Flexible networks to promote the development of industrial symbioses: A new optimization procedure

The development of eco-industrial parks is now proven a local efficient solution to reduce environmental impacts of industrial areas as well as increasing economic benefits. Furthermore, they are a good way to produce goods while reducing environmental footprints by exchanging energies, water and raw materials instead of consuming too much resources; that way it improves circularity of resources with no disruption of production. However, in an uncertain context caused by different factors such as environmental problems, sanitary crises or geopolitical conflicts, the design of an interconnected network must take into account possible deviations of nominal conditions. In this study, an integrated procedure is developed so that an optimal flexible network is designed with a given flexibility threshold. The flexibility is defined as the minimum of the maximal deviations for the nominal inlet parameters that can be handle by the network. Impact of such a flexibility is visible on the total investment cost of the network. Indeed, when adding some flexibility to the network, the investment cost is increased. The integrated procedure developed is able to find a global optimum in a very large computational time (up to 500 h); in this article, two other procedures are then tested over a large-scale case study (15 companies) to validate the approach. One of them consists in oversizing the network so that it can handle large deviations but the increase of the cost is important. The other procedure consists in a decomposition of the integrated approach in two steps: the first one aims at defining the structure of the nominal solution (minimum of the cost) and the second one evaluates several deviation scenarios to take into account disruption of inlet parameters. The results show that the two-steps procedure obtains results close to the integrated procedure with reasonable computational time (up to 40 min). This study helps to provide solutions to design optimal flexible networks, also adapted to solve large-scale eco-industrial parks; it can also be used as an optimization-decision aid tool for the territorial development of industrial ecology.