Abstract
Reliability is one of the most important properties of processing systems because of its high level direct effect on the investment and operational costs. Still there is no general method that is capable of simultaneously considering the reliability during the design procedure of processing systems. The main reason of the lack of general method is that the process systems design and the reliability engineering are based on different types of mathematical modeling tools. While process systems design is traditionally considered as mixed-integer optimization, reliability engineering is based on probability theory. The enumeration of the alternative feasible solutions of a redundant processing system is highly combinatorial, and, therefore, it is very difficult if possible to embed it into the widely used mixed-integer programming procedure.
For the simultaneous consideration of the costs and reliability in process systems design, general modeling tool is required that can conveniently cover the two areas. In the present work, it has been shown that the formerly developed axioms based combinatorial approach to process network synthesis (PNS), the so-called P-graph framework, can conveniently cover and integrate these two aspects. The interface between the synthesis and the reliability analysis steps of process systems design is established on the basis of the axioms of combinatorially feasible processing networks. Since the combinatorial feasibility and the structural operability are closely related terms, their simultaneous consideration hardly increases the complexity of the procedure.
The method for processing systems synthesis with simultaneous consideration of reliability is general and capable of effectively designing complex, highly interconnected processing networks with large number of redundant operations. The focus of the present work is on structures of processing systems, all statements and algorithms are general. The solutions of the synthesis of the HDA process have also been given.
