Get daily industry updates in your inbox. Free.
Your account already exists. Please login first to continue managing your settings.
One advantage of the integrated geomechanics process that has probably not received as much emphasis as it should have is the internal consistency of the resulting modeling output at all points along the model's evolutionary path. By comparison, conventional modeling methods that traditionally provide the data support for crucial decisions on pore pressure (Pp) and wellbore stability issues involve the combination of various separate models covering geology, geophysics, petrophysics and geomechanics. These can reflect differing starting assumptions, purposes, biases and data-quality levels, and often they are not consistent with each other.
These conventional practices can be problematic not only in the static sense of the quality of data viewed at any given time but in an evolutionary context as errors crop up along the way and revisions are needed. Accountability for revising the working model, in these cases, can be difficult to establish among the multiple sources of modeling data, and the process is quite time-consuming and costly to a major project.
However, the integrated geomechanics workflow generates a living mechanical earth model (MEM) that is internally consistent and eliminates these problems. Indeed, the workflow facilitates the revision process through a closed-loop feedback system that constantly updates the MEM - and incorporates real-time drilling data - as the project progresses.
The geomechanics workflow
The geomechanics workflow is a structured process to reduce operational risks from unplanned rock deformation, including wellbore instability and sand production. The four essential elements of the process are the data audit, the MEM, a closed-loop feedback system and software supporting process implementation over the life of the field or project, depending on the scope of work.
The data audit enables the asset team to gain an understanding of the geomechanics issues affecting operational risk and determine the scope of information available to mitigate that risk. The time an audit may take ranges from a few days for organizations with well-designed knowledge management infrastructure to some months for projects where data retrieval must extend to boxes, filing cabinets and personal computers. While data audits have value even if no subsequent work ensues, the most effective ones result in development of an MEM in support of ongoing field operations.
The MEM is a numerical representation of the state of stress and rock mechanical properties for a specific stratigraphic section in a field or basin. At its most basic level, the MEM would comprise depth profiles of the earth stresses and rock mechanical properties on an arbitrary well path tied to the local stratigraphic section. A simple 1-D MEM would include profiles of Poisson's ratio, Young's modulus, unconfined compressive strength, friction angle, pore pressure, minimum horizontal stress, maximum horizontal stress, vertical stress and the direction of the principle stress axes. A complete MEM would consist of a full 3-D description of these variables. The degree of detail can vary widely based on the availability of data. However, it is important that the initial model contains all the variables and their associated uncertainty.
Beginning even before the first well is drilled and running through field development, the MEM evolves in complexity and predictive power step by step as information is added and refined through the closed-loop feedback process. Once the initial well is spudded, revisions begin and the scope of the MEM can expand to become a fully calibrated model along the path to total depth (TD). After three or four wells, the MEM can evolve into a calibrated 3-D model, and ultimately the data from additional wells can transform it into a 3-D calibration of the entire field.
It is fundamental that an MEM honor all available data and be accessible in real time. For example, Pp predictions can be based on offset well data, mud weight data, sonic or resistivity logs, checkshot data and seismic velocity. An internally consistent Pp model will honor all of these datasets and in the process obtain an appropriate measure of uncertainty. A formal, quantitative specification of uncertainty in the Pp prediction is then incorporated in the well design, and the magnitude and location of the uncertainty is likewise factored into contingency plans.
The same essential conditions, although with different inputs, apply to a wellbore stability model. In either case, as one goes forward, one can reduce uncertainty and mitigate risk through the acquisition of new information while drilling, which becomes new input into the living MEM. The model can then recalculate the relevant predictions in time to support operational decisions. Without this MEM, it is difficult to accomplish this in the short time required - all the more so if data are scattered between several departments, contractors and consultants or if the model one does have is a curve or algorithm in a report.
While the technology for geomechanics modeling has been available for a decade or more, it has taken more recent technology breakthroughs to bring geomechanics software inputs to the level needed and enable the timely delivery of actionable information to the end-user. Powerful collaborative visualization methods have likewise been a major enhancement to the process.
The risks of not having an internally consistent model accessible in real time - such as the MEM generated through the geomechanics workflow - can be quite large.
Consider an operator drilling deepwater development wells who takes the conservative approach of using the highest Pp estimate among several unreconciled predictions, having never established the reasons for the differences in these profiles. After drilling a number of wells, the operator is unable to reach the target formations in part of the field, due to wellbore instability caused by drilling with excessive mud weight. This scenario represents the actual experience of one of our clients. Subsequently this operator was able to reach these targets, ahead of schedule, using the geomechanics approach.
In another instance involving a client drilling a deepwater exploration well, the well design reflected the most optimistic of several unreconciled Pp profiles. When the well took a kick after 27 days of drilling, indicating an error in the Pp forecast, the client initiated a real-time Pp monitoring program to revise the model using sonic logging-while-drilling (LWD) data instead of seismic interval velocities. After tracking the revised Pp model for 4,000 ft (1,220 m), the operator had to make a decision at half TD whether to continue drilling. At this point, there was confidence in the Pp prediction, and drilling proceeded successfully to TD. This case reveals both the value of obtaining a living Pp model and the cost of not having one when drilling began. The initial error in pressure forecasting resulted in the installation of additional casing strings and precious time lost, with daily rig costs running in the hundreds of thousands of dollars.
Following are two recent case studies where the internally consistent living MEM, generated through the geomechanics workflow, enabled clients to achieve results in drilling highly challenging development wells they might not have been able to achieve and certainly could not have achieved as efficiently, rapidly and economically without the MEM.
Directional drilling in the Andes
A MEM helped Pluspetrol drill the first directional development wells in the Camisea field, situated in a tectonically active foothills region of the Peruvian Andes. A MEM was developed from exploration well data acquired from a previous operator and used to forecast wellbore stability along specific well trajectories. In the process, it became evident that the in-situ stress magnitudes were not well constrained in the region and that this uncertainty affected the stable wellbore pressure window. Moreover, there were indications that wellbore instability stemmed from a number of factors beyond the high stress concentrations, including a drillpath that penetrated naturally fractured zones and chemically reactive shales.
Having determined that stress magnitude was uncertain, and noting as well its importance, the team quickly recognized the significance of an unexpectedly high leakoff pressure recorded at the 13-3/8-in. casing shoe in one of the wells. As this leakoff did not reach breakdown pressure, modeling showed that the differential horizontal stress was less than originally estimated. The MEM was quickly recalibrated to account for the new wellbore pressure constraint and other data, and the wellbore stability forecast was revised.
Deeper in the section, an updated geological model required that drillers quickly drop angle from 50? to vertical in order to hit their target. A new stability forecast was generated for the more complex trajectory, and drilling continued while monitoring surface and downhole measurements.
In all, application of an internally consistent MEM enabled a company with no previous experience in a field to drill the first deviated development well to TD 5 days ahead of plan. The achievement reflected the operator's confidence in the model, knowing it to be the most complete available representation of the region and that it was continuously monitored and updated. The rapid reengineering to meet the evolving operational situation would not have been possible without the living MEM. To date, the operator has drilled and completed five deviated development wells - all for less than their amounts authorized for expenditure (AFE).
Drilling in the Gulf of Mexico
In a deepwater field of the Gulf of Mexico, the operator used a geomechanics model in real time to help drill extended- reach development (ERD) wells with horizontal displacements greater than 19,000ft (5,800m). Although sediments there are normally pressured, they are poorly consolidated and create wellbore stability problems in high-angle wells drilled nearly parallel to bedding. Wellbore stability limited the wellbore pressure window to as little as 0.5 pounds per gallon (ppg) over some intervals. A living MEM and wellbore stability forecast defined the wellbore pressure window. Guided by continuous monitoring of drilling, LWD data and the wellbore stability forecast, drillers were able to control equivalent circulating density (ECD) to within 0.1 ppg. This allowed them to drill long intervals successfully, despite the high risks of hole collapse and/or lost circulation.
The predrill modeling identified the risk of lost circulation, due to fracturing, and stuck pipe, due to wellbore instability and/or poor hole cleaning. The risk mitigation strategies included close monitoring of tripping loads, ECD and resistivity for fracture identification. When the first ERD well encountered ballooning and losses, the drilling team was quickly able to determine that the losses were caused by induced tensile fractures, establish their location and effectively treat them with lost-circulation material. Within hours, the models were updated and able to generate a new wellbore-pressure window. The refined model indicated that drilling could continue, using a lower ECD, without risking wellbore instability or further losses.
What was true in Peru was equally true in the very different situation in the Gulf of Mexico. Only an internally consistent living MEM was capable of processing the large volumes of real-time drilling data rapidly enough and with the level of accuracy necessary to provide the continuing high degree of modeling predictability required to guide very difficult, exacting operations in a high-risk environment.
Application of this new technology is helping operators worldwide drill challenging wells below AFE consistently.