(356g) Optimization of a Reactive SMB Process Applied to the Purification of Fructose Syrup
AIChE Annual Meeting
2009
2009 Annual Meeting
Computing and Systems Technology Division
Dynamic Simulation and Optimization
Wednesday, November 11, 2009 - 10:30am to 10:50am
The simulated moving bed (SMB) process is a practical realization of a continuous adsorption system in which a counter-current flow is approximated by sequentially switching the ports of a certain number of chromatographic columns connected in series. The successful design and operation of SMB separation process depends on the correct choice of operating conditions and, in order to execute this task, a detailed and reliable dynamic model of the process is necessary. Rigorous modeling of SMB processes is widely documented in the literature (Dünnebier G. & Klatt, K.U. 2000, Klatt, K.U., Dünnebier G. & Engell, S. 2000, Schimidt-Traub, H. 2006 and Kawajiri & Biegler 2008). The focus of this work is to investigate the best system configuration and operation conditions for a reactive SMB process, the Hashimoto process (Hashimoto 1983), which is applied to the enrichment of fructose in fructose syrup.
The standard Hashimoto process is composed by chromatographic columns and chemical reactors which are alternately arranged within the SMB process. In Figure 1, an illustration of the process is provided. Four external streams are present: the feed stream QFe, the desorbent stream QDe, the extract stream QEx and the raffinate stream QRa. These external streams divide the system into four zones, which are also identified in Figure 1. In the Hashimoto process, the reactors do not take part in the simulated counter-current movement of the adsorption bed. In practice, it is implemented by switching the reactors together with the inlet and outlet ports in the direction of the liquid flow. Theoretically, a feed containing a mixture of two components can be completely separated by properly choosing the internal flow rates of each of the four zones as well as the velocity of the stationary phase, i.e. the period at which the ports are switched.
Figure 1: Illustration of the Hashimoto process.
In the considered case, a mixture composed of glucose and fructose is inserted into the system trough the feed port and a product with higher fructose content is collected at the extract port. It is desirable to operate the process in such a way that all the feed is converted into product, what can be realized by recycling the raffinate stream back to the feed, or by not using a raffinate port at all. Moreover, we investigate the potential gains which can be achieved when the process is operated in the VariCol mode, i.e., when the movement of the solid phase is simulated by switching the different ports at different times. Assuming that four columns and one reactor are available to perform the separation, the profit which can be achieved by operating the process by the eight different schemes described in Table 1 were compared for optimized operational degrees of freedom.
Table 1: Different process configurations.
A rigorous dynamic model which takes into account the continuous dynamics of the single columns and the discrete events resulting from cyclic process operation is employed to perform the optimization. Due to the cyclic nature of the process, the dynamic simulation leads to a periodic steady state (PSS) at which the concentration profiles change dynamically within a period but are identical from period to period. In order to find the PSS that leads to the maximal profit, the optimization solver iterates on the design variables in an outer loop which solves this dynamic model recursively. This optimization is difficult because of the existence of non-linear path constrains as well as of flat regions where the objective function has a weak dependency on the decision variables. These factors limit the performance of gradient-based optimizers and may lead to termination of the search before the global solution is found. This issue is addressed here by optimizing starting from several initializations both inside and at the boundaries of the operating range.
The results clearly show that the schemes 4Z-1111 and 3Z-112 are the more profitable ones. When the process is operated with synchronous port switching, these configurations leads to a profit at least 5% higher than the configurations 3Z-121 and 3Z-211. When the process is operated in the VariCol mode, the profits of the different schemes are increased up to 13%. Another important result is that the best solutions for the four zones process suggest a raffinate recycle stream with a flow rate almost equal to zero.
In order to identify the factors that limit the raffinate stream, the case in which the reactor is four times longer was also investigated for the schemes 4Z-1111 and 3Z-112. When the reactor is longer, the raffinate recycling schemes provide the best result. In the VariCol case, the four zone process is slightly more profitable than the three zones process. In the synchronous (SMB) case this difference increases to 10%. This shows that whether or not raffinate recycling is advantageous depends on the design of the reactor. If the reactor is able to convert the extra amount of glucose into fructose, i.e. if it is long enough, it is advantageous to recycle the raffinate stream. Otherwise the raffinate recycle overloads the reactor and reduces the process profitability.
References
1 - G. Dünnebier & K.U Klatt, Chemical Engineering Science, 55, 373-380, 2000.
2 - K.U. Klatt, G. Dünnebier & S. Engell, Mathematics and Computers in Simulation, 53, 449-455, 2000.
3 - H. Schmidt-Traub, Preparative Chromatography of Fine Chemical and Pharmaceiticals Agents, Ed. Wiley-WCH, Weinheim, 2006.
4 - Y. Kawajiri & L.T. Biegler, Computers & Chemical Engineering, 32, 135-144, 2008.
5 - K. Hashimoto, S. Adachi & Y. Shirai, Biotechnology & Bioengineering, 25, 2371-2393, 1983.
Checkout
This paper has an Extended Abstract file available; you must purchase the conference proceedings to access it.
Do you already own this?
Log In for instructions on accessing this content.
Pricing
Individuals
AIChE Pro Members | $150.00 |
AIChE Graduate Student Members | Free |
AIChE Undergraduate Student Members | Free |
AIChE Explorer Members | $225.00 |
Non-Members | $225.00 |