Vineet Chawla, Weatherford
Michael Villemarette and Siddharta Wahab, Apache Corp
The efficient management of gas lift systems is pivotal in minimizing operational costs and maximizing production for a large majority of unconventional wells. By leveraging automated workflows to efficiently build and tune physics based nodal analysis models, operators can optimize well performance and gas injection rates thus reducing operational expenses. A cornerstone of effective gas lift optimization is the seamless integration of real-time data with physics-based models. Automated assisted workflows streamline this process which enables continuous optimization of gas lift injection rates to compensate for changing production rates, gas liquid ratios, and reservoir pressures.
The author emphasizes the value of having an evergreen tuned well model to optimize every gas lifted well. Optimization can be realized in some cases by increasing or decreasing gas injection, as the model often shows over injection can reduce production. The challenges in realizing the value from a physics based well model for every well include staff time to build and maintain the models, time to tune the models, and time to make gas injection rate adjustments. The gas lift optimization workflow presented requires significantly reduced engineering staff time by letting automated processes continuously complete the majority of the workflow.
Automated Model Building
In order to efficiently build physics based well models for hundreds of wells, a unique data loader was developed through a collaborative effort between various teams. This process merges wellbore, completion, and production data from multiple databases into a centralized staging table used to create the model. Any missing model data such as fluid gravities, reservoir pressures, and pipe roughness factors are manually entered by the engineer to complete the well model generation. This workflow dramatically reduced the time required by engineering staff to build well models. In addition to building the initial model, the data loader automatically updates the model with any changes made to a well following workover activities.
Automatic Model Tuning
To keep the model evergreen, software automatically tunes the model using every well test. The Inflow performance relationship (IPR) and Vertical lift performance (VLP) variables are derived from the nodal well model, while Injection Rate, Tubing Head Pressure (THP), Casing Head Pressure (CHP), Water Cut (%) and GOR are extracted from production test data to construct an updated gas lift well performance curve. This performance curve facilitates the gas lift optimization process by ascertaining whether there is an under- or over-injection.
Autonomous Control Logic (ACL) through Enterprise
ACL, which was created through a collaborative effort of subject matter experts and computer programmers, was designed to use the tuned model’s performance curve to determine optimum injection rates for each well. The ACL accomplishes this by running solutions at rates above and below current injection rates and solving for total fluid rates and oil production. Based on these results, and parameters set by the Operator within the ACL control interface, the system automatically suggests an optimum injection rate. The frequency of optimization runs can be easily defined by the Operator but is typically done every 4 hrs as the ACL continuously adjusts to optimize the gas injection rate.
Results, Observations, Conclusions
As a result of this automated workflow, Operators can much more efficiently have all gas lift wells modeled, automatically tuned, and automatically optimized for production and associated gas injection rates. As a result of applying this workflow, Operators can realize either reductions in gas injection rates with no loss in production or incremental oil production associated with incremental gas injection.
In conclusion, the deployment of this highly automated workflow can create significant value for Operators by allowing them to efficiently utilize physics-based models to continuously optimize their gas lifted wells. Future improvements include enabling full ACL logic to continuously adjust gas injection rates via automated control valves without human intervention.