Steady-state design of CO2 pipeline networks for minimal cost per tonne of CO2 avoided

Abstract

Transporting captured carbon dioxide (CO2) to injection sites can be a significant component of the costs of CCS. This cost can make the difference between an economically viable and an unviable CCS project. Therefore, optimising transport costs is important in minimising overall costs. Although the timing of the deployment of CCS network components is a highly relevant aspect of network design, it also increases the complexity of the design problem significantly. As a first approximation, this paper describes a procedure for obtaining near-optimal CO2 pipeline networks with minimal cost per tonne of CO2 avoided, under the simplifying assumption of fixed steady state flow. The effect of a staged deployment of CCS over time is therefore not considered for this analysis.The procedure presented in this paper is a high-level decision tool that can help determine which characteristics of a CO2 transport network are most important from the point of view of reducing the cost of avoiding CO2. The design procedure allows the specification of multiple emission sources, capture plants and injection locations, and does not assume a predefined fluid velocity inside the pipelines, but uses a genetic algorithm for minimising the cost of the network per tonne of CO2 avoided. The total cost of the CCS network includes the costs of building, operating and decommissioning the capture plants, pipeline network and injection sites.As a case study, the procedure is used to design a CCS network for a set of emission sources in the south-eastern Queensland region in Australia. Different emission sources and injection locations are considered. All emission sources are assumed to commence deployment of CCS concurrently and all pipelines and injection facilities are assumed to begin operating at the same time. The pipeline network obtained from applying this procedure is compared to existing pipeline routes.The near-optimal network transports all of the captured CO2 in a branching network and stores the CO2 close to the emission sources. It was found that the near-optimal network design for this case is largely insensitive to variations of the modelling parameters within typical ranges.