Abstract
Reducing emissions from offshore oil and gas operations on the Norwegian Continental Shelf (NCS) is critical to meeting Norway's 2030 and 2050 climate targets. This paper presents a decision-support framework that combines machine learning–based forecasting with structured technology evaluation to quantify and benchmark alternatives during early-phase project planning and development as well as development of decarbonization strategies. Machine learning models trained on historical data for production and energy demand are used to forecast future energy demand, greenhouse gas (GHG) emissions, and carbon costs. The method enables rapid screening of decarbonisation alternatives using publicly available data. A selection of 16 topsides on the NCS was analysed using publicly available data. To evaluate forecasting performance, the model was first tested on historical data from these installations over the period January 2022 to March 2025. The model achieved strong predictive accuracy, with a mean absolute percentage error of 4.20% and an aggregated error of 1.76%. The same set of topsides was then used to explore the effects of six decarbonisation technology categories—power from shore, offshore wind, CCS, low-emission fuels, efficiency, and digitalisation—on emissions and carbon costs. The study also introduces a set of evaluation metrics—covering maturity, implementability, safety, and CO2 reduction potential—to qualitatively compare technologies at the field level. The framework is intended to support high-level scenario screening. While this study applies it to selected scenarios on the NCS, broader scenario development is ongoing, and the approach is adaptable for use in other offshore regions.