Key points are not available for this paper at this time.
Abstract Offshore and onshore oil rigs generate power from sets of diesel generators, which significantly contribute to upstream and drilling-related emissions. These generators often operate inefficiently due to the "all or nothing" fashion used to maintain a safety margin of power despite highly variable rig loads. This paper introduces a novel artificial intelligence (AI) framework that integrates data pipelining, machine learning (ML), and evolutionary optimization techniques to dynamically model and optimize the emissions profile of generator schedules based on rig load demand. Utilizing historical operations data from over 50 rigs worldwide, the framework forecasts power demand in real-time and classifies rig states to inform generator scheduling. This dynamic approach allows for load balancing, minimizing fuel consumption and greenhouse gas emissions by operating only the necessary generators at their best efficiency points (BEP). Comprehensive simulation results demonstrate a reduction in fuel consumption and emissions exceeding 20%, with some rigs achieving up to 30% savings. This innovative framework represents a significant step towards more efficient and environmentally responsible energy use in the oil and gas industry.
Marzban et al. (Fri,) studied this question.