Electrical Submersible Pumps (ESPs) are critical artificial lift assets whose unexpected failure causes significant non-productive time and workover costs, with unplanned shutdowns lasting up to several weeks. Existing monitoring systems rely on reactive, threshold-based alarms applied to surface measurements. These traditional methods cannot detect incipient faults—such as early-stage gas locking, progressive impeller erosion, and developing bearing friction—and struggle to distinguish between faults that produce overlapping, single-channel signal signatures. While the governing equations of ESP operation are well established, their direct application to fault detection remains impractical in the field. Fault-relevant parameters like effective fluid density are not directly measurable from surface instrumentation, and differentiating noisy speed measurements to recover analytical quantities amplifies uncertainty to impractical levels. To address these limitations, this paper presents MLSensor, an edge-deployed machine learning framework that bridges physics-based understanding with data-driven implementation. To overcome the scarcity of labelled field data, a multi-domain Digital Twin was developed in OpenModelica, coupling Kloss motor dynamics with Affinity Law hydraulics. This twin generated twelve labelled simulation runs across three fault types at three severity levels, incorporating Gaussian sensor noise at 10–45 dB SNR to mimic real-world conditions. A 46-element feature vector, including four novel cross-modal electrical-hydraulic decoupling features (such as the highly discriminative decouplingIQ, was extracted per 2-second window. A Random Forest classifier achieved a 100% precision, recall, and F1-score classification accuracy under a temporal train/test split, validating the discriminative completeness of the cross-modal feature representation on simulation data. Finally, the trained model was deployed on an ESP32 microcontroller, achieving a 427 µs average inference latency while utilizing only 23% of program flash and 6% of SRAM, proving the viability of cloud-independent, real-time edge inference.
T. M. Busoye, Q. A. Jokomba, O. M. Busoye et al.· SPE Nigeria Annual Internati...· 0 citations
Sand production has become a significant concern in the hydrocarbon recovery process from unconsolidated reservoirs which may result in equipment damage, flow restrictions, and costly operational downtime in vertical oil wells. Accurate prediction of sand production is vital to optimize well integrity. Conventional geomechanical and empirical models frequently fail to capture the highly non-linear interactions among reservoir pressure, multiphase flow rates, rock mechanical properties, and dynamic operating conditions. This study addresses the identified research gap by developing and systematically comparing four supervised machine learning classifiers for binary prediction of sand production occurrence using routine well-test data from a single vertical oil well in the Niger Delta basin. A total of 235 validated well test observations consisting of 19 recorded variables which include date and operational parameters such as production rates, pressure conditions, choke size, and fluid properties were pre-processed and analyzed. The target variable was formulated as a binary classification problem with the operational threshold sand rate > 0 lb/1000 bbl to enable early detection of any sanding event. Four machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine were developed and evaluated following feature optimization and model tuning. Model performance was assessed using accuracy, precision, recall, F1-score, and ROC–AUC metrics. The results show that ensemble and kernel-based methods significantly outperform linear and single-tree models, with the Random Forest classifier achieving the best model prediction accuracy of 93.62%. This strong performance demonstrates the model's robustness in capturing the complex, nonlinear interactions governing sand production behavior. This study demonstrates that machine learning classifiers can be effectively utilized in a manner that enables proactive sand management strategies, including choke adjustment, artificial lift optimization and selective sand control deployment, ensuring a minimal risk of equipment failure and increasing overall well productivity and hydrocarbon production.
S. E. Balogun, A. Joledo· SPE Nigeria Annual Internati...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
