Water Production Monitoring in Offshore Gas Fields with Virtual Flow Metering

Hi Community Members!

We’ve recently published a cloud-based Virtual Flow Metering (VFM) system tailored for a late-life well with high water cuts. This solution employs a hybrid physics-data approach, delivering precise and near-real-time estimates of water production per well in gas fields.

Link to the full article: https://www.sciencedirect.com/science/article/pii/S277250812300042X
 

Enhancing Production Optimization

In the oil and gas production domain, understanding the fluid composition and production rates of multiphase operations is crucial for effective planning and optimization. With wells yielding a diverse mix of oil, gas, and water, often with varying compositions, ensuring accurate measurements is essential for maximizing production efficiency. Ratios change in time, and this gets more and more important as the water cut grows in the late part of a water flooded well life. This article explores how Virtual Flow Meters (VFMs) offer a cost-effective alternative to traditional methods like physical multiphase flow meters.

The Power of Hybrid Physics-Data Approach

Unlike purely physics-based models or entirely data-driven approaches, the hybrid physics-data method showcased here leverages the strengths of both. By integrating a simplified multiphase flow model with reference well test data, this approach delivers accurate estimates without the computational demands of simulations or the potential drift associated with data-driven models.

Real-world Application: Offshore Gas Field Case Study

This innovative system was successfully deployed in a gas field situated offshore in shallow waters, consisting of three dry-tree production wells connected to an unmanned offshore platform. The wells produce gas, condensate, and water, with gas flow rates measured at the platform. The produced fluids are combined and transported over 100 km to the onshore processing plant for separation and individual measurement.

Seamless Integration with CDF

The implementation was kickstarted by the seamless integration of the VFM system with a cloud-based architecture, on by Cognite Data Fusion. This integration facilitated real-time availability of field sensor data for data science models, ensuring accurate and timely predictions.

Automated Well Test Identification

In a significant advancement, we developed an automatic detection algorithm for well test campaigns. This algorithm streamlines the process by identifying specific conditions, ensuring accurate well test data without manual intervention.

Efficient VFM Calculations

The VFM calculations are performed remotely on a cloud-based DataOps platform, guaranteeing uninterrupted access to live sensor data. Employing a hybrid physics-data system, this solution accurately estimates water production per well for allocation purposes, using a physics-based model tailored to high gas volume fraction scenarios. This study introduces three mathematical models for VFMs, combining machine learning algorithms with physics principles to ensure physically meaningful solutions. These models were trained on well test data and scheduled to run hourly, providing precise water rate calculations with a granularity of 10 minutes.

Conclusion

In summary, this work showcases the successful implementation of a hybrid physics-data approach to Virtual Flow Meters for monitoring water production in an offshore gas field. The VFM system offers accurate estimates of water rates from individual wells, even with limited well test data. Additionally, the deployment on cloud infrastructure, powered by CDF, ensures continuous prediction, self-calibration, and minimal maintenance requirements.

We'd love to hear your insights on this solution. How do you envision VFMs integrating offshore operations? We hope you’d like to share your thoughts below!