The state of the art for artificial lift optimization is still largely based on the decades-old Gibbs equations. As seen in numerous other industries, the application of Artificial Intelligence can help operators derive new insights, automate control, and improve efficiency.
Modern Artificial Intelligence engines work by training them using real world data. Rather than simply telling the system what conditions to monitor for, the system learns by analyzing all available data to identify parameters that are leading indicators of an issue, such as gas lock or paraffin buildup, that are not apparent to the average user. In addition to using best in class AI algorithms, Ambyint has the largest and rapidly growing of production optimization data.
Ambyint has a 12-year operating history delivering best-in-class artificial lift control and monitoring solutions to E&Ps and over that time has gathered a high-resolution data set data, currently totaling almost 100 million pump operating hours or 45 TBs, 67,000x larger than our nearest competitor.
Ambyint’s data lake includes over 10 years of operating data from 1000’s of horizontal and vertical wells, sampled every 5 milliseconds, including more than 33 million dynocards.
Enabled by a massive training data set, continuously updated and tuned models enable an inference engine to detect key production issues proactively, including detection, characterization, and prediction of well anomalies or prediction of wellhead leaks.
By digitizing the visual input from millions of dynocards, Ambyint’s AI platform is able to perform micro pattern analysis to diagnose downhole and surface anomalies and optimize well parameters.