Page 299 - Applied Process Design For Chemical And Petrochemical Plants Volume II
P. 299

288                      Applied Process Design for Chemical and Petrochemical Plants

                                                                  more sensitive this condition becomes as the extreme right
                                                                  hand side of the graph is approached.
             a                                                      Kister’s [93] study indicates that flooding with the cur-
             Li                                                   rent  newerdesigned  packings  occurs just  below  2  in.
             I                                                    water/ft of packing, somewhat below the prediction given
                                                t                 by  the earlier Eckert chart, Figure 921C or D. The data
                                    Design max. oper.’   D,       plotted by Kister indicate that for larger packings (2 in. and
                                    oapacily = 8O%F   Min. HEW
                                                                  3 in., for example) the flooding point is noted at much
                                                                  lower pressure drops than suggested by Figure 9-21C. The
                            c,=v[      Pr PG 1                    pressure drop at flood point has been found to be inde-
                                                On5
                                        pG
                                                                  pendent of the Flow Parameter on the charts, but does vary
                                                                  with the packing family and packing types [93,95].
            Figure 9-22.  Typical  HETP curve illustrating operating and  design
            relationships.  Nomally  reasonably constant  over  wide  range  of   Kister [93] has developed a new approach at establish-
            vapor flows.  Note: C,  = capacity factor; V = gas mass velocity =   ing the flood point that appears to suit the available data
            Ib/ftz/sec. Adapted by permission from Strigle, R. E Jr. and Rukove-   and is  obviously more accurate than reading the upper
            na, F.,  and reproduced with permission of the American Institute of   curve on Figure 9-21C.
            Chemical Engineers, Chemical Engineering Progress, Mar. (1 979); all
            rights resewed.
                                                                  APflood - 0.115 Fp0.7 (do not extrapolate below Fp = 14)   (920)
                                                                  where APflood = pressure drop at flood point for all random
            (text continucdbm page 284)
                                                                               packings, in. water/ft  of packing
                                                                           Fp = packing factor, empirical, based on packing size
              Thus, for mass transfer performance design a specific            and shape, l/ft,  see Tables 9-26A-E.
            design HETP value should be established, which in effect
            represents the range from point B through E for C, values   By calculating the @flood,  the capacity parameter can be
            above point F,  the HETP values will be greater (and thus   determined using the calculated flow parameter and Fig-
            less efficient contact).                              ure 9-21E, and, if available, the SIX (Shenvood, Leva, Eck-
              Some recent evaluations of data by  other investigators   ert) charts of  Reference 93. For  Fp > 60, the calculated
            indicate that a so-called loading region does not exist as   APflood  using the equation coincides with Eckert’s Figure
            clearly as may be suggested by some data, and therefore   9-21C flood line. Figure 9-23 illustrates the relationship of
            they  suggest  operations essentially up  to  the  flooding   packing factors to flooding pressure drop and is represent-
            point. For a good, reliable design that must allow for fluc-   ed by  the APflood equation. The predicted results of  the
            tuations in  feed, and possibly column back-pressure up   equation are k10 to 15% based on the plotted data [93].
            sets, designing to the flood region cannot be recommend-
            ed. Design limits are discussed later.                Foaming liquid Systems
            Flooding Point                                          For an accurate design, the effects of foaming of the Iiq-
                                                                  uid as it flows through the random packed/structured bed
              At the second sharp change in the slope of the pressure
            drop curve, Figure  9-20, the packing tends to  hold up   should be known, estimated, or determined by experimen-
            more and more liquid as the gas flow increases. This cre-   tation. There is little published data on the subject except
            ates a rapid increase in pressure drop. The flooding point   Eckert [24] and Strigle [82]. Generally, foaming systems
            of the system is said to be the point of the second inflec-   produce higher pressure. drops than  non-foaming, most
            tion of the pressure drop curve. Here the liquid build-up   probably due to the blocking of packing voids by the foam.
            on top of the packing becomes increasing higher and the   Therefore, it  is  wise  to  determine the  foaming related
            pressure  drop  essentially goes  to  infinity  for  a  finite   nature of  the specific system, and because there are no
            increase in gas rate. In many actual cases operation can be   numerical data published, the designer must use judgment
            maintained at the flooding point, but it will be erratic, the   and make allowances for a pressure drop greater than read
            performance  (efficiency) of  operation  poor,  and  the   from any of the charts, perhaps even 2 to 3 times greater.
            entrainment carry-over excessively high. It is obvious that   Hsu  [ 1241  presents equations for  directly calculating
            towers are not designed for flood point operations, but at   random packings based on published data and which are
            -0%     of gas and liquid rates associated with this point.   adaptable for computer programming and thereby study-
            Figure 9-21D indicates that the flooding region usually is   ing the effects of variables. The basic data are essentially a
            above 2.5 in. water/ft,  but note how cramped and much   match with Figure 9-21D.
   294   295   296   297   298   299   300   301   302   303   304