One of the jobs of exploration and production (E&P) asset teams is to make long-term commitments of funds in an environment where uncertainties exist that can increase project risk. Usually made during the planning phase, these decisions involve the asset's wells, facilities, scheduling and production strategies. Often, the uncertainties lead to undesired outcomes relative to predictions of performance such as Net Present Value (NPV), rate of return, cumulative oil production, and gas plateau period.

Over the 20 years prior to 2003, the E&P industry's overall Return-On-Investment (ROI) averaged less than 7% despite the fact that when individual assets were approved for production, the minimum ROI required to gain approval typically has a hurdle rate of 18% or more. Such wide disparities between actual and predicted ROI have become commonplace.

In order to make sound E&P investment decisions, asset teams need a decision management system that optimizes E&P decision-making by allowing the discovery of more optimal reservoir development decision-making alternatives in uncertain environments so associated risks can be managed.

Landmark's DecisionSpace™ Decision Management System (DMS) was developed to address the E&P industry's need to narrow the gap between the predicted and actual performance of an asset, from both a production and economic perspective. The approach is to perform integrated simulation of multiple reservoir, drilling, facilities and economic scenarios. . While no technology (including DMS) can guarantee a favorable outcome, DMS enables E&P professionals to make faster, better qualified decisions with reduced risk to asset performance.

DMS is unique in that it looks at all aspects of an E&P asset, from the subsurface, to well planning; from facilities scenarios to production prediction and economics and evaluates real field development scenarios using rigorous tools like earth modeling and reservoir simulation. It is designed to help asset teams integrate uncertainty, risk factors and business objectives into the field development planning workflow. Given a set of business objectives and constraints, asset teams can quickly evaluate multiple decision alternatives for any prospect, or field and the cycle time for development project planning is accelerated from months to weeks.

DecisionSpace DMS performs integrated simulation of multiple reservoir, drilling, facilities, and economic scenarios to enable faster, better qualified decisions with reduced risk to asset performance

Integration Across Disciplines
The traditional asset decision-making process requires that geosciences and engineering disciplines follow a sequential work process that generates "best estimate" asset models which are fine-tuned and potentially subjected to Monte Carlo economics analysis.

Traditional Approach

Unlike the traditional process, the DecisionSpace™ DMS system integrates across the E&P value chain and allows high fidelity and technical rigor in each of the chain's components. DMS uses probabilistic methods to provide a better understanding of the potential distribution of outcomes for the entire asset. Mathematical optimization methods show the best scenarios when given a set of decision drivers and constraints. The result is improved asset development planning and asset performance. The following case studies best illustrate this solution.

DecisionSpace DMS approach

Case Study 1: Better Decisions Under Uncertainty
Background: An offshore operator's reservoir is a faulted structure covering about 30 square miles and has a 450 ft thick oil column at about 5,000 ft subsea depth. The operator wants to evaluate the range of Net Present Value (NPV) as a consequence of the underlying state parameter uncertainties for each of many possible development choices (decision variables) and better understand the important parameter sensitivities.

Results: Large uncertainty ranges, as seen from Monte Carlo simulations, mean that a few iterations with fine resolution models will not be sufficient to resolve the uncertainty in the NPV. In order to refine the NPV distributions, a re-evaluation of the well plan scenarios, the injector well locations and scheduling would be a possible next step. Reducing the uncertainty in the relative permeabilities and fault transmissibilites also might help reduce the uncertainty in the NPV. Consequently, drilling the next appraisal well close to a fault and obtaining core samples could also help reduce the uncertainty in the NPV.

Conclusion: The case study presented above evaluates risks to production and NPV for a set of project decision choices and seeks to understand the major sensitivities. Numerous underlying uncertainties are included. The process described leads to an understanding of how uncertainties drive overall evaluation of the value of the asset and the need for multiple scenario analysis.

Case Study 2: Optimizing Multiple-Field Scheduling and Production Strategies with Reduced Risk
Background: At the beginning of 2002, a company is considering developing a new asset with three separate reservoirs—Unit 1, Unit 2 and Unit 3. Each unit can be produced with up to four wells connecting to individual gathering centers (GC). The company's asset team wants to determine the best way to invest in the asset, but is constrained by an existing pipeline with a capacity of 20,000 STB/D over the next 10 years. Ultimately, the team wants to maximize NPV based on three decisions: (1) which year (2003 through 2012) should production begin for each unit; (2) what are the number of wells to be drilled in each unit (0 to 4); and (3) what is the oil processing capacity for each GC (0 to 20,000 STB/D). The company also wants to manage the potential risks.

The actual number of possible decision combinations for unit scheduling and the number of wells is 125,000 (103 dates x 53 wells); in addition, there are the three continuous GC rate capacity decision variables. Obviously, the company cannot evaluate all possibilities in a decision tree. In practice, the company might consider a small number of the decision alternatives, based on its best judgments and experience, and then apply a probabilistic analysis to account for uncertainty for some of the cases.

Instead, using the parameters above, the company employs the DecisionSpace DMS to look at all aspects of the asset and then generates possible decision combinations that maximize the asset's NPV while managing risk.

Results: The results from the DMS optimization solution are evaluated in three different scenarios about the asset, each reflecting different assumptions about the asset's characteristics and uncertainty that lead to different asset development decisions for the asset team to consider.

Scenario 1 – Given that Unit 1 has high reservoir volume, productivity and low development costs, Unit 2 has medium characteristics (volume, productivity and development costs), and Unit 3 has low volume, and productivity, but high development costs, the decision solution, as indicated by DMS, is to drill and produce Unit 1 early, start up Unit 2 to help fill the pipeline, and produce Unit 3 only if it can add value relative to its higher cost. The NPV is shown to be $247M with Unit 3, but $279M without Unit 3. Thus, producing from Unit 3 the first year is less valuable than not producing it.

Scenario 2 – Given that Unit 1 has high volume and productivity and development costs, Unit 2 has medium characteristics, and Unit 3 has the smallest volume, productivity and development costs, DMS indicates that all units should be produced, but that Unit 1 startup should be delayed until 2005 and that the number of wells to be drilled in Unit 1 should be reduced and production capacity of Unit 2 should be increased. This yields an increase in oil recovery of 4 million bbl, an increase in NPV of about $20 million and a decrease of cash out of $65 million, and an increase of rate of return from 18.15 to 21.6%.

Scenario 3 – In this scenario, the Units have the same characteristics as in Scenario 2, but now the company must manage risk, Unit 1 has high uncertainty, Unit 2 medium uncertainty and Unit 3, low uncertainty. Also, the company mandates that development decision alternatives meet its risk tolerance: that the standard deviation of NPV must be no greater than 15% of the mean NPV. Using these constraints, DMS indicates that producing all three units together has a risk of 20%, producing Unit 1 in isolation has a risk of 47%, but producing Unit 3 by itself only has a risk of 12%. The DMS then searches for a combination of production start, well number, and rate that maximizes mean NPV, while meeting the overall risk requirement of 15%. Ultimately, DMS finds a solution with about the same mean NPV, but which meets the risk tolerance, requires much less cash outlay, and produces a higher rate of return. Taking into account the company's risk requirement, the optimal decision is to only drill two wells on Unit 1 at a relatively low production rate of 8,000 STB/D, and to produce Unit 2 at 6,500 STB/D.

Conclusion: The case study presented above illustrates the complexities involved in making realistic asset planning decisions and the competing tradeoffs and multiple uncertainties that must be considered. Even the most experienced engineers are limited to evaluating only a few alternatives using common processes such as decision trees. Optimization using DecisionSpace DMS improves the evaluation process by searching a very large number of alternatives quickly and, at the same time, accounts for data uncertainty to manage risk.