Artificial Lift Techbook 2019

Issue link:

Contents of this Issue


Page 51 of 59

50 | April 2019 | ARTIFICIAL LIFT: TECHNOLOGY A nalytics has a very obvious attraction for the oil and gas industry—how can more wells be managed effectively with fewer experienced engineers? If analytics can help improve performance in sports such as stock car racing and football, or even diagnose when an elevator needs maintenance, all based on real-time data, how can analytics be used to improve the ability to produce oil wells using artificial lift? An article published by Bain & Co., a global management consultancy, states that "analytic advantages could help oil and gas companies improve production by 6% to 8%," which is what operators typically gain when they implement Artificial Lift Performance's artificial lift optimization software. When using artificial lift, oil companies pay to achieve a certain drawdown and achieve additional production. If the artificial lift method underperforms, production is lost. Using analytics, it is possible to analyze and diagnose the artificial lift performance for every production test automatically and identify when the lift system has a problem, resulting in lower production. Operators can spend about $10 million fracking and com- pleting an unconventional well. The goal is to then produce the well at an appropriate rate to ensure rapid payout. When a well is planned to produce 2,000 bbl/d and then only produces 1,500 bbl/d, production is lost. Artificial lift analytics can tell whether this is a result of a well inflow problem or an issue with the artificial lift. The ability to identify poorly perform- ing artificial lift systems is critical to optimizing production. Gas-lift analytics Ideally, a downhole pressure and temperature sensor should be installed on all gas-lift wells. Every production test could then be analyzed to verify the injection point and address the challenges: determining how much more production could be achieved by more and deeper injection. On existing wells that do not have a downhole sensor, a good practice is to perform a gradient survey monthly to verify the injection depth and ensure that the well is producing optimally. Because of the uncertainty around inflow performance relationship (IPR) on unconventional wells, traditional nodal analysis cannot predict well performance. Instead, it is import- ant to have a top-down process to verify injection depth and predict well performance. Unconventional wells can often have 10 to 15 mandrels in the wellbore to account for inflow uncertainty and to be able to produce the depleted well later in life. Having so many mandrels in the wellbore means the design information has to be extremely accurate, which inevitably results in valves staying open when they should be closed and injection at multiple valves (also known as multi-pointing). Figure 1 shows a gas-lift well with a production test and data from a gradient survey. The software determines that the deepest possible injection point is at mandrel 4 and that all the mandrels above that point are open. The software also predicts if additional production can be obtained by injecting more gas in the upper four valves versus injection at the orifice and compares the results. Performing these calculations automatically every time there is a production test on the well allows for constant screen- ing and flagging of wells where • Injection is very shallow; • Flowline differential pressure is excessive; • Back pressure is high; • The injection rate is low and may result in slugging; • The injection rate is high; or • The injection flowline is plugged/closed (hydrates). Having this knowledge constantly across all wells is the holy grail of production optimization and is the true value of analytics with respect to production optimization. ESP analytics In a report by Stephen Rassenfoss in the Journal of Petroleum Technology, it was indicated that gas-lift production losses can go undiagnosed. However, there is a similar phenomenon with electric submersible pumps (ESP). If an ESP is worn, has a blocked intake, is running in reverse or has deposition, it can be running but not producing as it should. Most operators never identify these scenarios, which results in lost production. An ESP, like any other artificial lift method, should reduce bottomhole flowing pressure. Any time the ESP has a problem such as deposition in the pump or pump wear, the ESP creates less drawdown and well production is reduced. Using analytic tools, it is possible to analyze the pump performance every production test and quantify pump degradation, identify poorly performing pumps and provide recommendations related to recuperating a lost amount of production. Figure 2 shows a well that has declined rapidly in produc- tion over three months. The operator believed the decline was associated with the well's IPR. However, implementation of the ESP analytic tool proved categorically that the production Better leveraging data can result in improved production. By Sandy Williams, Artificial Lift Performance The Advantages of Analytics

Articles in this issue

Links on this page

Archives of this issue

view archives of Playbooks - Artificial Lift Techbook 2019