An operator leveraged high-performance edge and cloud computing to achieve a fully autonomous chemical injection system and more than tripled electric submersible pump (ESP) run lives.

Minimize ESP failures due to corrosion and scale

An operator in Ecuador has 148 wells that use ESPs for artificial lift. Corrosion and scale issues were causing an annual ESP failure rate of 20% to 30%. Although the wells were chemically treated to protect against these issues, monitoring the chemical performance was a manual process. Field personnel drove to each well pad to check chemical pumping rates daily, collect produced fluid samples monthly, and retrieve corrosion and scale coupons every 60 to 90 days. Samples and data were subsequently sent to a laboratory for analysis of treatment effectiveness.

To implement any recommendations for modifying the chemical injection plan, engineers had to return to the well pad and manually adjust the injection pump. The delay between data acquisition and remedial action could be as much as 2 to 3 months. Because well conditions are very dynamic, the time lag meant that chemical injection decisions were based on data that was usually out of date, and wells were overtreated or undertreated more than 40% of the time. The results were significant unplanned workover costs and loss of production revenue. A more effective chemical treatment plan was required.

Enhance chemical injection precision via remote, real-time monitoring and control

SLB recommended its Production Chemicals Optimization on Delfi application, which was implemented on two wells in a series of steps.

1. Identify the data required at higher frequency. For corrosion, the rate of water production and the corrosion rate are key, while produced water rate and pressure and temperature changes have the largest impact on scale deposition risk.
2. Determine what elements of this data can be accessed by digitally connecting existing equipment. An AgoraGateway™ ruggedized edge computing device is used to integrate and process downhole pressure and temperature from the ESP controller and wellhead pressure and temperature from sensors already in place.
3. Investigate implementation of data-driven, physics-based models where real-time measurements are not feasible. There is no commercially viable technique for continuously measuring produced fluid rates in real time, so a virtual flow model (VFM) was developed for the two wells, using the Pipesim™ steady-state multiphase flow simulator. Ingesting data into real-time scale and corrosion prediction software applications enables forecasting the risks based on changes in produced fluid rates, temperature, pressure, and water chemistry.
4. Procure and install additional digitally enabled equipment to bridge data gaps. The operator had experienced several issues with the outdated chemical injection pumps over the past few years. They were replaced with new pumps that include a variable speed controller for precise monitoring and control. In addition, because of the lack of an adequate model for predicting surface corrosion rate, a corrosion probe with a real-time wireless transmitter was procured.
5. Develop techniques to ingest data from the operator’s internal systems, such as water chemistry analyses and well test data from its production data management system (PDMS). Water chemistry is required for scale prediction and well test data for ongoing calibration of the VFM.
6. Use the data—process it as close to the equipment as possible and implement a closed-loop control system. An application was developed using Agora™ edge AI and IoT solutions to process in real time
   • produced water rate (from the VFM)
   • scale prediction (from the model)
   • surface and downhole corrosion rates (from the wireless probe and corrosion model, respectively)
   • current chemical injection rate (from the chemical pumps).
   The output is a recommended chemical injection rate. Any discrepancy with the existing rate is flagged in the cloud application, an adjustment can be approved manually or autonomously, and the new setpoint is sent via the Agora IoT solution to the chemical pump.
7. Deploy a cloud-based interactive visualization application to provide full visibility to the end user. Production Chemicals Optimization on Delfi enables real-time visualization of all transmitted, predicted, and calculated data.

More than tripled ESP run life and minimized risks

The new system has been in place for 18 months, and ESPs that were previously failing every 6 months on average continue operating flawlessly. Compliance between actual and target injection rates has increased from 60% to 99%, ensuring continuous optimal treatment, with chemical injection adjusted about 275 times per day on average. The time from detecting a risk to performing the relevant adjustment has decreased by 99%. Monitoring workflows that were typically performed manually on a monthly basis now run every minute, enabling quick detection of system changes and fully autonomous chemical optimization. The real-time models for virtual flowmetering and scale prediction are periodically validated against manual methods and have produced results that agree within 3%.

Currently the solution is deployed on two wells, but with 20% to 30% of ESPs in this field failing annually because of issues related to corrosion and scale, the operator is planning deployment on additional high‑risk wells as well as on new ones. The goal is to realize substantial reductions in opex, workover costs, and production losses.

Field trips are only required if the system detects a mechanical issue, decreasing vehicle emissions by as much as 90% and reducing the overall carbon footprint. Eliminating multiple trips to the wellsite also significantly lowers HSE risk. Potentially harmful events, such as chemical leaks, are detected in real time before secondary containment is breached, further reducing environmental impact.

For more details

Read technical paper SPE 207732.