Integration and collaboration were keys to identifying an optimal field development plan.

Schlumberger Data and Consulting Services conducted a multidisciplinary integrated study on a highly overpressured gas reservoir, Kela 2 Field, Tarim Basin, China. The study, in collaboration with and on behalf of PetroChina, was to identify an optimum field development plan (FDP) for this sizeable resource. With reserves exceeding 10 Tcf, the field is being developed to supply gas to the Shanghai region.

The technical aspects of the study consisted of constructing a project database; building comprehensive geologic, reservoir and single-well models; estimating reserves; performing dynamic production and performance analyses on the reservoir; conducting economic and risk analyses; and recommending an optimal FDP. The final FDP included drilling and completion designs, production engineering, reservoir monitoring and management, optimal implementation methods and best practices.

Achieving high-quality output hinged on designing a workflow process that supported seamless integration of the various disciplines required, as well as close collaboration between the consultant (Schlumberger) and client (PetroChina).

Background

Kela 2 gas field, situated on the northern edge of the Tarim basin, is a thick, overpressured dry gas reservoir with normal temperature, located in the tectonically active Tian Shan orogenic belt. Discovered in 1997, the productive structure was originally defined by 2-D seismic surveys. Additional seismic guided subsequent drilling of five delineation wells.

The productive formation is of fluvial origin, consisting of sandstone (fine- to medium-grained sublitharenite and feldspathic litharenite) with minor amounts of silty mudstone. Matrix materials consist largely of pseudomatrix formed by deformation of mudstone intraclasts.

Pressures in the Kela 2 reservoir measure twice hydrostatic pressure, although the origin of the overpressure remains unclear. Ninety-seven percent of reservoir fluid is methane; and other reservoir properties, including porosity, permeability, temperature and saturations, are typical.

Technical challenges

When beginning the reservoir analysis and field development project, it was clear that balancing safety (managing high pressures) and quality (formation damage control) would be the major technical challenges. As for the analysis project itself, while PetroChina provided all available field data, some very important information was missing, including the impact of pressure-dependent reservoir properties (permeability and porosity) and unknown aquifer and boundary water, gas/water relative permeabilities and capillary pressure data, and a pressure survey.

Methodology

Based on the overall project scope and the technical challenges identified, an integrated workflow process was chosen, as seen in Figure 1. The project team consisted of a geologist, geophysicist, petrophysicist, modeling expert, reservoir engineer, production engineer, drilling and completion engineer, and economist. A project coordinator and a team leader helped ensure close collaboration among the disciplines. For the results of one discipline to be used fully and integrated with results from the other disciplines, the work in different disciplines was planned to occur simultaneously whenever possible.

The study was divided into six phases, as listed in Table 1. Each phase had clear milestone deliverables and go/no-go decision points defined to ensure that the intermediate and final work product would be timely, relevant, economically justified and on objective. After the phases were defined, following the Figure 1 process, the group developed the workflow plan shown in Figure 2.

The first phase focused on data collection and review and development of a comprehensive project database, including standardization of data formats and interpretation/integration processes. A Schlumberger team traveled to Tarim to collect and review all the data, determine data deficiencies (including a plan to collect additional data or evaluate uncertainties), clarify project objectives and agree on the final project workflow.

After completing each phase, the team held a formal milestone go/no-go meeting. The project also underwent several peer reviews; reviewers were selected from senior technical staff to serve a quality assurance/quality control function. The reviewers often brought different viewpoints, as they were not involved in day-to-day project work. Together, go/no-go meetings and peer reviews aimed to: evaluate project status and planned activities, ensure that the conclusions/recommendations were supportable based on study activities; and ensure that any uncertainties arising from data deficiencies or project constraints were fairly represented and adequately described. At a couple of critical milestones, a team of Schlumberger and PetroChina peer experts reviewed the status of the project and made recommendations to improve the expected outcome.

Since the field had no commercial gas production yet, no reservoir-based calibration of the static model with dynamic data could be performed. The calibration was limited to well productivities only, which had two impacts on the project:

1. efforts were focused on getting a very good static reservoir model, and
2. uncertainty analyses would be required in the economic phase of the FDP.
The following sections detail the methodology used for each major discipline involved.

Geophysical interpretation

The geophysical interpretation focused on defining the shape of the Kela 2 structure. The original 2-D seismic data were reinterpreted, and results were validated and improved with the 3-D seismic data acquired in the late stage of the project.

The team studied the seismic data to determine if reservoir property attributes - particularly porosity - could be extracted and mapped, or if the available seismic could be used for direct fluids detection. However, neither approach was conclusive. The seismic ultimately was used only for structural mapping.

Geological interpretation

Geological evaluation provided a basis for distributing the reservoir properties in the 3-D reservoir model as well as a means to evaluate the influence of faults and fractures on reservoir development. In addition, geological modeling of the reservoir and areas surrounding it provided input for creating a geo-mechanical model for use by the drilling engineers.

Petrophysical analysis

The crew performed the petrophysical analysis to identify gas and water zones and study reservoir properties. Special logging data such as microscanner images were available for only two wells, but dipmeters were available for all wells.

All five wells contained sufficient data to allow detailed analysis, including porosity, saturation, permeability, and net-to-gross and net-pay thickness in the main-pay sand reservoir. The group confirmed the consistency of the results using all available core, geological and engineering information.

Geological modeling

A simple "layer-cake" model of the Kela 2 reservoir was found to adequately describe the hydrocarbons in place but was not sufficiently detailed to predict early production rates of wells or provide a means to simulate proper fluid flow paths given the limited long-term test data. Therefore, a geo-cellular model was built to provide estimates of early well production rates and gas-water flow behaviors.

The fluvial facies architecture for the distribution of reservoir properties was determined by using both object modeling and Sequential Indicator Simulations. The geologic model was then conditioned to the data from the four wells that penetrated the productive part of the Kela 2 structure. Because the number of sand bodies included in the reservoir was extremely large, the resulting model was believed to be an excellent representation of the reservoir itself. Porosity, permeability and fluid saturations were distributed in a geologically realistic manner. Effective porosity values were assigned within each of the stratigraphic intervals found in the geological model using the distribution of log-derived porosities that occur within the unit.

The permeability model was created using Sequential Gaussian Co-Simulation. As in the process of populating porosity, permeability was populated by stratigraphic intervals based on the distribution of log-derived permeability values compiled for each unit.

Neither water saturation values nor the value of total gas in place were calculated at this stage. In the workflow process used, these reservoir properties were treated in the dynamic reservoir simulation model, where the effects of capillarity are better handled.

Dynamic reservoir modeling

The team used a dynamic, numeric reservoir simulation model that fit the following criteria:

• a grid system for discretizing the 3-D space occupied by the reservoir;
• placement of petrophysical properties in each grid cell;
• use of relative permeability data to characterize multiphase flow;
• gas-water J-Function to correlate capillary force to fluid (gas-water) saturation;
• rock compressibility for pressure or stress dependence of rock properties;
• fluid properties for information on fluid behavior;
• datum pressure for distributing pressure throughout the model;
• fluid contact location(s);
• external aquifer properties for defining additional aquifer support;
• development well completion depths and locations;
• well control criteria (target production rate, bottomhole pressure threshold, etc.); and
• other control elements (pressure maintenance schemes, abandonment criteria, etc.).
Each of these components was prepared systematically while making the best use of existing data. Extensive analyses were performed on the wide range of available reservoir engineering data. These analyses helped establish a better understanding of reservoir rock and fluid properties as well as the uncertainties associated with them.

Following creation of the geologic model, a simulation grid was designed to:

• maximize horizontal resolution for characterizing inter-well fluid flow;
• maximize vertical resolution for predicting water coning;
• limit number of active grid cells to provide reasonable simulation run time;
• minimize grid distortions;
• accurately reproduce major horizon structures; and
• represent fault juxtapositions accurately.

Simulation

Typically, following the construction of a simulation model calibration work is needed to ensure that the model produces historical production data. However, in this case, production from the field was limited to short-term well tests. Furthermore, overbalanced drilling damaged the existing wells.

Among the production tests carried out, only one was subsequently suitable for calibrating a simulation model. The calibration work resulted in an upward adjustment in the permeability values.

Optimization

Optimization work was performed to maximize production efficiency; that is, the length of the production plateau and ultimate recovery. Very few wells were initially input in the model to determine locations of major areas of remaining gas. Additional wells were inserted later to increase production efficiency. Scenario runs were then performed to finalize the choice of other operational parameters (e.g., tubing size and perforation length). Each scenario was evaluated from an economic perspective to determine its relative merit.

Sensitivity simulation studies were also performed varying the number of wells, tubing size, perforation length, plateau production rate, compressor installation and sand control measures. These studies were helpful in identifying a "base case" recommendation, which was included as a guideline for future development plans.

Earth mechanical model

Geomechanics covers a large scope of work during the life of a field. The basic approach is to use available data to interpret earth strength, stress and pressures. Examining the available data and understanding key issues previously encountered in drilling, testing or production can focus efforts on specific issues that will have the greatest impact on field development. The key is ensuring an internally consistent approach to data integration and interpretation. A series of steps must be followed to grasp fully the quality of the available data and determine uncertainty in conclusions drawn from data used. Skipping or ignoring the importance of any one of these steps can lead to inconsistencies or poor assumptions in the results.

Developing a mechanical earth model

(MEM) is essential to make the best use of a field's geo-mechanical information. The MEM
is a description of strengths, stresses and pressure, as a function of depth, referenced to a stratigraphic column. Once constructed, a MEM is used to estimate and predict the best possible methods for safely drilling and completing both a single borehole and a full field.

The Kela 2 MEM was developed primarily to determine the safe mud-weight window for drilling, understand the geo-stress distributions in the field and factors controlling abnormal pressures, evaluate wellbore stability and analyze the natural fractures.

As well, the geo-mechanical analysis provided input to other parts of the project.

Well engineering

The main goals of the field development design were to determine:
• the type and number of wells required to optimally produce the gas field;
• drilling fluid type and drilling practices to minimize reservoir impairment, increase drilling speed, etc.;
• the potential for sand production and, if required, a sand control treatment;
• perforation and completion parameters and completion fluid requirements;
• optimum hole, tubing and casing design as well as cementing techniques; and
• optimum well completion design to guarantee desired production rates in a safe and cost-effective manner.

The well engineering process designed for this stage of the study addressed each of these criteria. Given the fact that Kela 2 is a highly overpressured, thick gas reservoir, it also focused on balancing safety and quality (formation damage control) in the drilling and completion phases of field development.

Production engineering

The production engineering methodology chosen was designed to optimize and safely control the production of gas from each well and to transport this gas to a surface production facility, then into a pipeline. Most of the analyses were performed using nodal-based programs taking into consideration gas inflow performance relationships. Different models were evaluated to determine the combination that would best fit the actual flow rates, surface pressures and bottomhole pressures and ultimately predict actual well performance. These models were then used to optimize the tubing design and future production forecasts. A strategy for surface production monitoring in combination with reservoir monitoring was established to ensure proper data acquisition for the subsequent reservoir management process.

Reservoir reserves

A study of the original gas in place (OGIP) and recoverable reserves for the field was made using probabilistic or risk analysis. The value distributions for each parameter came from the results of the geophysical, geological, petrophysical and engineering stages. This analysis revealed the expected range of the reserves and provided a valuable crosscheck on the single value derived from the geo-cellular or dynamic modeling results. As both of these values fell well in the middle of the probabilistic distribution, the analysis gave added assurance of the validity of the results.

Economics and risk analysis

An economic model was built to calculate the gross revenues less operating costs, capital investments, theoretical royalties and theoretical taxes. Probability distribution factors for royalties, taxes and other key factors were used to account for uncertainties during field development and production.

For each economic case, 5,000 iterations were performed in a Monte Carlo simulation with Latin Hypercube sampling. The results for each case included distributions of undiscounted cash flow, rate of return (ROR), net present value (NPV) discounted at 10% and discounted profit-to-investment (DPI) ratio.
NPV was viewed as the main criterion for ranking the various FDP schemes considered. However, the DPI was used to reveal investment efficiencies. A third key factor considered was the lengths of the production plateau periods. Cases with longer plateaus were preferred given the marketing plans for the Kela 2 production.

A variation of the base case was included to model the economics from inception of exploration at Kela 2 field. High-side estimates for the geological and geophysical costs and the exploration and delineation wells were used in the absence of actual costs to reflect a "worst-case" economic picture.

A total of 33 cases were analyzed. The first 14 cases represented possible development scenarios comparing controllable variations: number of wells, completion types, well production rates, field production rates and costs. The next 13 cases demonstrated the economic impact of potential, uncontrollable changes in reservoir parameters: residual gas saturations, water drive mechanisms, faults effects on reservoir flow, overburden stress effects on permeability and lower OGIP.

Lessons learned

The following elements are critical to the success of a multidisciplinary, comprehensive workflow process:
• peer review provides QC/QA as well as new perspectives that can positively impact final outcomes;
• unencumbered information exchange and communication among all team members (client and consultant) is essential;
• project managers and systems that aid coordination among different departments add critical efficiencies;

• an appropriate workflow sequence is:
1. seismic interpretation for formation tops and fault surfaces within the entire stratigraphic section (not just reservoir);

2. geology-generated type stratigraphic sections from petrophysics; and

3. engineering (the first two steps are needed before geo-mechanical work can proceed efficiently).
Acknowledgment

This article was excerpted from SPE 77668, originally presented at the 2002 Annual Technical Conference and Exhibition.

Comments