Robust Integrity Management Planning of Anthropogenic Dense CO2 Pipelines - Leveraging MFL Data Fusion

Abstract Dense CO2 transportation by pipelines is a key element of the CCUS value chain. Currently, there are significant efforts to design safe / practical CO2 compositional limits. Despite this, the historic experience from the natural gas transmission industry shows that, despite all care undertaken, operational upsets could be still expected. The arrangement of CCUS infrastructures into clusters with multiple feeders of various nature also increases the likelihood of anomalous scenarios. Finally, there is not yet a fully operational pipeline transporting man-made CO2. MFL inspection programs are critical to confirm that internal corrosion is safely managed to prevent uncontrolled loss of containments. However, historical and conventional inspection and integrity assessment practices introduce conservatisms and uncertainties that can lead to unnecessary mitigation costs or at the other end of the spectrum potentially unsafe conclusions. These limits are further compounded in the case of CO2 pipelines, for which the combination of complex corrosion morphologies, aggressive rates, and high operational pressures can exert an onerous burden on the fitness-for-service and remaining life. This paper shows how the use of a machine-learning-based data fusion model to aligned MFL-A and MFL-C datasets can help to produce more robust integrity management planning of anthropogenic CO2 pipelines.