| Autor: |
Rail Salimov, Benoit Jaffres, Jamal Alblooshi, Arlen Sarsekov |
| Rok vydání: |
2022 |
| Zdroj: |
Day 3 Wed, November 02, 2022. |
| Popis: |
In offshore field due to pessimistic design in one of the field, water injection flowmeters are not capable to measure full potential injection rates. Highly precise Virtual Metering methodology was developed in-house to estimate rates real-time using advanced analytics and machine learning algorithms. Main application of developed virtual meters are improvement in water injection allocation and Injectivity Index monitoring. For Injectivity Index monitoring a comprehensive methodology was developed to minimize uncertainties such as reservoir pressure and friction forces. Virtual metering algorithms are based on choke position and pressure loss across the choke (Injection Header Pressure – Well Head Pressure). For all choke sizes with available data a quadratic regression (dP = A*Q^2+B) is fitted to observed data. For intermediate choke position a polynomial is used to estimate "A" coefficient for any choke position. In case when no data is available to calibrate choke curves, additional step is introduced: total well head tower rate is estimated based on pressure loss between the wellhead towers. Developed algorithms showed very high precision and accuracy. Backtested algorithm showed < 2.5 % error between estimated and measured rates. Developed Virtual Meter was main enabler for injecting at full capacity without limiting injection rates which had positive impact on field injection and production potential. Practical application of VM are rates allocation as well as injectivity index estimation. Injectivity index calculation methodology discussed in the paper takes into account transient and interference effect by building numerical model in well test interpretation software. Friction forces in the tubing are estimated with prosper model or pressure drop at the beginning of any PFO. Matched tubing friction curve will be used for continuous BHP calculation and precise II estimation. Transient II estimation is very important to understand injectivity index evolution and prompt any injectivity restoration actions if necessary. Such Virtual Meters are novel in ADNOC group as discussed field is one of the first pilot green fields with available real-time data. Virtual Meters have zero CAPEX and OPEX as only existing hardware and software is used. Developed Algorithms can be upscaled in any asset as proven to be as a success story in discussed field in Offshore Abu Dhabi. All developed algorithms are embedded in commercial real-time monitoring systems and used for daily monitoring. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
