Algorithm for Real-Time Field-Data-Driven Automation for Intelligent ESP-Lifted Multilateral Wells

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.