Artificial lift optimization is one of the core activities Production Engineers and Techs are asked to perform on a daily basis. Rod lift is the most widely used artificial lift type, deployed on horizontal and vertical wells alike. Due to the widespread use of rod pumps, consistent best practices have been established over time. When those operational best practices are adhered to, meaningful increases in field profitability of even the lowest producing wells have followed.
Despite their differences in the horizontal development and vertical legacy well context, industry best practices to optimize rod lift around efficiency, and thus profitability, have a consistent logic and methodology. First, wells are diagnosed as underpumping, dialed in, or overpumping. Second, based on the categorical classification, a consistent remediation workflow is applied based on the available levers to pull that either increase production or lower the number of damaging strokes into the system.
The primary daily operational set point levers to optimize rod lift wells are on/off time, SPM, and pump fillage. Following the Pareto principle, these levers represent the lion’s share of the value-added changes that an engineer or tech can make to optimize wells. Thus, the third and final step is to apply a change to the control system (POC, VFD, timer) and observe the before-and-after to judge success.
Ideally, this workflow is performed for every well on a regular basis. However, this involves significant time and resources to accomplish, and even when done daily can still neglect over 99% of the strokes put into a system during a well’s run-life. The reality is that field personnel have too many wells, too little time, and not enough of the right data or technology to enable a step-change in profitability.
The question, then, is: why have humans perform such a sisyphean task? Or rather, in the era of modern technology, why are humans still tasked with carrying out a repetitive, defined logic workflow such as set point changes when we are much better suited to solve more complex and nuanced field problems? Wouldn’t E&P operations be better served following the manufacturing model by letting machines automate optimization set point changes? This would allow operators, techs, and engineers to do what they do best: solve problems, design solutions, and think strategically, while keeping wells dialed in at all times.
This post (originally released as a white paper) outlines how machine learning combined with rod pump domain expertise delivers operators a step-change in operating leverage and optimization capabilities in the context of horizontal wells. If you are interested in reading more about vertical wells and legacy stripper well optimization, you can download our other white paper in this series here.
In the context of unconventional horizontal wells, conditions are dynamic. This makes it difficult for legacy control systems to keep wells properly optimized, and requires significant manual intervention to plug the gap. One operator we work with, a power user of XSPOC, estimates a deep-dive technical analysis of a well in XSPOC takes upwards of 30 minutes to complete using the below optimization workflow (left).
With the quickening pace of unconventional development and an ever-growing well count, the required optimization time per well necessitates the complete focus of the team in order to keep up with the entire field. We know in reality this is impossible due to daily field demands, and leads to the top 20% of producing wells receiving the majority of the staff’s attention, while the other 80% of the base is left largely un-optimized.
Ambyint’s intuitive Production Optimization Platform (POP) design and use of modern technology has collapsed the cycle time for manually analyzing and optimizing wells. Now the equivalent workflow in XSPOC takes less than 5 minutes to complete in Ambyint, as outlined above. On a route of 100 wells, the optimization time drops from 48 hours down to less than 8 hours. This also includes being able to look at more dynocards. Thus, the POP creates a higher probability of accurately making the right decision and eliminating analysis rework.
With a strong foundation built, we are taking it a leap forward by layering on a level of machine learning and deep learning to further reduce and even eliminate optimization cycle time. This feature, called Autonomous Speed Range Management , is enabled by our High-Resolution Adaptive Controller (HRAC), which engages with the POP via its embedded communications protocols and the Cloud.
The machine learning algorithm first classifies whether wells are underpumping, overpumping, or dialed in, based on well conditions over various time series. Then, it runs the proper optimization diagnosis and remediation based on the classification. Finally, it either sends a recommendation to the user to accept and engage the change, or, if enabled by the user, the system automatically applies a change to the SPM set points to optimize well conditions in real time.
Unlike humans, machine learning algorithms are able to look at every stroke and identify well behaviors. But, like humans, machines also require training by domain experts to become truly effective at their jobs. Studies across multiple industries show that even the most accurate machine learning systems still require “humans in the loop” to account for up to 20% of the solution. This is primarily accomplished either through helping label training datasets or correcting inaccurate predictions to continually refine the algorithm. We subscribe to this methodology across both of these functions, having production engineers collaborate with developers and data scientists.
In the example below (Figure 1), a machine learning algorithm simultaneously looks at multiple patterns to detect if a human is walking or running. This parallels well optimization, in which a computer is able to detect underpumping (walking) or overpumping (sprinting) conditions, shown in the platform screencaps in Figure 2.
Fig. 1: A Machine Learning is used to determine walking/running state.
Fig. 2: The Ambyint platform uses ML to detect well state.
Next, by combining domain expertise with machine learning and deep learning techniques, we are able to advance from classification of well types into remediation algorithms for each well. Image classification or natural language processing algorithms work by taking in data, applying predefined system and rules, delivering an outcome, and improving over time. Following the same methodology, we are able to read in multiple streams of rod pump well data and execute a structured optimization logic. This is akin to how an engineer makes an optimization decision, with similar outcomes.
A machine can execute well optimization set point adjustments more quickly, objectively, consistently, and routinely. This is a win-win for field personnel because, since they are unburdened by routine optimization changes, they can focus on delivering higher value tasks. Additionally, management is happy because wells are dialed in.
In early testing of automatic speed range management, customers have seen production gains of up to 10% on underpumping pilot wells with no drop in well efficiency metrics (e.g. pump fillage). On overpumping wells, we are cutting strokes into the system by up to 15% without losing production, which leads to a reduction in power costs and failure rates on a linear, stroke-for-stroke basis.
Ambyint has been delivering best-in-class artificial lift optimization solutions to E&P customers since 2004. Ambyint delivers remote optimization and management that cost-effectively reduces level of effort and increases production from both horizontal and vertical wells on artificial lift. By using a combination of traditional physics-based techniques and modern artificial intelligence, Ambyint automates diagnosis, remotely adjusts operating parameters, and predicts and prevents failures without requiring time consuming human intervention.