump rates, and pressure pulses. Other factors which may have some influence on the bottom hole pressure are: Well stability problems leading to excess fill and loading of the well, and unstable mud. An additional method to reduce the ECD is to increase the fluid column diameter by installing liner instead of casing or by drilling holes with a larger diameter. Also, an exact calculation of the ECD is usefull. Three cases from three different wells are discussed, and theoretical calculations and observations are compared. We also observed that the fracture gradient of the borehole wall increased with time, which is usefull if it is not possible to go below the minimum fracturation gradient.

Abstract This paper presents the introduction, application and verification of a new drilling optimization approach. During the past decades rate of penetration (ROP) models for all types of drilling bits were developed and verified. The unique software uses these models in a new approach that optimizes the drilling cost on each hole section. The simulation of ROP every foot gives the capability to simulate the effects of all operating conditions, bit design, bit wear and formation properties. All the penetration rate models used can be inverted to solve for rock strength if ROP, operating conditions, lithology percentage and field reported bit wear is known. This imply that if one well has been drilled, and the penetration rates are inverted, a foot by foot strength is obtained, by integrating over each individual bit run until the field reported bit wear matches the calculated. This strength is then used as input for simulation, and it is shown that the drilling cost for one particular 12,25" section can be reduced as much as 45-50 percent by changing bit design and operating conditions. This approach has been verified with North Sea data where the generated rock strength from the 12,25" section has been used as input to simulate the ROP and rotating time. These simulations compare well with the measurements of ROP, made in the two following wellbores drilled in the same field. Simulating ROP using drilling data from a well in a nearby field also gave promising comparisons. Introduction The DRilling OPtimzation Simulator (DROPS) is developed to reduce the cost of future wells based on a Geological Drilling Log (GDL), created from the data collected in a previous well drilled in the same area. The GDL is created using ROP models inverted to calculate rock compressive strength. The simulator has the capability of simulating any combination of operating conditions, bit designs, pull depths, hydraulics, WOB and RPM. The basic idea behind DROPS is to simulate the drilling operation prior to the actual drilling, and to find the optimum cost level. Simulations to verify the software (drill behinds) have been performed using actual field data from the North Sea, provided by Saga Petroleum ASA, well A and B. Data from well C, provided by Statoil, have also been used in the drill behind simulations. The simulated ROP results from the drill behind compares well with the results for actual measured ROP and cumulative rotating time, which are promising for further development and application of the method. Based on the comparison some cost optimizations for future wells in this region have also been performed. The cost reductions obtained indicate potential cost savings of 45-50 percent of total drilling cost per well, compared to the current industry practice in the area. By using this methodology in the planning process the most cost effective drilling program can be obtained.

Abstract Newly developed software has illustrated that the safe mudweight window to avoid borehole collapse and fracturing can be obtained from the direct use of drilling and/or the use of sonic data. Instantaneous or post analysis using inverted drilling operational data like WOB, RPM, flowrate, and mud properties can be done by applying an inverted ROP model to predict rock compressive strength. This rock compressive strength is then used in conjunction with survey, in-situ stresses and rock failure criteria to predict the safe mud weight window. The sonic logs are also used to obtain the rock compressional properties through published correlations. The rock compressive strength is then used the same way as the drilling data approach. This paper gives a solid field example from a North Sea well where both methods are applied and compared. The field case provides proof that the new software is a very good tool for predicting a safe mud weight window to avoid borehole collapse and fracturing while drilling. Introduction During drilling operations it is crucial to maintain the hydrostatic mud weight pressure between the fracture and collapse pressure at any depth to ensure trouble free drilling. The casing program is also dictated mainly from the fracture and collapse gradient. Therefore, the knowledge and use of the correct fracture and collapse gradients can save the operating companies large sums of money. When estimating the safe operational mud weight window two important parameters arethe strength of the rock, andthe stress conditions the rock is subjected to. The question is; how can these parameters be monitored while drilling a well. Today there is no way of measuring compressive rock strength or stresses while drilling. The best method of determining compressive rock strength is by triaxial testing of core samples, but the problem is that these data will then only be available after the well is completed. These triaxial data are then used in conjunction with logs through correlations when planning the next well. This paper describes a decision support system for well planning, and follow-up of well integrity (safe operating mud weight window) during drilling operations. The system, is a Windows based modular system, using Internet as the communication network for data exchange and technological support. One module in the system is the inverted ROP module which back calculates the compressive strength of the formation while drilling. This in-situ compressive strength will be used dynamically to calibrate the in-situ stresses which in conjunction with knowledge of the overburden and pore pressure, and the wellbore survey dictates the collapse and fracture pressure for the wellbore instantaneously. If the well has already been drilled and the above information is available, the true vertical depth collapse and fracture gradient can be predicted as part of the post analysis or for pre-planning the next well in the area. Another approach to obtain the collapse and fracture gradient is from log data, where the rock compressive strength of the rock is estimated from the sonic travel times. After the compressive strengths as a function of depth are estimated the same approach as for the use of drilling data to obtain in-situ stresses is applied. Using such a decision support software system, data may also be more readily available for display instantaneously or for use in future planning, thus providing a more complete set of data for design prognosis. P. 253

ABSTRACTThe regional stress field in the northern North Sea (offshore western Norway) has been studied through the acquisition and analysis of directions of maximum horizontal compression (s̀H) as extracted from borehole breakouts and from earthquake focal mechanism solutions.The results indicate that the regional stress field is dominated by NW‐SE compression, with good consistency between shallow borehole breakouts (2–5 km depth) and deeper earthquakes (10–25 km depth). The broad spatial consistency in stress direction indicates that the main stress field is related to factors of primarily plate tectonic origin, and the results are in good agreement with the western Europe trend found in earlier investigations.The Tampen Spur region in the northern North Sea has been subjected to particularly complex deformation, with two dominating fault directions trending NW‐SE and NE‐SW. From Tampen Spur in the west to the Sogn graben in the east an anomalous stress field is indicated, with NE‐SW oriented maximum horizontal compressions. This anomaly is clearly seen both in the borehole breakout data and in the earthquake data. Possible sources for this anomaly are discussed, and include postglacial uplift and/or lateral variations in the physical properties of the crust.

ABSTRACT Models for prediction of in-situ rock behaviour often suffer from lack of relevant input data, especially rock strength data. Rock mechanical testing of cores provides only discrete data points. This paper presents a method which produces a continuous plot of uniaxial compressive strength. The method is based on correlating standard log parameters with laboratory measured rock strength data. When a correlation has been established, it can be applied to new wells in the same field, giving a strength plot by computer processing of the necessary logs. The method has been applied to two wells in a North Sea field, resulting in a statistically significant correlation. Proper depth adjustments between core depth and log depth are essential. It is further important that the lithologies of interest are represented among the core material.

This paper studies the stress distribution around an injection well drilled through a layer of poorly consolidated sand. As a plastically strained zone may form around a hole drilled through this type of rock, both elastic and plastic theories must be applied.