Risk-Based Surveillance Planning: A Practical Value-of-Information Approach for Data Acquisition in Producing Fields

Abstract

Summary Surveillance data can be critically important in managing producing oil and gas fields. To maximize value, it is necessary to identify those surveillance opportunities that are not only informative, but materially value-adding. For surveillance decisions in producing fields, applying conventional value-of-information (VoI) methods to every individual surveillance opportunity can be too time-consuming to be practical, considering the large number of relatively small data decisions to be evaluated, each possibly addressing multiple reservoir uncertainties and supporting numerous business decisions. This paper outlines risk-based surveillance planning (RBSP), a simple approach that is based on the observation that the vast majority of surveillance opportunities in producing fields are designed to manage risks. RBSP then evaluates surveillance opportunities with VoI principles, but exploits the fact that much relevant value data already exist as byproducts of risk-management processes. RBSP has been used successfully in dozens of oil and gas fields; two case studies are described herein. Many companies use a risk-management framework that documents risk events and their causes and consequences, assesses risk probability and impact, and identifies prevention measures to reduce the probability of the risk events occurring and mitigation measures to reduce the impact of risk events should they occur. If such a framework has been used, the risk assessment provides an estimate of the expected monetary value (EMV) lost because of a risk event (= probability × impact). Risk reassessment taking into account the planned prevention and mitigation measures reveals the EMV increment attributable to the risk-management measures. RBSP assumes this to be the theoretical value of “perfect” information (VoIperfect) for the package of surveillance that supports these measures. RBSP links each potential surveillance opportunity to the risk-management plan by determining how the surveillance would (a) help create or strengthen prevention measures; (b) detect the impending occurrence of a risk event, thus enabling mitigation measures to be triggered; or (c) help monitor both prevention and mitigation measures to detect any weaknesses that require improvement. The VoIperfect of a surveillance opportunity with respect to a given risk can then be estimated by attributing to it a proportion of the total VoIperfect of the surveillance package by use of a simple criticality score to estimate how dependent the risk-management measures are on each item of surveillance. A reliability score then discounts this VoIperfect on the basis of how likely the surveillance is to deliver the necessary information in reality, yielding the value of imperfect information. Where an item of surveillance affects several risk-management measures and multiple risks, the values are summed, providing a VoI for each item of surveillance. In case studies of an asphaltene-precipitation risk in a deepwater oil field and a water-influx risk in an offshore gas field, RBSP helped to create a logical value-based rationale for surveillance decisions, and the surveillance, when implemented, did facilitate effective management of the risks and added value.

Event-based risk management for subsurface risks: An approach to protect value generation from oil and gas fields

“…The proposed risk management measures included P r e l i m i n a r y V e r s i o n managing production rate and pressure to stay above the (uncertain) asphaltene onset pressure to prevent asphaltene deposition (prevention measure), whilst chemical soak or stimulation treatments could be used to remove asphaltenes should they deposit (mitigation measure). In this particular example, the risk did actually occur (see Grose and Smalley (2016) for more detail), with repeated zonal pressure build-up tests diagnosing asphaltene precipitation in one reservoir interval. The mitigation measures were then implemented, including shutting in the affected zone to prevent further damage and then removing asphaltene build-ups using xylene chemical soaks.…”

“…The risk management plan is implemented in its entirety. Managing the risks effectively should add value to the project or asset; if full implementation of the risk management plan is subsequently challenged (for example on the grounds of cost), this challenge may be rebutted (or avoided) if the risk management plan demonstrates the relation between the cost of risk management measures versus the overall increase in project or asset value they generate (see calculation examples in Grose & Smalley, 2016). 5.…”

“…Where there are risk types that are particularly prevalent at a certain stage of field life, for example, wells-related risks when moving into early production, or recovery factor risks late in field life (Fig. 7), knowledge of the patterns of risk relevance may help with planning appropriate and timely data acquisition or surveillance (Grose & Smalley, 2016). Appraisal wells, for example may provide the only opportunity to gather data that could help manage wells-related risks that will only emerge in early production, or surveillance during early production may be the only chance to obtain data that can help manage recovery factor risks late in field life.…”

Sensitivity analysis applied to fuzzy inference on the value of information in the oil and gas industry

“…VoI has been applied on expensive and complex projects such as 4D seismic acquisition, as shown by Ferreira (2015) using a probabilistic approach. Grose and Smalley (2017) discussed a risk-based surveillance planning method that uses the methodology of VoI in producing oil fields. For fracture reservoirs under waterflooding, Shbair et al (2017) discussed the application of VoI for reservoir surveillance.…”

A practical approach for applying Bayesian logic to determine the probabilities of subsurface scenarios: Example from an offshore oilfield

“…This could provide a quantitative basis for estimates of data reliability that are needed for Value of Information studies (e.g. Coopersmith & Cunningham, 2002;Grose & Smalley, 2017).…”