The start of an electric submersible pump (ESP) is the most dynamic event in the life of the ESP, and one that has been shown to be the main contributor to the premature failure of the ESP; yet it is clearly unavoidable. This article introduces an algorithm comprising of a model-predictive controller and a moving horizon estimator for automating the well startup. Objectives and constraints related to the startup are considered for the whole well system, including the reservoir, the ESP, the tubing etc. A lumped-parameter model is established to model the fluid dynamics in the system. The estimator recalibrates the model and provides estimates (virtual measurements) in lieu of unavailable physical measurements. The operating sequences for the ESP and choke are then updated step-by-step by the controller, considering the model of the system, the startup objectives and constraints, and the measured feedback information from the wellbore gauges. The startup algorithm was implemented on a field edge device and deployed to a well in the Permian Basin. The algorithm executed two successful startups. A model recalibration was conducted before the second startup which improved the accuracy of setpoint tracking.
Multilateral wells with electric submersible pumps and intelligent completions are notoriously difficult to operate and require long testing and frequent retests due to production condition changes and significant transients resulting from the horizontal undulating drains. For human operators, this task is very time-consuming and extremely challenging given the multidimensional and multi timescale system characteristics. However, the process can be automated via optimisation and control, with the proposed algorithm responding to observed production and system changes throughout the well’s life. To that end, a reduced-order well model is derived and validated with real-well-matched synthetic model data, and subsequently an automation algorithm is developed. This innovative and integrated approach to real-time lift and inflow automated control offers the prospect of boosting operators’ production value and investment returns. The algorithm utilises existing or new intelligent completion hardware and instrumentation and the wellsite-deployable smart algorithm, capable of adjusting to varying well conditions and optimally managing the production throughout the well’s life. To achieve that, the algorithm allocates flow-rate and water cut contributions from each lateral or zone and as such recalibrates the well model on the fly using the real-time field data. We present simulation results using a field-data-matched synthetic model and are working with an operator to implement the technology in the field. All in all, such a data-driven automation to autopilot intelligent production is now within sight and could in the future scale towards multiwell/fieldwide solution.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.