- Gain accurate well-by-well flow rate intelligence to optimize production and volumes
- Full automation and self-adjustment allow for a more scalable, economically-viable solution for late life assets
- Predictions of wells with very high water cut result typically in 95% accuracy of all three phase flows
Most assets undergo major changes towards the end of their field life, predominantly driven by the shift in characteristics of their maturing wells. One typical characteristic of late life wells is a low reservoir pressure and hence a low tubing pressure drop; another characteristic is the increasing water cut, with wells sometimes producing at water cuts in excess of 99%.
In order to continue running an asset economically, this may result in the drilling of additional wells, the revamping of wells from oil to gas producers, or the switching from pressure support by gas / water injection to natural flow, amongst others. Such changes make production monitoring and optimization models temporarily obsolete. In addition, accurately estimating the oil flow rate when the water cut is very high is indeed extremely challenging. Typically, production models in late life assets are then not further maintained, often driven by the perception that solutions may not provide continuous accuracy but at economically-exhaustive price points.
The fully automated, self-adjusting capabilities of our FLUX VFM apply a combination of optimization and machine learning to estimate and continuously update the reservoir inflow model. The automated tuning of the tubing pressure drop and reservoir inflow performance continuously ensures that the models are up to date, as the reservoir pressure and tubing pressure drop declines, and the tubing flow regime changes. Full automation and self-adjustment allow for a more scalable, economically-viable solution in the market. We therefore speak of our real-time FLUX VFM as a living asset model. Setting up and tuning additional VFMs for new wells is undertaken in a couple of days. Typical late life changes, such as the installation of ESP or the switch to gas lift, can be accommodated in a few hours.
Our hybrid technology combines the predictive capabilities of physics models with the self-adjusting capabilities of machine learning over time. First, the physics models in our transient multiphase flow simulator (FLUX Simulator) are calibrated based on available reference and historical test data, for example fluid data, temperature and pressure, in the well, on the surface or on the topsides. The available, historical well test data is utilized to develop an automated model that self-adjusts over time. New well test data is routinely captured in the system in real-time to update the model.
A pure physics-based solution can deliver very accurate estimates for the gas and total liquid flow rates. In addition, machine-learning models incorporating high-level feature engineering accurately predict the oil/water ratio. The best possible accuracy for the three-phase well rates is attained when physics-based simulations are combined with the machine-learning models. The answer, a hybrid model orchestrated in a unified solution that accurately predicts all three phases independent of constantly changing operating conditions.
This results in water cut predictions of typically 95+% accuracy, which again sets the base to retrieve an accurate and usable estimate of the oil rate at extreme water cuts.
The Value Gain
In late life assets, with declining profitability, the FLUX VFM will provide accurate, reliable and economical flow rate predictions. We have seen that this can typically help improve production by up to 1%. As an example, at a production of 50 mmboe/y with an increase of production by 1%, an additional 500,000 mmboe/y can be produced. At a 70 USD oil price the annual gain will be 35 MUSD.
Our FLUX Solutions, including FLUX VFM, can in addition enable other use cases, including: