Resilience of Processing Systems: Analysis and Synthesis

Several aspects are considered in designing processing systems, these aspects are usually evaluated via assessment of expected events on the basis of cause-effect relations, e.g., the determination of the output of an operating unit by a simulation program under known values of its inputs and parameters. However, such evaluation is not appropriate for determining the resilience, since resilience is based on unexpected events in addition to the expected ones. Consequently, the cause part of the cause-effect relation is not known or not effective. Natural way of circumventing this issue is to enumerate all possible “effects” as the set of possible failures.

In the current work, the general formulation for determining the resilience of a system is embedded into the P-graph based mathematical modelling framework. The enumeration of all possible failures is performed by the P-graph based synthesis algorithm, i.e., resilience analysis is done by a synthesis tool. The proposed algorithm is general, there is no limitation on the structure of the processing system (e.g., on the number of loops, raw materials, and products) and on the types of operating units.

Embedding this method into the synthesis algorithm of the P-graph framework, resilient processing systems can directly be synthesized. Since most of the critical infrastructures operate as processing systems, the proposed method can contribute to their analysis and design. Three case studies illustrate the method: a process with a single product, a processing system with multiple products, and a network of industrial plants.