Liquid purification or separation – Processes – Chromatography
Reexamination Certificate
2000-11-21
2003-11-25
Cintins, Ivars (Department: 1724)
Liquid purification or separation
Processes
Chromatography
C210S662000, C210S739000
Reexamination Certificate
active
06652754
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a process for determining by simulation the optimum stabilization conditions of a simulated moving bed separation system allowing to rapidly reach set values fixed for the purity of an extracted constituent and for the plant efficiency.
The method is notably suited for separation of aromatic hydrocarbons or of optical isomers.
BACKGROUND OF THE INVENTION
In industry, there are many continuous separation processes using selective adsorption of at least one component among several within a mixture of fluids, notably chromatography processes where the property of certain porous solids, in the presence of liquid, gaseous or supercritical mixtures, of retaining, to a more or less high degree, the various constituents of a mixture is used.
Separation or fractionation processes based on chromatography are most often implemented in a device comprising (
FIG. 1
) a series of k chromatographic columns or column sections connected in series (generally 4≦k≦24), forming an open or closed loop. A porous solid of determined grain size, distributed over various beds, constitutes the stationary phase.
Injection points for the mixture to be separated F, comprising at least two constituents A, B and the solvent or desorbent, and fluid draw-off points delimiting most often four zones Z
1
, Z
2
, Z
3
, Z
4
are distributed along this loop, the constituent that is preferably wanted being in most cases either in the extract (Ex) or in the raffinate (Raf). An identical flow of liquid runs through all the columns or column sections of a zone. In a real countercurrent process, a fixed and constant concentration profile develops along a separation column
1
(FIG.
1
), where the position of the points of injection of a feed A+B, of an eluent S, and of draw-off of an extract Ex and of a raffinate Raf remains fixed. Adsorbent solid
3
and liquid
2
circulate in a countercurrent flow. A solid carrying system and a recycling pump P, both placed on the location of the column (at the junction of zones Z
1
and Z
4
where the only species present in the liquid as well as in the solid is the elution carrier fluid), allow respectively to drive the solid from the bottom to the top, and conversely the liquid from the top to the bottom.
Processes known as simulated countercurrent (SCC) or simulated moving bed (SMB) processes allow, as it is well known, to avoid a major difficulty inherent in real moving bed processes, which consists in correctly circulating the solid phase without creating attrition and without considerably increasing the bed porosity in relation to that of a fixed bed. In order to simulate the displacement thereof, the solid is placed in a certain number n of fixed beds arranged in series and it is the concentration profile which is displaced at a substantially uniform rate all around an open or closed loop, by moving the injection and draw-off points.
In practice, successive shifting of these injection and draw-off points is performed by means of a rotary valve or, more finely, of a series of suitably controlled on-off valves. This circular shifting, performed at each period, of the various incoming-outgoing liquid flows in a given direction, amounts to simulating a displacement of the solid adsorbent in the opposite direction.
The main incoming liquid flows are as follows: the flow of feed Q
F
and the flow of eluent Q
El
. The outgoing flows are the flow of extract and the flow of raffinate. At least one of these flows (raffinate, eluent, extract) is withdrawn or introduced under pressure control. The flow of raffinate Q
Raf
is the sum of the incoming flows Q
F
and Q
El
minus the flow of extract Q
Ex
. A recycle rate Q
Rec
adds further to these controlled flows. The respective locations of the four flows around the beds thus define four distinct zones.
An example of well-known processes is the Eluxyl™ process, which is notably described in the following patents: EP-415,822 (U.S. Pat. No. 5,114,590) or FR-2,762,793 (U.S. Pat. No. 5,902,486).
The parameters necessary for operation of a separation loop are difficult to determine because the variables involved in the process are numerous.
For a given supply concentration, these flow rates can be found empirically but the optimum solution lies in a restricted zone of a four-dimensional space (three liquid flow rates and either the solid rate in the case of a real moving bed, or the permutation period T in the case of a simulated moving bed, the raffinate flow rate being deduced from the other flow rates), which can only be obtained by trial and error after a considerable length of time, without being certain that the optimum point has been reached.
In order to find optimum conditions for controlling or dimensioning a real moving bed (RMB) or a simulated moving bed (SMB) separation system, it is preferable to find a model representative of the separation process, taking account of adsorption phenomena, matter transfer and the flow properties of the fluid flowing through the porous solid phase, and to replace the unwieldy empirical approach by simulations. This approach by simulation can however be just as unwieldy if it is not carried out properly.
Processes are generally optimized according to a static optimization, by means of the well-known trial and error technique. One starts from a set of data (flow rates for example) selected a priori. A simulation is performed until stabilized conditions are obtained (hence the term static), and the performances obtained are recorded. After slightly modifying an input datum, a new simulation is performed in order to measure the sensitivity of the performances to this modification. This is iterated for all the data in order to similarly determine their respective influences on the evolution of the performances. The input data are then modified in the sense of a performance improvement.
There are many strategies for carrying out sensitivity calculation and for determining the way to modify the data. However, all of them have more or less the same drawbacks. They require many simulations, therefore an appreciable calculating time. They are blind and do not use knowledge that one might have a priori. They find a solution if they are initialized close to this solution but they diverge most of the time if they are initialized too far from the solution sought. Furthermore, they may <<fall>> into a local minimum and remain within.
SUMMARY OF THE INVENTION
The process according to the invention relates to a simulated moving bed (SMB) type separation system comprising a series of beds containing an adsorbent solid matter, divided into several zones delimited by points of injection of a feed and of an eluent and by points of withdrawal of an extract and of a raffinate, fluid injection means, fluid extraction means, means intended for permutation of the injection points and of the draw-off points, and means for measuring operating variables. The process allows fast determination of the optimum injected fluid flow rates and withdrawn fluid flow rates in order to obtain a purity degree (P) and an extraction efficiency (R) set for at least one constituent in extract (Ex).
It comprises a single dynamic type simulation with a succession of stages intended for rectification of the fluid flow rates at limited time intervals comprising each, after each time interval (T) the steps of:
comparing the respective purity degrees and efficiencies obtained, resulting from situations stemming from simulation stages over a determined time interval (&Dgr;t) where rectifications are applied to at least one flow rate, with those of a reference situation resulting from a simulation stage without flow rate rectification, over the same time interval,
selecting flow rate rectifications minimizing the quantity of impurities converging to the extract draw-off point, as long as the desired purity degree has not been reached, and
selecting flow rate rectifications minimizing the quantity of said constituent converging to the raffinate draw-off point, as long as th
Cintins Ivars
Institut Francais du Pe'trole
LandOfFree
Process for determining by simulation the optimum... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Process for determining by simulation the optimum..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Process for determining by simulation the optimum... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3157433