Alexander Fokin, Acting Exploration Assurance Dept. Director, International Projects and Exploration Division Group, Upstream
The concepts of ‘exploration risk’, ‘uncertainty’ and ‘probabilistic evaluation’ play a key part throughout the exploration process. These are critical factors that provide the basis for planning, assessment of exploration and appraisal (E&A) success rate, screening of new license blocks and regions, selecting E&A well locations, prospect and license ranking. Appropriate methodologies are the key to credible definition of uncertain parameters.
Multiple risk factors that control the technical and commercial success of an oil and gas project can be roughly grouped into technical, commercial, organizational, and political risks. The technical category includes subsurface risks that are used to describe the probability of presence or absence of hydrocarbon accumulations for E&A planning as well as potential quantity and quality of hydrocarbons.
Subsurface analysis and E&A planning are based on such interrelated concepts as ‘probability’, ‘risk’ and ‘uncertainty’. Probability and risk describe the likelihood of something happening or not happening, or of a factor being present or absent. These parameters are reciprocal and are expressed in fractions of a unit (Probability = 1 – Risk).
Uncertainty describes a situation where the expected outcome cannot be definitely predicted because there is a range of possible outcomes. Subsurface risk and its inverse parameter, probability of hydrocarbon discovery (chance of success), stem from the uncertainty in the geological structure and in the history of an area under evaluation. Greater geological complexity and lack of data translate into increased exploration uncertainty and risks. Hydrocarbon exploration and appraisal are based on the analysis of subsurface information of broadly varying quantity and quality. So the accuracy of risk assessment depends on the availability of comprehensive and reliable data. Any hydrocarbon development project is largely constrained by subsurface risks, which makes their evaluation a key element of the exploration process.
TNK-BP Risking Methodologies
Prospect analysis aims to achieve two principal objectives: assess the probability of a hydrocarbon discovery and evaluate its potential resources. Petroleum industry worldwide uses a range of approaches, most of which are part of the exploration workflow in ТNК-ВР.
Regardless of the specific methodology, each approach is based on a structured subsurface analysis of hydrocarbon play elements that control hydrocarbon presence. The risk analysis is the closing phase of a comprehensive study of geological data.
Geological and geophysical investigations should be designed to define the following prospect risk factors:
» Container (trap) presence risk
» Reservoir presence and effectiveness risk
» Seal presence and effectiveness risk
» Hydrocarbon charge access risk
A number of methodologies are used to quantify and visualize each risk factor.
Standard Risking Process
Prospect risk analysis based on critical risk elements measured as coefficients is the most common technique, which is an integral part of the E&A planning process in ТNК-ВР. For an exploration prospect (i.e. a potential hydrocarbon accumulation contained in a reservoir formation, which is expected to be productive), each risk factor is quantified (in unit fractions) and then all the individual risks are multiplied to produce a composite prospect risk (Table 1).
To a considerable extent, such analysis relies on expert judgement as the evaluation of key risks is generally qualitative and based on comparisons of geological parameters between similar prospects within a limited area. Geologists largely draw on their expertise in the area of study and each may come up with their own view of the summary risk for a particular prospect that could be vastly at variance with other estimates. Logically, the discrepancy is likely to be yet bigger between prospect analogues in different regions.
Note that most oil and gas companies that run their own exploration business face this dilemma. Composite risks provide valuable inputs at different stages of E&A planning and impact prospect ranking, and subsurface teams should make sure that their risk analysis is as comprehensive and objective as possible.
Exploration Division has developed a subsurface risk template to standardize the prospect evaluation process and minimize subjective influence. It is a matrix containing key prospect risk factors: the presence of trap, reservoir, seal and charge access. Each factor is defined by seven to ten critical parameters with possible variation and values in unit fractions. These numbers are multiplied to estimate key risks and overall prospect risk.
As an example, we can look at eight controls of trap presence divided into three groups. The first group includes parameters related to data used for the subsurface analysis and its quality: seismic data, the number of wells around the prospect, and vertical seismic profiling (VSP) data. The second group comprises parameters that describe the type of trap and confidence in the trap closure. The third includes the parameters, on which the credibility of structural mapping depends, such as vertical closure versus the error margin of structural interpretation, the quality of seismic data, uncertainties of time-to-depth conversion, and the accuracy of ties between seismic reflecting horizons and known stratigraphy (based on well data). Each of the eight parameters has a specific range of variation.
In the course of the analysis, each parameter should be assigned an appropriate value that reflects the characteristics of the evaluated prospect. The product of multiplied individual risks is the composite prospect risk. This approach will help narrow down the variation between risk assessments made by different specialists.
Common Risk Segment Mapping
The lateral variation of geological risks across evaluated horizons is visualized on common risk segment (CRS) maps reflecting the distribution of reservoirs, seals and hydrocarbon source rocks. Depending on the parameters, different areas on CRS maps are colour-coded with red, yellow or green for highest, medium and low risk respectively. Individual CRS maps (of reservoir, seal risks, etc.) are multiplied to produce a composite CRS map, which in combination with a relevant structural map helps high-grade the most likely hydrocarbon prospects. CRS mapping is based on gross depositional environment (GDE) maps, seismic facies, geochemistry data, and geographic information systems data. The CRS mapping workflow is shown in Fig. 1.
The CRS approach has the benefit of providing a consistent framework of subsurface risk analysis, which can be applied across large areas to identify exploration prospects with the highest chance of success.
The CRS methodology is widely used in Tyumen Petroleum Research Center (TNNC) for building risk maps based on regional GDE maps to plan exploration and appraisal drilling. Further on, these maps get increasingly detailed and accurate as new seismic and well data become available.
Probabilistic method has become a standard practice in TNK-BP for pre-drill resource evaluations for exploration prospects and appraisal targets. It is used to generate a probability distribution to support base case, upside and downside resource estimates that are subsequently used as a basis for commercial project scenarios.
Unlike the deterministic approach, this method recognizes that there is uncertainty around the estimates of each of the input parameters. While using conventional reserves equations, the input variables are assigned values from statistical distributions rather than averaged parameters. The estimate calculations are run with specialist software tools for Monte Carlo stochastic modeling and their output is plotted as a reserves distribution curve where each probability value from 0 percent to 100 percent corresponds to a specific volume of potential hydrocarbon resources.
Although the probabilistic method has been globally applied as a proven tool, its application requires caution, for instance, with deriving the overall probability for several reservoir formations and combining probabilistic estimates with subsurface risking. This is where typical mistakes are frequently made by specialists in many companies.
Geologists who are new to the methodology have no problem with using software like Crystal Ball for probabilistic analysis. It takes about an hour to learn the functionality, but in order to make sure the generated results are meaningful, it is important to follow some rules of handling both input and output probabilistic data.
Table 2 shows an example of probabilistic resource estimate for a prospect containing three reservoir units, with volume predictions made for each unit and expressed in probabilities (percentiles): Р10, Р20, …, Р90. In this example, the total prospect resource represents the sum of volumes based on specific percentiles for each target horizon, a mistake made by exploration teams when this methodology was being introduced in TNK-BP.
This problem is described in a publication posted on the SPE website (see E.C. Capen. “Probabilistic Reserves. Here at Last?”, SPE, 2001), which discusses the rules of probabilistic analysis and a proper way of estimating total resources of a stacked pay prospect. In Crystal Ball, along with probabilistic evaluation of individual reservoir horizons, it is necessary to make a probabilistic estimate of the total amount in all the horizons, which will be considerably different from a simple addition of individual volumes. While oil resource probabilities in separate horizons are independent of each other, the total estimate will show a much narrower distribution range. This can be easily explained by rules evaluating probabilities of independent events. In our case, the probability Р that at least one of the three independent events with probabilities of Р1, Р2, and Р3, or that their combination will occur is as follows: Р=1-(1-Р1)*(1-Р2)*(1-Р3). With this approach, the overall probability increases relative to Р1, Р2, and Р3, and the uncertainty range becomes narrower. If horizons are dependent due to their geological characteristics, the probabilistic evaluation has to take into account the relationships between parameters that describe the geological linkage, resulting in a wider distribution range of the total hydrocarbon amount.
Probabilistic Reserves and Subsurface Risks
Also noteworthy is the method of calibrating probabilistic evaluation against subsurface risking. Fig. 2 shows a distribution of unrisked total oil-in-place (OIP) for a prospect. It clearly illustrates the relationship between probabilistic estimate and prospect risk (chance of a hydrocarbon discovery); the subsurface risk is essentially a qualifying coefficient for the probability axis, and not the OIP axis.
Risk factors are estimated and applied to forecast reserve additions and describe the probability of hydrocarbon discoveries. Correct risking of expected volumes minimizes the variance between plan and actual reserves delivered by subsequent E&A drilling. If reserves estimate is based on the conventional deterministic method, the overall prospect volume should be multiplied by the risk factor. With the advent of probabilistic tools, exploration teams in TNK-BP commonly fell into the pitfall of risking probabilistic outcomes in the same manner.
Deterministic results are average values that give the right answer when multiplied by risk. But applying risk factors to resource volumes that correspond to Р10, Р20, …, Р90 probabilities is a flawed methodology. This is where Crystal Ball helps to get things right. At first, its output may seem strange. For example, with a risk of 0.5, Р50 or lower probability volumes can be zero. But note that the chance of success of 0.5 means that there is a 50-percent risk of prospect failure.
Smart risk assessment helps companies avoid direct economic and reputational losses, leading to reduced market capitalization, so the application of efficient risking workflows to all aspects of subsurface management is a critical activity for any oil and gas business.
Published with thanks to TNK-BP and Innovator Magazine