Abstract
Process synthesis is the initial stage in process design. There are potentially many different flowsheets that can be created for transforming the raw materials into products. Process synthesis endeavors to find the more favorable flowsheets with respect to the process objectives. Due to the market potential for purified proteins, especially therapeutic proteins, there is currently great interest in large scale protein production systems. As the protein separation process can account for a substantial portion of the total production cost, there has been a lot of activity in trying to develop more economical unit operations and processes for protein separations. In this research, a process synthesis strategy has been developed for separating soluble, predominantly non-aggregated proteins from bioreactor broths. A generic flowsheet is established and broken up into the following six processing levels: cell separation, cell disruption, cell debris and nucleic acid removal, concentration/dilution, low resolution purification, and high resolution purification. A depth first search is performed where knowledge in the form of heuristics, experimentally determined information and the results from unit operation models are applied at each processing level to determine which alternative appears most favorable. Once this initial base case flowsheet is created, other alternatives at each of the levels that also looked promising are considered one by one to create other alternative flowsheets. These flowsheets can then be compared to establish which is the most favorable flowsheet. The strategy is implemented, where possible, in a Turbo C program on an IBM compatible PC. This includes the generic flowsheet, the heuristic knowledge and the unit operation models. Short-cut calculation models are used in order to reduce the computer run time and the amount of experimental data required. The need for obtaining some critical information experimentally makes the strategy cumbersome but the results are more justifiable than just using heuristic knowledge or making inferences based on the values of a few major properties.
Fitzpatrick, John Joseph (1992). Process synthesis for the large scale separation of soluble, predominantly non-aggregated proteins from bioreactor broths. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -1292999.