"Chrom.Sim": simulation software for modeling chromatographic separations for the sugar industry.

New applications are being developed for the sugar industry that use industrial scale chromatographic separations to produce a range of chemica1.s. Computer models provide excellent tools ,for assessing and optimising the separations. A package of prograntmes has been developed fir modeling four principle systems available to the I sugar industry. The paper outlines the modeling of the four systems:l. Single column batch 2. Simulated moving bed 3. Discontinuous moving bed and 4. Sequential chromatographic. The U.S. corn sweetener and beet sugar industries are presently among the largest users of industrial chromatographic systems, but new applications are on the drawing board also for sugarcane and refinery systems as well as wood and biomass based speciality sugars and sugar alcohols (polyols) separations. Several designs of various degrees of complexity have been available from coinmercial suppliers for years (Saska, 1996), yet an objective evaluation of each design's effectiveness for a particular separation is difficult if not impossible as the effort required for piloting is considerable and little if any published data exist that would allow comparison of the various systems in comparable applications. Computer models can be excellent tools that, within limitations, provide for unbiased evaluation of the process, optimization of the process variables, quick answers to "what if '?" questions to their effects on the recovery and purity of the products and frequently as a tool for educating the personnel in fundamentals of the not-soobvious processes. A versatile, user-friendly software package of programs was developed for modeling of four principle systems for chromatographic separations available to the sugar industw: a single-column batch separation ("PULSE"), a conventional twoor threeproduct simulated moving bed chromatography ("SMB"), a discontinuous inoving bed chromatography ("DSMB") specially designed for three-fraction separations, and the sequential chromatographic process ("SCP"). Each model requires a number of parameters, most of them common to all four programs, such as the system (column) dinl'knsions and configuration (Figure I), flow rates, parameters describing the behavior of the major components maximum of four are allowed which is usually found sufficient to approximate behavior of the real, multicomponent feeds sucrose, ash, color, invert, betain, raffinose, etc. The isotherm coefficients are generally not available to the user but are best obtained by matching results from quick simple experiinents with those from the PULSE program. The numerical integration algorithm at the heart of the programs is common to the four programs. Thus a good experiment-model match in PULSE provides confidence for reliability of the prediction for the other systems that are not readily accessible to the industry for experimentation. This manuscript has been approved for publication by the Director of Louisiana Agricultural Experiment Station as &lanuscript ~Vunrber 99-61 -0099. Zone: 0 I II 111 IV Number of colurnnr: / b --componqnts: Suqors Invert Ash 0rg.nonsus. cqncentratbn of cornponeno In fgmd [q/106m0: Number of w i r n ~ t s in cdt~mp Unoar Psrpmear at c lon4nhtbn Isdthrrm , .--. r* P o --. ,P.o"-'. qwdr24 pprnmebr efconohffatlon Irgthcrm: F.0 b 0 F~OW. dWnt nuact lnvsn WCIQ p w m t b s ~ I ~ ~ I N : ~ ~ j ~ ~ l ~ F -Column Ienlthttml: I T Column di~me&r tcml: -.-. Brd voldage tepsl: / ~ . m lntegrarlon time Incrument Imld: IF---Run tlme tminl: [bma.o swlhhlnq time [mink $G---.-lntarval f a prlntlng detalled results [mink rj3.0 --Animation interval Cmlnl: Ihm.0 Purity and recovery graph tlmo rtsp fmlnl: jb20.0----. Convsrgsncs 1 matchins bottom af zone I with top of zone 0 (concsntratlon) tg/100m0: !p.oi Convergenca 2 non-linear Isotherm loop (concenQutlon) tg/lOOmO: Fig. 1 : Parameter input table, " S M B program Each program provides several windows, some for easy entry of input parameters (Figures 1 and 2) others for graphical depiction of the dynamic (time-dependent) models. Generally, an "animation" window (Figures 3 and 4) provides display of the system (configuration of the columns, location of the input and output ports, etc) and the concentrations of the principle components of the feed throughout the columns.