(587p) Single-Objective Versus Multi-Objective Optimization of Integrated Fermentation and in situ Product Recovery Based on Time-Dependent Fermentation Models

Process simulation provides a platform that can be used to systematically analyze the process at hand, evaluate the feasibility of different process alternatives, and provide the most informative scenarios that can be experimentally investigated to aid in process design and optimization. The batch fermentation is characterized by low final product concentration, yield, and reactor productivities as a result of product toxicity to the microorganisms. Approaches to reduce product toxicity include integrated batch fermentation and in situ product recovery techniques, such as gas stripping, adsorption, membrane separation, etc. A broad range of experimental conditions in terms of gas recycle flow rate relative to fermentation volume, initial substrate concentration, gas stripping start time relative to fermentation start times, etc. have been investigated experimentally for fermentation and in situ gas stripping processes. Process optimization will reveal the optimal conditions of the process to guide the design of high performing integrated fermentation and in situ gas stripping processes across the broad range of experimental conditions investigated.

Typically, there exist a trade-off among the performance parameters of integrated fermentation and in situ gas stripping processes. For example, the total product produced increases with increasing gas flow rate whereas the concentration of products in the stripped stream (condensate) decreases with increasing gas flow rate in an integrated fermentation and in situ gas stripping process. The trade-off in these performance parameters stem from the fact that the parameters are conflicting; conflicting parameters mean as one parameter is enhanced, another parameter is simultaneously made worse. This phenomenon typically leads to conflicting objectives in optimization of the process. Single-objective optimization translate a multiple objective system into one objective system by either using relative weights to combine different objective functions into one objective function or choosing one objective function as the main objective function while transforming other objective functions into additional constraints. In contrast, multi-objective optimization is able to optimize multiple objectives simultaneously even in the presence of conflicting objectives, revealing multiple optimal solutions relative to the single-objective optimization that finds only one optimal solution.

To this end, this study will use and compare (a) single-objective optimization and (b) multi-objective optimization approaches for fermentation and in situ gas stripping processes under different conditions. An unsteady state batch fermentation and in situ gas stripping (simulated with a batch reactor in Aspen Plus linked to a Fortran user kinetics subroutine) will be connected to the multi-objective genetic algorithm in MATLAB for the multi-objective optimization. This study will highlight the advantages of multi-objective optimization to reveal and understand the trade-off that exist among conflicting objectives over traditional single-objective optimizati