A survey of pipeline industry publications generated a database of over 80 collapse tests of cold-expanded line pipe. The data are compared to casing collapse ratings per
Further investigation revealed that, for two operators, the vast majority of offshore surface casing is line pipe that has been cold-expanded without stress relief. Several risk mitigation alternatives were considered. In the short term, the risk can be managed through learning bulletins, design guidelines and operational procedures. The preferred mitigation is to change the collapse rating for cold-expanded line pipe used as casing. This is a long-term solution involving industry standards and the subsequent adoption through commercial design software.
The work described in this paper has led to a ballot change for the next edition of API TR 5C3. This paper is presented to provide drilling industry awareness of the lower collapse performance of cold-expanded line pipe and to add context for selection of an appropriate alternative rating.
A drill string in a wellbore always contacts a casing. However, in a curved section of casing the contact force between the drill string and the casing can be significant. Friction between the casing and a rotating drill string tool joint in contact with the casing generates a heat source at the interface between the two objects. The generated heat energy is a function of rotational speed of the drill string, side force and friction coefficient between the hard-banding layer covering the tool joint and the casing. Heat partition between hard banding layer and casing depends on the thermal properties of both. When there is no mud circulation, e.g. due to a pack-off in the annulus or lost circulation, and the contact region stays in the same section, the resulting temperature increase can lead to degradation of the mechanical strength of both the drill string tool joint and the casing. In addition, the casing strength reduction can facilitate casing wear, which may lead to leak and tool joint heating can lead to heat checking cracks or mechanical strength weakening which may result in a parted drill string due to brittle or ductile fracture.
When there is no mud circulation, rotation of the drill string leads to mud angular rotation inside and outside the drill string. Convection heat transfer occurs due to mud rotation and convection heat transfer coefficient depends on mud flow regime. CFD simulations were performed to compute the convection heat transfer coefficient. Two and three-dimensional steady state and transient finite element simulations were performed to compute the temperature distribution in the casing and the drill string tool joint when there is no mud circulation.
Results show that, when there is no mud circulation, conduction through the drill string and casing has the highest impact on the maximum temperature generated due to frictional heating. Two graphs are plotted, one shows the steady state temperature versus side load at different rotational speeds while the other shows casing yield and ultimate stresses degradation versus increase in temperature. Both graphs can be used by drilling engineers at the well design phase to select the appropriate rotational speed of drill string to avoid failure when there is no mud circulation.
Novelty of this paper is in thermal analysis of a tool joint hard banding layer rubbing against casing. In the analysis the convection heat transfer through mud rotation is involved.
Chen, Chaohui (Shell International E&P Co.) | Gao, Guohua (Shell Global Solutions US Inc.) | Li, Ruijian (Shell Exploration & Production Co) | Cao, Richard (Shell Exploration & Production Co) | Chen, Tianhong (Shell Exploration & Production Co) | Vink, Jeroen C. (Shell Global Solutions International) | Gelderblom, Paul (Shell Global Solutions International)
Although it is possible to apply traditional optimization algorithms together with the Randomized Maximum Likelihood (RML) method to generate multiple conditional realizations, the computation cost is high. This paper presents a novel method that integrates the Distributed Gauss-Newton (DGN) method with the RML method to generate multiple realizations conditioned to production data synchronously.
RML generates samples from an approximate posterior by finding a large ensemble of maximum posteriori points, from a distribution function in which the data and prior mean values have been perturbed with Gaussian noise. Rather than performing these optimizations in isolation, using large sets of simulations to evaluate the finite difference approximations of the gradients used to optimize each perturbed realization, we use a concurrent implementation, in which simulation results are shared among optimizations whenever these results are helping to converge a specific optimization. In order to improve sharing of results, we relax the accuracy of the finite difference approximations for the gradients, by using more widely spaced simulation results. To avoid trapping in local optima, a novel global search algorithm integrated with DGN and RML is applied. In this way we can significantly increase the number of conditional realizations that sample the approximate posterior, while reducing the total number of simulations needed to converge the optimization processes needed to obtain these realizations.
The proposed workflow has been applied to field examples on liquid rich shale or tight oil reservoirs developed with hydraulically fractured horizontal wells. The uncertain parameters include stimulated rock volume (SRV) and matrix properties, such as permeability and porosity, and hydraulic-fracture properties, such as conductivity, height, and half length. The case studies involve a sensitivity analysis to identify key parameters, a history matching study to generate history-matched realizations with the proposed method, and an uncertainty quantification of production forecasting based on those conditioned models. The new approach is able to enhance the confidence level of the Estimated Ultimate Recovery (EUR) assessment by accounting for production forecasting results generated from all history-matched realizations. Numerical results indicate that the new method is very efficient compared with traditional methods. Hundreds of history-matched, or rather data-conditioned, realizations can be generated in parallel within 20-40 iterations. The computational cost (CPU usage) is reduced by a factor of 10 to 25 when compared to the traditional RML approach.
This paper proposes an approach for assessing a reservoir-simulation model for use in estimating reserves. A simulation model can integrate complex static data, the physical description of displacement processes, production constraints, and schedules. Hence, it can provide important information for business decisions and reserves estimation. Confidence in simulation predictions depends on the strength of evidence for the input data, quality control of the model, robustness of the history match, and whether there is independent evidence supporting predictions. We explain the principles for evaluating a simulation model and propose requirements for simulation predictions to be considered as proved reserves. This involves evaluation against different strands of evidence, such as static and dynamic characterization, wells and facilities description, reservoir performance, and analogs. Simulation models are often built to support business decisions by use of the best technical estimates for inputs. There can be instances where a simulation model may be reasonable and reliable but it only represents a “best technical” outcome. There may not be sufficient evidence to count the whole predicted recovery as proved reserves. We propose how such a model may be modified to also provide proved-reserves estimates. The approach can be used with different available data and at different stages of field life. It is illustrated through a case study that shows how the principles may be applied.
Cornelisse, Pieter Marinus Willem (Shell Exploration & Production Co)
On the basis of PVT data from 10 well/reservoirs, fluid models were developed for subsurface and surface requirements. The present paper described a methodology that can be used to maximize consistency between the data used for subsurface as well as surface network modeling. The models were developed to honor all of the experimental data, even though in some cases inconsistencies between the different experimental data sets were found. The workflow is not preliminary aimed at the best match between model and experiment of a single sample but getting the most representative model for the reservoir fluid based on all knowledge. The basis of the model is a PR78 with constant Peneloux volume shift.
The paper describes the workflow for low CGR gas condensate systems from low permeability formations and focuses on three different steps tackling the impact of sampling and experimental issues. It is shown that for this example system many bottom hole samples, both open and cased hole samples, have some external contamination in the C34-C36 range that artificially increases the dew point and liquid drop out at pressures above 2500 psia, while having little to no effect on the overall CGR. Some separator sample sets seem to be depleted in the C5-C10 range compared to all bottom hole samples. It is likely that incomplete equilibration in the separator in combination with some water washing is the reason for this behavior. This observation is supported by the fact that this issue becomes more pronounced for leaner fluids. The issue has a quite strong effect on the CGR (~30%). So the most representative compositions were obtained from cased hole and open hole fluids, corrected for contamination and draw down effects.
In the modeling part of the workflow several steps can be identified: 1) Generation of a 1st model used for QC of data and definition of the parameter ranges (22 comp), 2) Create a model as above but doing a compositional correction for contamination or other compositional irregularities like H2S content, specific contaminants, effects of depletion/enrichment due to 2 phase sampling (22 comp), 3) Create a unified model (32 comp). The model is derived from model 2 but includes BTEX and is mapped onto a single ABC set of pseudo components that are the same for all reservoirs/wells. 4) As 3, but with S-components included. 5) Create the compositional 10 component lumped reservoir model (6 light end + 4 C7+ components).
The models generated show good agreement to the data and are believed to be the most realistic estimate of the fluid behavior of the reservoir fluids. On top of that we can swap from one model to another using a simple lumping/delumping scheme.
As part of the work processes of dealing with fluids samples, generating PVT data and performing QC, several PVT models were generated for fluid samples from the exploration and first appraisal well. Later other parts of the field were appraised, with strongly different compositions. As a result the question arose whether it is possible to describe all those fluids with the same models, and if so could we use that characterization to model the surface processing steps at least to some degree without having to go to a completely new fluid model with pseudo components and interaction parameters. Several approaches in the open literature are available but none did exactly meet the needs we had in mind for this project. As far as lumping/delumping procedures are concerned, the method as presented by Leibovici [2, 4, and 5] seems most suitable.
This paper proposes an approach for assessing a reservoir simulation model for use in estimating reserves. A simulation model can integrate complex static data, the physical description of displacement processes, production constraints and schedules. Hence it can provide important input to business decisions and to reserves estimation. Confidence in simulation predictions depends on the strength of evidence for the input data, quality control of the model, robustness of the history match, and whether there is independent evidence supporting predictions. We explain the principles for evaluating a simulation model and propose requirements for simulation predictions to be considered as proved reserves. This involves evaluation against different strands of evidence e.g. static and dynamic characterisation, wells and facilities description, reservoir performance and analogues. Simulation models are often built to support business decisions using best technical estimates for inputs. There can be instances where a simulation model may be reasonable and reliable but it only represents a ‘best technical’ outcome. There may not be sufficient evidence to count the whole predicted recovery as proved reserves. We propose how such a model may be modified to also provide proved reserves estimates. The approach is illustrated through a case study which shows how the principles may be applied with different available data and at different stages of field life.
A simulation model can integrate complex static data, the physical description of displacement processes, production constraints and schedules. Hence it can provide important input to business decisions and to reserves estimation. For estimates to be reported as proved reserves sufficient confidence must be established in the predicted production. Companies and institutions commonly assess reserves using definitions either from Petroleum Resource Management System (PRMS) (2007) or from United States Securities Exchange Commission (SEC) (Modernisation 2008). Under these definitions there must be Reasonable Certainty in estimated production from a commercial project for this to be classified as proved reserves. If deterministic methods are used then Reasonable Certainty means for PRMS ‘a high degree of confidence that estimated quantities will be recovered’ and for SEC ‘much more likely to increase or remain constant than to decrease’. The challenge is to assess whether model predictions from a particular simulation meet the standard of Reasonable Certainty and how to adapt less certain predictions to meet the standard.
Several authors have proposed approaches to assessing or modifying simulation models for use in reserves estimation. Rietz and Usmani (2005) focus on the situation in which there is sufficient performance data to test a model against history. They highlight assessment of:
The Perdido development is one of the most complex deepwater projects in the world. It is operated by Shell in partnership with Chevron and BP. It currently produces hydrocarbons from twelve subsea wells penetrating four separate reservoirs. The properties of produced fluid vary per reservoir as well as spatially. The producing wells display a relatively wide range of fluid gravities, between 17 and 41 °API, and producing Gas Oil Ratios (GOR), between 480 and 3000 scf/bbl. The fluids produced from the subsea wells are blended in the subsea system and lifted to the topsides facilities via five seabed Caisson Electrical Submersible Pumps (ESPs). In the topsides facility, gas and oil are separated, treated and exported via dedicated subsea pipelines. The fluid compositions and properties across the various elements of the production system are used as input to the respective simulation models, and the corresponding outcomes (i.e. fluid properties, compositions, etc.) vary upon the well/caisson lineup and daily operating conditions.
Given the wide spectrum of fluids produced through the Perdido spar, a special Equation of State (EOS) characterization of the fluids had to be developed. As a common EOS model was utilized to characterize the fluids, we will call this the Unified Fluid Model (UFM) throughout this manusctript. This approach enables accurate and efficient prediction of the properties of blended fluids and is suitable for use in an Integrated Production System Model (IPSM) that connects reservoirs, wells, subsea flowlines network and topside facilities models. Such modeling scheme enables effective integration among relevant engineering disciplines and can represent production and fluid data from field history with high confidence.
The IPSM uses a black oil fluid description for the well and subsea flowlines network models. Based on the initial composition and producing GOR of each well, the fluid composition is estimated via a simple delumping scheme. The resulting compositon is tracked through the subsea network to the topsides facilities model, where compositional flash calculations are performed. The IPSM model can forecast production rates together with fluid properties and actual oil and gas volumetric rates across the whole production system. The model can be used to optimize production under constrained conditions, such as limited gas compression capacity or plateau oil production.
Projects in the global deepwater environment are increasingly complex, costly and often face significant schedule pressures to reach both a “commitment to develop” decision and “first oil/gas milestone”. Typical industry deepwater project costs ranges from USD$1 billion for a small scale subsea tie-back to upward of USD$10 billion for a custom design standalone hub class development. Therefore ensuring that the appraisal campaign addresses the data acquisition required to reduce both key subsurface, and key development uncertainties as well as enable a robust development decision is critical. The appraisal phase of the project (post discovery) is the most important early phase in the development cycle which sets the stage for the remainder of the concept selection, detailed design, execution and production phases of the development.
This paper explores and discusses, with the use of relevant examples, the key aspects of appraisal planning and execution specifically related to reducing uncertainties and determining the ultimate development scope for deepwater oil and gas development projects. The paper does not attempt to expand on building understanding or managing the non-technical aspects of early phase deepwater project management which can have an equally significant role in shaping the ultimate deepwater development project.
Van Den Haak, Arno L.M. (Shell International Exploration & Production Inc) | Cameron, Wylie J (Shell Exploration & Production Co.) | Grant, Lisa S (Shell Exploration & Production Co.) | Japar, Nor Janiah H (Shell Exploration & Production Co) | Reagins, Deandre R (Shell Exploration & Production Co.)
The Mars B Development announced Final Investment Decision (FID) in 2010 and represents one the first major new development to be executed during a period of evolving technical and regulatory requirement. Shell’s move towards FID demonstrated our commitment to the Gulf of Mexico. The project aims to unlock these resources over the next 50 years thru the deployment of a new floating TLP Deepwater structure?the Olympus TLP a 24 well platform and six subsea wells.
Unique challenges pertaining to the well design extend beyond the regulatory environment and include numerous technical challenges from the development of the 24 slot TLP wells targeting over 50 horizons in a 10.000’ vertical sequence to approx 22,000 ft TVD, to designing and delivering six West Boreas/South Deimos 15K subsea wells..
The well trajectories cover a range from low angle, high angle, extended reach and complex through salt penetrations. Multiple design challenges are present and must be managed to ensure well integrity, reservoir isolation, and desired well construction objectives; all requiring successful development and deployment of novel technologies and world class systems to achieve top quartile. Design requirements have resulted in an evolution of traditional well designs. This has resulted in the use of unique casing sizes, larger casing wall thicknesses, tighter clearances and the reduction of traditional contingency options. The new requirements increase the already complex well construction.
The presentation will provide a high level overview of the well and project challenges as well as details of a few of the specific challenges. Some of the challenges to be discussed are the development of novel tieback systems and solutions, zonal isolation strategy, integration and automation of new surface rig equipment, development of an innovative drilling riser concept, advanced casing and liner construction techniques, and equipment to accommodate the 50 year design life. The design life is the longest design life of a tension leg platform in Shell’s history.
The development and deployment of these new technologies will not only have benefits for the Mars B Development, but has already identified potential benefits for brown-field environments and the next generation of Gulf of Mexico projects.
Gomez Alonso, Hector (Shell Exploration & Production Brazil) | Avila, Runer (Shell Exploration & Production Co) | Peppard, Bret (Shell Exploration & Production Co) | Gee, Vicki (Shell Exploration & Production Co) | Baima, Joao (Shell Exploration & Production Co) | Simone, Alessandra (Shell Exploration & Production Co) | Tejera-Cuesta, Pablo (Shell Exploration & Production Co)
This paper describes how the integration of basic, analytical and numerical reservoir engineering techniques in the Salema Field (brownfield offshore Brazil under waterflooding), has proven to find/support additional opportunities, and to really “understand reservoir dynamics”.
Salema is a mature field currently with only one oil producer and one water injector. Current recovery factor (RF) is 29%, and under NFA conditions the field was supposed to “economically” die in 2016 with a final RF of 32%. Under those circumstances, it was key to determine the presence of additional opportunities or let the field face its destiny.
The application of traditional techniques such as conventional production analysis, diagnostic and Chan plots, material balance, Build-ups/Fall-off tests allowed the confirmation of different key events and water production mechanisms, to characterize the basic description of reservoir dynamics. All this understanding was incorporated into the simulation model.
The methodology helped to minimize the number of uncertainties, narrowing down the variables that had a considerable impact on the history match, diminishing the number of potential realizations, and more importantly ended up in a more consistent and better understanding of the production behavior of the field.
Previous models were not able to reproduce the water production/reservoir pressure observed in some of the existing wells, clearly suggesting that the injected water volumes were not in the right place in the reservoir.
The methodology can be considered as an integrated workflow towards the construction of any 3D reservoir simulation model, being simple and easy to apply. The 4D results obtained six months later confirmed the results obtained from what the model predicted.
This case study is another example that success in subsurface disciplines depends more on the correct understanding of the reservoir than the degree of complexity with which it is modeled, using reservoir simulation to confirm the achieved understanding rather than just getting a history match.
All the integrated work served to identify two additional opportunities that will increase the final RF to close to 40% and contributed to extend the field life until 2023. Such opportunities have later been confirmed by the 4D.