The unconventional oil and gas industry has made a huge impact to North American energy output. The segment has been successful at increasing the recovery rate from wells to a point where production rates continue to increase while the rig count decreases. A significant contributor to this trend is the adoption of new technology and innovation in fracking techniques. To continue the pace of efficiency the fracking industry will need to address reliability and integrity issues to ensure production is maintained and maintenance costs are optimized. This is especially true as the productions from unconventional wells are being extended from months a few years ago to years in scale.
The growth in the unconventional industry has been so rapid that equipment and infrastructure are being redeployed or reused in places where they were not intended to be used beyond the immediate production forecast. The original design basis for this equipment may not have accounted for such a scenario which results in significant uncertainty in the equipment's integrity and reliability. Inspection and test data may be missing or incomplete for the same reason. Combine these two effects with more recent innovations in technology and fracking techniques then there is the possibility for loss of containment and releases for an industry that is highly scrutinized by the public and regulators.
To address the issue of managing integrity for large scale infrastructure where uncertainties exist in design criteria, condition and operating environment DNV GL has piloted the use of Bayesian networks in a model called MARV. The MARV model is a risk assessment methodology that utilizes sensor data; physics based models, past failure statistics and expert knowledge to create probabilistic forecasts. The method overcomes the limitations of incomplete or imprecise data and offers a very accurate prediction of where failures may occur.