UCLA

This work addresses for the first time in chapter 1, the synthesis of globally minimum volume reactor networks featuring SFR/MMR, with the same normalized residence time density function. Global optimality is ascertained by demonstrating that the input-output information maps of SFR and MMR with general RTd/RTD models satisfy all properties required for the application of the IDEAS to the RTd/RTD reactor network problem. The resulting formulation is shown to possess a number of novel properties, which can be used to facilitate its solution. The proposed methodology is demonstrated on three case studies featuring SLFR model in which the Trambouze reaction scheme is carried out.

In chapter 2, the concept of NRT is defined, as a production normalized, capital cost measure for a reactor network. For networks consisting of CSTR's, PFR's, and RTD-SFR/MMR, described within the IDEAS conceptual framework, it is shown that NRT is independent of the network's inlet flowrate. The novel concept of Network Residence Time Constrained Attainable Region for Reactor Networks is then introduced. It is shown to be a convex set, points on the boundary of which are identified through repeated solution of increasingly accurate finite linear program approximations of infinite linear programs. A case study featuring a network of reactors in which the Trambouze reaction scheme is carried out.

In chapter 3, the optimization of an isothermal monolith reactor is carried out. First, a reaction-diffusion 3-D mathematical model for a monolith reactor is developed. The analytical nature of the obtained solution enables the optimization of a multi-channel 3-D monolith reactor to be carried out. The obtained optimization results are discussed and conclusions are drawn.

Chapter 4 presents a novel method for determining reaction kinetics using the reaction invariant reduction method for any set of complex chemical reactions. An effective way for reducing the dimension of chemical reaction mechanisms and to predict the kinetics from a given set of data. The new method will be developed based on a convex formulation of the associated optimization problem. A case study on the Trambouze reaction scheme carried out in a PFR will be used to illustrate the proposed methodology