Automated intelligent chemical injection technology minimizes corrosion/scaling, extends ESP run life, increases Permian production

DEAN E. GADDY, Permian Production Solutions, Baker Hughes Co. 

Fig. 1. A typical Permian Basin ESP response to dynamic changes in 3-phase flow behavior.

In the Permian Basin, electric submersible pumps (ESP) are designed to operate in wells with rapidly changing three-phase flow behavior, transitioning from gas to liquid within a matter of minutes, sometimes seconds, Fig. 1. Advanced gas handling pump configurations, variable-speed drive control modes and automated back pressure valves are some of the tools used to help extend run lives and maximize production over the lifetime of a well. 

However, rapid changes in three-phase flow behavior can also challenge the ability to inhibit scale and corrosion efficiently and effectively with traditional injection methods. For example, “set-it-and-leave-it" chemical injection methods lack modern digital capabilities, requiring manual set point changes at the wellsite. Meanwhile, “smart” pumps offer improvements via remote monitoring and adjustment capabilities, but they still require personnel to manually reset injection rates in response to dynamic changes in well behavior. 

A new digital approach, implemented on three Lea County, N.M., well pads (with eight wells), uses “virtual” flowrate outputs—derived from operational inputs—to automatically calculate and adjust chemical injection rates in near real time. This capability represents a step-change, compared to current chemical treatment practices. 

CURRENT INJECTION CONTROL METHODS  

Beginning in late 2023, Baker Hughes collaborated with an independent Permian operator in an initiative that combined the expertise of the artificial lift system and upstream chemicals divisions. The initiative aimed to increase oil production through extended ESP run life, achieved by limiting corrosion and scaling through the tailored application of specialty chemicals. 

As part of this effort, the chemicals team conducted field audits and discovered large target variances from well to well. As the team compared current production data with injection rate set points, they found previously established target rates had become invalid within days of being set. In fact, on many wells operating in unconventional formations like Wolfcamp, Bone Springs and Harkey, production levels and phase ratios are in a constant state of flux. In other words, essential chemicals were often used in quantities too small or too large, leading to sub-optimal operations. 

Because of this, minimizing such disparities became a top priority, as it became clear that current steady-state chemical injection practices would not be effective in solving the problem. 

For example, with the set-it-and-leave-it approach, a service technician must manually conduct a chemical drawdown rate (typically monthly) and then set it to inject according to the most recent production data. However, in most instances, the technician would return only to find large target variances, due to the well’s constantly changing three-phase flow behavior and decline rates. 

More advanced chemical injection methods, such as smart-pump technologies, which apply a constant injection rate that is monitored and managed remotely (Fig. 2), work efficiently for wells that have little or no gas interference. However, these methods require constant attention from experienced engineers when operating below the bubble point. In short, the limitations of these methods to minimize target variance discrepancies are readily apparent. As liquid production rates vary from targeted set points, the system either undertreats during events of water influx or overtreats during gassy phases, leading to both increased lifting costs and shortened ESP run life. While a controller can be set up to inject at the highest possible production rate, this would become quite costly on the chemical side, and it could lead to emulsion issues, if saturation levels remain elevated. 

Fig. 2. Steady State Auto Volume Control vs Virtual Production Rate Alignment.

PRECISE CHEMICAL CONTROL 

Fortunately, one of the advantages of ESPs, as compared to other artificial lift types, is the ability to capture unique data sets, such as amps, pump intake pressure, pump discharge pressure and motor vibrations, to understand downhole conditions. Experienced production engineers can then interpret that data, so as to select the best control mode to handle gas interference events. 

With the advent of machine learning, these data can also be put to good use in calculating virtual flowrates, which in turn can automatically align chemical injection rates as well conditions change. When the ESP design models are match-point calibrated to production and pressure data, we now have a means of establishing accurately derived virtual flowrates that can be trended, analyzed, and used as a benchmark mode. 

To achieve near instantaneous virtual flowrate response with the chemical controller, there are three elements that must work in coordination (Fig. 3): 

• A smart pump chemical controller to monitor pump accuracy and performance. 

• Downhole ESP and SCADA data inputs, which are critical for the advanced models that calculate inputs to the controller. 

• An advanced algorithm matched to the ESP design curves. 

Fig. 3. How InjectRT functions across oilfield operations.

INTELLIGENT CHEMICAL CONTROLLER 

The intelligent pumps deployed in this application had to be capable of changing injection rates in direct relationship to changing three-phase flow behavior. That means it must be able to 1) provide continuous calibration and system quality testing; 2) facilitate communications between the control systems; and 3) enable remote or automated injection rate control. 

At its most fundamental level, this intelligent system must replicate the drawdown test that chemical field service technicians manually perform weekly, bi-weekly or monthly. However, in this application, instead of sending out a technician to check the pump factor, the system is designed to remotely check the pump rate at regular intervals to monitor performance, variance and electrical status. Through cutting-edge machine learning techniques, the intelligent system leverages institutional knowledge honed over decades to discern exactly how much fluid the smart pump is capable of displacing and then adjust it to meet the desired rate needed for an optimized chemical treatment level, Fig. 4.  

Fig. 4. An ESP trend plot of virtual flowrates mirroring the chemical pump injection rate.

By repeating this process on a regular basis, the system can immediately identify issues, transmit alerts and, most importantly, control the pump to eliminate chemical variances. With years of learning and in-house experience, the intelligent system can then apply a hybrid solution to adjust chemical dosing rates. 

While closed-loop control of any parameter influencing ESP performance introduces risk, Baker Hughes’ advanced digital systems, leveraged by InjectRT, enable a thoroughly robust command chain from cloud to wellsite, ensuring that the injection control is always performed in a safe manner. 

AUTOMATED CHEMICAL TREATMENTS  

By themselves, neither smart pump technology nor machine learning virtual flowrate calculations are new concepts; both have been used for many years. However, what makes this novel approach unique is the integration of the two technologies, as part of an advanced digital solution. These intelligent injection pumps can now automatically calculate and deploy optimized chemical treatments in real time. 

In February 2024, Baker Hughes worked with the operator to implement the InjectRT™ intelligent chemical treatment optimization service, powered by Leucipa, on three wells in Lea County, N.M. Injecting a combination water-based corrosion and scale inhibitor at 150 ppm showed promise. 

Comparing virtual production rate output to SCADA and LACT production numbers showed accuracy at greater than 90% for wells producing below the bubble point. For those wells operating above the bubble point, accuracies within 5% of actual were observed. 

Moreover, focusing on water phase inhibition, the operator recorded significant daily cost savings, as a result of the injection rates slowing down during gassy conditions, Fig. 2. If the operation had simply allowed the pilot well to run in smart pump mode, applying a steady-state injection rate from the prior week’s average, the system would under-inject during certain production rates and conversely, over-inject during gas events. 

After seven months, corrosion coupon results coming in from the use of intelligent chemical-injection technology continue to show negligible metal loss using this technology. 

BENEFITS 

The intelligent chemical treatment optimization service, which ensures the proper chemical dosage by instantaneously aligning the chemical injection rates with the VSD-derived measurements, was implemented successfully and is being rolled out by the operator to other wells to increase production. This approach also significantly reduced the required number of wellsite visits—along with the associated reduction in fuel consumption, HSE risks, and carbon emissions—and allowed those field personnel to be redeployed on other applications and wells. Field trips were further reduced through the automated monitoring of tank levels and continuous quality checks on the injection pump efficiency and solar and/or battery performance. 
 

Fig. 5. Chemical Injection halts when ESP shuts down.

As an added cost saving for the operator, the system automatically shut down the chemical pump when the ESP was not operating and restarted it immediately when production flow resumed, Fig. 5. Finally, the system provided an analytic capability for optimizing and troubleshooting purposes through the integration of ESP, SCADA and chemical pump data. 

The InjectRT technology, powered by Leucipa, provides an innovative approach to addressing persistent operational challenges in upstream unconventional fields, delivering optimized operations and sustained production rates and unlocking value from short-cycle barrels. 

About the author

DEAN GADDY is a 45-year oil and gas industry veteran with extensive experience in West Texas, the U.S. Rockies, Russia, Ukraine, China and Australia. Having worked for Baker Hughes since 2011, he has also worked on the operator side for Unocal, EOG, Devon Energy and BP. He holds a bachelor’s degree in geology, a master’s degree in petroleum engineering and an MBA in Technology Management.