A Digital Twin and Machine Learning Approach for Optimizing Coiled Tubing Nitrogen Gas Lift Operations

Abstract In situations where a well experiences cessation of flow due to heavier liquid accumulation in the wellbore, a well kickoff operation may be required to revive the flow. An effective method involves employing coiled tubing (CT) to inject high-pressure nitrogen gas (N2), thereby lifting the accumulated heavier liquid and reducing the hydrostatic pressure in the wellbore prior to restarting production. Aligning both business and technical objectives, optimizing well kickoff operation is crucial to achieve more efficient job, lower cost, improve resource allocation, maximized oil production, and environmental responsible practices. Using extensive well kickoff operation datasets, this paper proposes leveraging big data analytics to identify optimal parameters of successful well kickoff operation. The proposed methodology employs a digital twin model to establish baseline (executed) well kickoff operation parameters (including CT kick-off depths, CT size and N2 volume usage) across diverse executed well conditions (such as water cut, reservoir pressure, gas-oil-ratio, productivity index, fluid properties, tubing size, wellbore deviation and perforation depth). Subsequently, the model is optimized to determine the optimal combination of kick-off depths, CT size and N2 volume that yields the minimum Bottom Hole Pressure (BHP). Both of baseline model and optimized model are trained to machine learning program to predict the future optimized well kickoff operation (CT kick-off depths, CT size and N2 volume usage) based on input well candidate parameters and prioritized well candidates sorted by success rate of well kickoff operation. The analysis results demonstrated that the methodology efficiently improve success rate of well kickoff operation by better well candidate prioritization, maximize oil recovery and enhance operational efficiency. Notably, the optimize solution offers technically viable kick-off depth and reduce volume of N2 usage, facilitating more efficient job executions, lower costs, and reduce environmental impact. The used approach, alternatively, could be utilized to improve success rate and optimize well kickoff operations.