Using Artificial Intelligence to optimise pipeline network design
CAPEX of pipeline networks can be minimised through the simultaneous optimization of pipeline sizing, pipeline route selection, and geographic location of the tees and any compression/pumping stations within a network, whilst ensuring that flow and pressure-loss requirements are respected. However, such an optimization process presents itself as a complex problem that is rarely addressed due to it requiring an unfeasibly large number of manual workflow iterations spanning multiple engineering disciplines. This lack of field layout optimization at the concept-select phase can result in invalid conclusions being made when ranking concepts for further development, with consequential loss in project value being incurred at the earliest of phases. Augmented Engineering Ltd is pioneering technology which leverages cloud-scale compute capacity to apply a combination of engineering calculations, bespoke algorithms and artificial intelligence to automate the optimization process. This paper outlines the complexity of the challenges involved and demonstrates how the aforementioned technologies can be applied to automatically determine the optimum network configuration within a reasonable computational timescale, allowing such optimizations to be performed during the early stages of a project lifecycle.