A new design methodology recently introduced for looped networks is now applied to branched networks. The methodology is based on cost analysis minimisation, network mathematical simulation under normal and failure states and reliability estimates. It is capable of obtaining not only an adequate value for the global reliability of a network but also the appropiate total design demand. This paper then presents a deeper insight into the comparison between looped and branched networks by applying the methodology to both types of networks and comparing the results in terms of cost and reliability. Considerations regarding available reliability definitions are also included which seem to favor volume-type definitions in practice. Results provide a rational foundation to reveal the superiority of looped networks over comparable branched networks by applying quantitative cost evaluations.

This content is only available as a PDF.