Wells that are drilled today are becoming deeper and more complex, and actions to make these wells economically justifiable are desired by oil companies. One method is to reduce the non-productive time during drilling by analysing the hook load response to restrictions in the well. The goal is to recognize abnormal behavior as early as possible to prevent stuck pipe or other incidents that may lead to non-productive time. In present thesis a mathematical model for simulating hook load during normal tripping conditions is made. The model is based on theory about forces acting in the borehole, and properties of the drill string and the wellbore are inserted into it. The goal is to make good hook load simulations for simple tripping conditions, as restrictions most often cause problems during tripping operations. If restrictions were to occur, the changes in hook load during tripping would be identified quickly, and remedies can be started early. A laboratory that simulates tripping out of a nearly horizontal well is buildt and experiments are done. The laboratory experiments are carried out with and without mud flow, to identify also the forces that depends on mud circulation. Washout has been imitated in order to see how restrictions affect the hook load. The model consists of mathematical equations that describe all forces acting on the drill string. The laboratory results were used to adjust the equations, so they match hook load behavior even better. From comparison with the laboratory results it was found that the model creates accurate hook load simulations under simple conditions. Simulations with mud flow during tripping were inaccurate because the forces that occur during circulation were poorly mathematically described in the model. It is not possible to run real time drilling data in the model to see if the model is realistic because block position measurements are missing and some measurement rates are too low, so the equations in the model is not general enough. With improved position measurements and following adjustments in the model, it is likely that the hook load behavior model could be used to reduce the non-productive time during drilling.

To cover the world’s future consume of hydrocarbons, technological improvements are needed, turning currently unreachable and unprofitable reservoirs into the opposite. The main focus of this thesis, Dual Gradient Drilling (DGD), is a drilling technology that is ideal for drilling in ultra-deep waters, and could prove vital in drilling the reservoirs that are currently undrillable. This will increase the possible supply of hydrocarbons available for consumption. DGD is not fully accepted in the drilling industry, and DGD is still considered unconventional. In this thesis, simulations investigating pressure control in a DGD system are done, showing possibilities and limitations when using DGD. Also, it is done a study of how the challenges faced on the Macondo prospect in the Gulf of Mexico could have been solved better by using DGD instead of the conventional drilling methods. MATLAB was used for mathematical simulation of the control of the hydraulic pumping system. The program reads field- and equipment specific input data from an excel sheet. The different input parameters are changed separately, and simulations are run for each parameter change, showing each parameter’s effect caused to the system abilities. Criteria for approved well control are set, and by trial and error with the program, requirements for pump rates, well bore design and pressure safety margins needed are found, presented and discussed. It is here shown that the currently available pumping rates are not able to control the pressure as agile as other available drilling technology can, using the marine riser size currently developed. A subsea pump producing rates of more than 13000LPM, the double of what can be produced today, is needed. Utilizing a narrower marine drilling riser, with an inner diameter (ID) of 12¼” instead of the conventional riser with 19½” ID, can increase the speed of a pressure change with 171%; from 0.045bar/s to 0.121bar/s, when other parameters are kept constant. Because of linearity, a doubling of pumprates will result in a doubling of pressure change speed. It is here concluded that sufficiently agile wellbore control, depends on the development of a narrower riser and higher subsea pump rates. Other unconventional drilling currently available, like Managed Pressure Drilling (MPD), is far superior to DGD when it comes to quickness and accuracy in keeping the BHP at a desired level.

This thesis summarizes and concludes my master thesis research work. The main objective of this research was to develop a rigorous and generic forecast model for all the fields on the Norwegian Continental Shelf based on publically available data and free software. Pipe-It Norwegian Continental Shelf Integrator and Forecaster solution provides the opportunity to forecast oil and gas production rate and economics for all assets on the Norwegian Continental Shelf. The solution is automatically synchronized with an on line database, that is maintained by the Norwegian Petroleum Department. The solution currently contains 87 fields and handles thousands of application launches in parallel. The results can be filtered and aggregated for multiple engineering purposes, like the impact of new discoveries on future production rates and economics.

Many accident investigation techniques and other methods used by the petroleum industry today list a set of underlying human related causes and subsequent improvement suggestions. Do these techniques address the root cause behind the problem so that the appropriate initiatives can be implemented? The focus of the present thesis was to determine the human related root cause of two major accidents in the North Sea. This in order to give recommendations to improve the safety levels in the organisation. In order to achieve the above-mentioned goals, the IPT Knowledge Model was adapted to the given accidents. The data input into the model was based on interpreted observations from former investigation reports. The analysis of the blowout on Snorre A and the well control incident on Gullfaks C resulted in 49 and 63 observations respectively. For both accidents, the Human Factor that was indicated to have the largest affect on the accidents was Training and Competency (29% for Snorre A and 19% for Gullfaks C). Lack of competence was indicated as the majority subclass. Collectively, management and supervision, or lack thereof, was also indicated as being a contributing factor to the accidents. These final results coincide with the findings in other investigation reports. However, these are more acute, indicating a specific area of improvement within the company. By increasing the competency levels within the company and ensuring that the leaders and management have the proper tools to follow-up their employees and their operations, the safety levels and culture will improve.

Denne oppgaven omhandler teoretisk modellering av trykktap grunnet rotasjon i eksentriske borehull med Herschel-Bulkley strømning. En eksisterende modell for trykktap for konsentriske borehull ble videreført til å inkorporere borerrørets eksentrisitet ved å fremstille ett konsentrisk borehull med flere konsentriske borehull med varierende radier. Gjennom å studere numeriske resultater for trykktap for både Newtonske og ikke-Newtonske strømninger i ringrommet, ble en korrelasjon for trykktap grunnet akselerasjon av strømningen laget som en funksjon av viktige parametere. Disse inkluderer Taylor nummer, eksentrisitet og radiusforhold (borerør- mot ringrom-radius). Den nye modellen kan altså beregne trykktap i ringrommet i eksentriske borehull med effekten av rotasjon (skjær-fortynnende og akselerasjon av strømmen inkludert); eller enklere scenarioer hvis det er ønskelig. Modellen ble skrevet til et enkelt program som kan brukes i felten. Programmet ble videre validert mot eksisterende numeriske data og feltdata. Resultatene var svært gode for konsentriske borehull og trendene ble riktig regnet. Trendene for Herschel-Bulkley strømninger var ikke like gode som for de andre fluidene i eksentriske borehull, men anses fortsatt som tilfredsstillende.

En av de viktigste operasjonene i en brønnkonstruksjon er primærsementering. En dårlig utført primærsementering kan være en fare for både mennesker og miljøet rundt. En reduksjon i produktivitet, dyre sementering løsninger for å reparere skader og tap av brønnen kan komme av en dårlig primærsementering. Migrering av olje og gas grunnet mangel på isolasjon kan føre til forurensninger, i verste fall kan mangel på isolasjon føre til en farlig utblåsning. I dag er det slik etterspørselen etter olje og gass økes mens oljen og gassen som er ”lett tilgjengelig” minkes fort. Dette fører til at operatører blir tvunget ut mot dypere vann, arktiske områder, reservoarer med høytrykk og høytemperaturer og andre ukonvensjonelle resurser blir oppsøkt. Konsekvensen av dette er at de ulike operasjonene som er involvert blir mer komplisert, spesielt primærsementeringen. Dagens sementeringsteknologi gir ikke tilstrekkelig isolasjon over lenger tid. For å forbedre på primærsementering har alternative materialer med egenskaper som gir bedre langtidsisolering blitt evaluert. En vurdering av ThermaSet og Sandaband som primærsementerings materiale ble gjennomført, forslag til plassering av materialet og evaluering metoder ble også diskutert. Materialene ble sammenlignet opp mot konvensjonell sement. Vurderingen av materialene viste at det kan være mulig å bruke både Sandaband og ThermaSet som erstatning for sement i en primærsementerings operasjon. Begge materialene hadde egenskaper som gjorde dem passende for en langtidsvarende primærsementering, men sement vil fortsatt være det eneste materialet for primærsementering i nær fremtid. Dette er på grunn av den overlegne evne til å støtte foringsrøret, mangfoldet av sement å velge mellom, og de lave kostnader av sement. Bruken av Sandaband vil være begrenset til enklere brønnløsninger, et eksempel på bruk av Sandaband kan være vertikale brønner i nærliggende gassreservoarer. Når det gjelder ThermaSet så er det fortsatt en del tester som per dags ikke er blitt utført. Et forslag til å redusere svakheten av sement og best mulig forbedre på primærsementeringen er å benytte seg av forebyggende tiltak, et eksempel på et forebyggende tiltak er ”swellable elastomer packers”. SEP i bruk sammen med sement kan styrke sementen. En vurdering av muligheter til å kombinere ulike sement systemer sammen for å oppnå en bedre og langvarig sement design kan også være en løsning til å styrke svakheten med sement.

The world’s total oil reserves are to some extent dominated by heavy oil. The heavy oil reserves are doubled in volume compared to conventional oil reserves. As conventional oil reservoirs are depleting, heavy oil and bitumen possesses a great potential in covering parts of the future energy demand. The possibility of horizontal drilling has created a pathway for SAGD (Steam Assisted Gravity Drainage), which is the most preferred heavy oil and bitumen recovery method. The mechanism of SAGD involves two parallel horizontal wells, one for production and one for injection. The production well is situated at the bottom of the reservoir and the injection well is placed above. Steam is injected and heats up the oil which is then able to flow to the production well by gravity drainage. In the present thesis, a numerical study of parameters has been performed in relation to SAGD implementation in the Athabasca field. The thermal simulator utilized is CMG STARS. The Athabasca field is located in Northern Alberta in the Western Canada Sedimentary basin. Due to the complexity of core extraction in bitumen reservoirs, a comprehensive sensitivity analysis is significant in order to determine the appropriate production approach. The present study confirmed that a decrease in viscosity and increase in porosity yielded higher oil recoveries. All oil recoveries found in 3D simulations were within model uncertainties compared to the 2D result. Increase in horizontal and vertical permeabilities resulted in higher oil recovery up to a certain limit, where exceeding permeabilities provided limited increase in oil recovery. The effect of different vertical well spacing proved to have minor effect on amount of oil produced. Yet, based on cumulative steam oil ratio (CSOR) it was proposed to maintain a vertical well spacing in the range of 3.5 to 7 meters.

Abstract Pre stack inversion of seismic data consists of numerous difficulties. Two of the problems of greatest concern are the problems of non-uniqueness and non-linearity of the inversion. There may exist several solutions to any given inversion problem, and to be able to choose the correct solution we are dependent on a priori information. This thesis will explain how a priori information can be implemented with the seismic data using Bayesian modeling and fractal based initial methods in order to obtain the most likely solution for the inversion. This thesis will also explain the theory behind global optimization routines, such as the random walk Monte Carlo, the Metropolis algorithm and Simulated Annealing. A Simulated Annealing routine has been made, and this is used to solve optimization problems. The routine is analyzed for its capability of finding global optimums and the requirements for its success. It is then implemented to simulate the inversion of a seismic dataset. The solutions of the inverted data is then analyzed and compared to the actual solution. This is done for an uncontaminated dataset, and for a dataset containing noise. The work has shown that Simulated Annealing can be a good method for finding a global optimum, but that the global optimization routine is unable to produce good results without good constraints and a good initial model, due to the problem of non-uniqueness.

The main objective of drilling a well is the potential economic profit. The total cost of a drilling operation result from a complex interplay of many factors (Graham and Muench, 1959). Work that can optimize some of these factors is desirable, and the need for information that can eliminate unfortunate incidents resulting in increased operational time and make the operation more efficient is constant. Well cemented layers and the composition of the cemented materials are the main causes of hard stringers. Hard stringers may lead to severe downhole problems like washouts, severe local doglegs and keyseats. Such downhole problems are often related to the unexpected appearance of hard stringers and may cause the drill string to stick and result in lost time. Hard and soft stringers may be detected through real-time drilling data. This is achieved by comparing different drilling parameters, like rotary speed (RPM), weight on bit (WOB) and block position (BPOS). A decrease in RPM followed by an increase in WOB indicates hard formations. However, manual analyses are time consuming and inefficient. A modified and improved version of the hardness detection program previously developed by Solberg (2011) is presented in present thesis. The model is based on a simplified version of the rate of penetration (ROP) equation proposed by Bourgoyne and Young (1986). Drillability is the desirable result from the equation, as drillability is the inverse of hardness. Not all the functional relations that initially are part of the ROP equation have been assumed relevant for the detection of soft and hard stringers. Some of them alter only gradually, either with depth or as the drilling operation progresses, and do not affect the sudden change in drillability as the well is being drilled. WOB, RPM and ROP (ΔBPOS/Δt) are drilling parameters chosen as relevant for present work. They are obtained through real-time drilling data and utilized in the calculations of drillability. Exponents related to the WOB and the RPM functions will frequently vary with the hardness in the formation. However, frequent manipulation of the exponents is difficult to achieve and only one soft and one hard formation exponent related to each exponent type have been applied. The final result shows a plot of hardness variation with depth. The plot has been proved to correlate well with experienced hard stringers stated in the Final Well Report. The program has been proved to be able to detect lithology transitions. The boundaries of particularly the Utsira Fm are evident on the hardness curve. The detection of Utsira Fm is enhanced by both gamma ray and sonic log from the same formation in wells nearby. Correlations between the two logs and hardness have been established. However, due to absence of sonic and gamma ray data from Well 0, the trustworthiness of the establishment is hard to evaluate.

Petrographic composition, diagenesis and reservoir quality of the Upper Triassic fluvial to marginal marine Fruholmen Formation, the Middle and Upper Triassic estuarine and coastal plain Snadd Formation, and the Middle Triassic estuarine Kobbe Formation have been examined in wells 7131/4-1 and 7222/11-1 located on the western Bjarmeland Platform and on the eastern Finnmark Platform in the southern Barents Sea. Twenty-nine thin-sections from the Fruholmen Formation, sixty-eight from the Snadd Formation and eleven from the Kobbe Formation were studied with a petrographic microscope, forty were point-counted, and six were studied with a cathodoluminescence microscope. The petrographic observations were compared with plug measurements of porosity and permeability.The Fruholmen Formation sandstones are moderately well and well sorted fine-, medium- and coarse-grained quartz arenites. The Snadd Formation sandstones are moderately well and well sorted and fine- and medium-grained, and the Kobbe Formation sandstones are mostly well sorted and fine-grained. Both the Snadd and Kobbe Formations comprise sublitharenites and lithic arenites with common to abundant metasedimentary rock fragments, but the Snadd Formation contains both K-feldspar and plagioclase, whilst the Kobbe Formation contains plagioclase only. These compositional differences between the three formations may largely be a result of the Snadd and Kobbe formations having had an eastern source area (Uralides), and the Fruholmen Formation a more southerly source area (Scandinavia), although a shift to a more humid climate in the Late Triassic may also have influenced sand composition.The Fruholmen Formation sandstones contain very little diagenetic cement, typically 1-2% quartz overgrowths, traces of pyrite cement, and occasionally 1-2% authigenic kaolin. The main diagenetic cement in the Snadd Formation is early diagenetic chlorite that occurs as grain coatings and more rarely as pervasive microporous pore-filling cement. Siderite is present in most Snadd Formation samples in amounts of 1-6%, and a few thin zones are strongly calcite-cemented. Up to 3% authigenic kaolin is commonly present, and traces of pyrite cement and quartz overgrowths occur. The Kobbe Formation sandstones contain the same diagenetic minerals as the Snadd Formation, plus a few albitic overgrowths on plagioclase. However, the total volumes of diagenetic minerals in the Kobbe Formation are typically very low, 1-5% in the point-counted samples.Diagenetic chlorite or poorly crystalline chlorite precursors may have formed from iron-rich colloidal material brought in by rivers and flocculated where fluvial waters mixed with marine waters. These colloids may also have been the main source for iron in early diagenetic siderite. Partly dissolved biogenic carbonate is still present in the Snadd and Kobbe formations, and carbonate fossils are probably the source of the calcite cement. The calcite cement engulfs and therefore postdates chlorite grain coats, siderite cement and authigenic kaolin. Quartz overgrowths were sourced from dissolution of quartz grains at stylolites evolved from clay laminae when temperatures reached 70-80°C.Reservoir quality is largely excellent in the Fruholmen Formation sandstones (25-32% helium porosity, 1 000-38 000mD permeability) because contents of detrital clay and diagenetic minerals are almost zero. Snadd Formation porosities are also in most cases high to very high, 26-36%, partly due to the chlorite coatings inhibiting quartz cementation. Permeabilities are mostly 100 to 5 000mD, but where microporous diagenetic chlorite fills the pore system permeabilities are very low to low, 0.1-15mD. The Kobbe Formation sandstones have been more deeply buried than the overlying formations, approximately 3.5km, content of soft components (detrital and authigenic clay, mica-rich rock fragments) is high, and compaction has therefore been severe. Porosities are consequently quite low, 15-21%, even in the best of the cored Kobbe Formation sandstones, and together with the fine grain size this results in low permeabillities, 1-20mD.Burial depth for the shallowest examined cores is only 0.4 and 0.56km, and present temperatures are around 30°C. The consolidated nature of these cores and the presence of quartz overgrowths that normally start forming at 70-80°C therefore suggest that the sandstones have been more deeply buried than at present. Lack of illitization of kaolin in the deepest samples indicates that they have not been subjected to temperatures above 130°C. Together with the degree of quartz cementation in the various examined samples this suggests uplift of around 1.5km in well 7131/4-1 and 1.7km in well 7222/11-1. Comparison of the present porosities in the quartz arenites of the Fruholmen Formation with the porosity depth trend for the Garn Formation also suggests 1.5km of uplift in well 7131/4-1.