Reconfiguration of large-scale systems using back-up components

Abstract

Abstract Large-scale control systems tend to present a large number of alternative and back-up elements that, although not used in nominal operation, could be brought into play if necessary. Such hardware redundancy extends the fault tolerant capabilities of the system, while posing the problem of selecting the most suitable new system configuration. Accordingly, this paper aims to formulate the problem of, after a fault occurrence, selecting the (in some sense) optimal system configuration that maintains an admissible system performance until the restoration of the nominal service. The configuration selection is posed as a multi-objective mixed-integer program (MIP) solved using a lexicographic approach. Aiming at reducing the worst-case execution time, the analysis of necessary properties for the existence of an admissible solution, as well as how to manage the information obtained by evaluating such properties to be included in the MIP, are investigated. A portion of a water distribution network is used in order to validate the proposed solution.