Reactor network synthesis with guaranteed robust stability

• We propose a dynamic model serving as a superstructure for reactor networks without integers.
• Existent as well as non-existent reactors are rigorously treated in the stability analysis.
• The discontinuous behavior of the spectral abscissae of the model is analyzed.
• A method for solving semi-infinite mixed-integer problems with eigenvalue constraints.
• A case study with up to 10 reactors in the superstructure is successfully treated.

This paper proposes a systematic approach to design reactor networks with guaranteed robust stability. The approach is based on the superstructure approach for reactor network synthesis. A structured dynamic model for reactor network modeling is formulated and embedded in a MINLP with robust eigenvalue constraints. Design parameters, structural alternatives and parametric uncertainty are considered simultaneously as design degrees of freedom. Structural alternatives result from decisions on the existence of reactors and flow connections in the superstructure. Parametric uncertainty may either result from model uncertainties such as reaction kinetic constants or heat transfer coefficients, or from process uncertainties including slow disturbances in load or quality of raw materials. A tailored two-step solution strategy is proposed to tackle the robust mixed-integer optimization problem. A case study with five continuous stirred-tank reactors (CSTR) and five plug flow reactors (PFR) is presented for illustration.