Huybert Groenendaal (EpiX Analytics) speaking at the Palisade Risk Conference in New Orleans late last year offered some insights into building ‘decision-focused’ @Risk models based on his experience with a large multinational client. First, at the design stage, a model needs to be appropriately sized for computability and still offer the desired output resolution. Also, every iteration must be a possible future scenario. Often, modelers fail to take account of relationships between inputs that may bias the Monte Carlo results. To help keep modelers on-track, EpiX offers ‘Risk Optimizer,’ a free @RISK add-on that helps modelers chose the best method for characterizing relationships between variables. Failure to do so typically results in the under-estimation of risk.
One large multi national oil and gas company called on Groenendaal when they found that their attempts to optimize a large drilling program were taking three days to run and that ‘the answer wasn’t that optimal!’ Model constraints included budget, rigs, human resources and access (some sites could only be reached in the winter). The company’s large optimizer included Monte Carlo simulation for uncertainty evaluation. Risk Optimizer allowed the client to check if model constraints were being met early on in a simulation, avoiding unnecessarily long run times and to identify and capture scenarios that were producing better results. The approach greatly reduces the number of runs compared with a brute force approach.
Mark Krahn (Revay & Associates) presented his company’s work on a variety of oil and gas projects such as the Keystone XL pipeline, the Fort Hills and Horizon oil sands and Sunrise SAGD projects. Risk management is ‘the fastest growing area of interest in project management.’ Krahn’s presentation included some spectacular images of what happens when things go wrong—from the 2007 Dubai Infinity Tower flood.
Sujay Karkhanis showed how Woodside uses @Risk to improve ‘cost confidence’ in major projects. Woodside has developed an @Risk-based methodology for new projects. The Woodside cost confidence modeling tool (CCMT) has been used to build an international exploration portfolio characterized by ‘materiality and depth,’ with an emphasis on emerging petroleum provinces. Novel locations and projects make for limited control over market conditions and mean that ‘cost and schedule certainty will be paramount.’ Conventional single point estimates will be inadequate for decision making. The CMMT process has been successfully deployed on projects ranging from $20 million to ‘mega.’
A presentation from Palisade, including a working spreadsheet, showed how oil offloading from an oil rig in the Gulf of Mexico can be optimized in the face of weather disruptions. The facility serves as an inventory stocking point for several other platforms. On a daily basis oil may be shipped to the coast, sent through a pipeline or stored at the rig. A downloadable Excel spreadsheet lets users play with the model using a weather behavior generator that leverages historical data. Access the Palisade presentations.
© Oil IT Journal - all rights reserved.