Page 534 - Fundamentals of Water Treatment Unit Processes : Physical, Chemical, and Biological

Adsorption 489

process variables for sizing are HLR and L(bed). The latter, 15.4.2.1.3 Assumed Data
i.e., L(bed), depends on the duration of run desired, i.e., The assumed data may take on any values desired but
t(run) ¼ L(bed)=v wf . The general ranges, usual in practice, usually are constrained by practical guidelines such as given
2
are 4.9 HLR 24.6 m=h (2.0 HLR 10 gpm=ft ) and in Table 15.7.
3.0 L(bed) 9.0 m (10 L(bed) 30 ft), respectively
(USEPA, 1973). . Adsorbent, e.g., Norit, Fitrasorb 300, Wittcarb 950,
etc. The GAC is selected based upon the pore struc-
15.4.2 DESIGN PROTOCOL ture, cost, and results of isotherm tests or upon given
data from the literature.
An objective of process design is to size the reactor columns,
. Hydraulic loading rate, HLR. The HLR is assumed
i.e., to determine the total packed-bed cross-section area and
the number of reactors in parallel; the total length of packed as a means to size the total bed area, i.e., A(bed) ¼
Q=HLR.
bed, i.e., L(reactor) and the associated number of reactors in
. Length of saturated zone of reactor, L(sat). The
series. Ancillary objectives include: to select a length of run,
value of L(sat) is assumed as a basis for calculating
i.e., t(run); to determine the layout of the system, whether the
t(run). If t(run) is very long, e.g., t(run > 300 days,
reactors are upflow or downflow; and to estimate the cost of
then a shorter value for L(sat) may be used. Alterna-
the virgin carbon for the packed bed. Incorporation of a
tively, t(run) may be assumed and L(sat) may be
design algorithm into a spreadsheet facilitates exploration of
calculated.
alternative scenarios.
. Length of wave front, L wf , Ideally, L wf would be
15.4.2.1 Spreadsheet Layout obtained from pilot plant data. If assumed, then an
Table CD15.8 is a spreadsheet that incorporates a design estimate of L wf 1.0 m is probably adequate, espe-
algorithm for a GAC reactor column. Inputs are required at cially if L(sat) is large and if t(run) is long, which
several points along the sequence of steps and are categorized would compensate for uncertainty.
as ‘‘laboratory,’’ ‘‘given,’’ ‘‘assumed,’’ and ‘‘calculated’’ and . Number of columns, n(col), to be operated in
are explained in the following paragraphs. parallel. The value of n(col) operated in parallel
depends upon such factors as the maximum bed
15.4.2.1.1 Laboratory Data
area desired per column, redundancy, i.e., n(col)
The laboratory generated data include the following: 2, economy, etc.
. Isotherm constants for the given adsorbate–adsorbent
pair, e.g., K and 1=n for Freundlich isotherm 15.4.2.1.4 Calculated Results
. Particle density, r s , and porosity, P (or, alternatively The calculated values, as given in the spreadsheet, are listed
apparent density, r(apparent)) below.

15.4.2.1.2 Given Data . The adsorbent capacity to adsorb adsorbate, X*(C 0 ),
1=n
The given data describe the circumstances of the situation and i.e., Equation 15.19, X* ¼ KC . The Freundlich
0
include the following: relation is given because values for the coefficients
appear in the literature most frequently.
. Adsorbate species, e.g., TCE, chloroform, benzene, . Velocity of wave front, v wf , Equation 15.48,
etc. The calculation protocol, i.e., Table CD15.8, v wf ¼ HLR C 0 =[X* r(1 P)].
applies only to a single adsorbate. . Length of run, i.e., time duration until breakthrough
. Influent adsorbate concentration, C 0 . The value of C 0 occurs, t(run), Equation 15.46, i.e., t(run) ¼ L(sat)=v wf .
may vary, depending upon the situation. For the . Length of packed-bed portion of reactor, i.e.,
purposes of design, a non-varying value of C 0 is L(reactor), Equation 15.50, L(reactor) ¼ L(sat) þ L wf .
used, as a rule. . The total area of the packed bed for all columns,
. Again, the flow to the plant, Q, is whatever occurs, A(total), Equation 15.2, A(total) ¼ Q=HLR.
but may also be set at a certain level, depending upon . The area of a single column, A(col), in a collection of
the circumstances. packed beds that are operated in parallel, A(col) ¼
. Unit cost of GAC. The unit cost of the GAC is usually Q=n(col).
based on a bid price and depends upon quantity. . Diameter of one of several parallel columns, D(col),
For reference, a bulk price range for virgin activated of area, A(col), i.e., D(col) ¼ (4=p) A(col).
carbon was given by USEPA (2000) as $1.54–2.64= . Volume of GAC packed bed, V(bed), that includes
kg ($0.70–1.20=Lb). all beds in system, V(beds) ¼ L(reactor) A(total).
. Unit cost of GAC regeneration. For large installa- . Mass of GAC in all packed beds of system, M(car-
tions, on-site regeneration is usually recommended. bon), i.e., M(carbon) ¼ r(apparent) V(beds).
For smaller installations, the carbon may be trans- . Cost of virgin GAC, i.e., Cost(GAC) ¼ unit cost
ported for regeneration by special tanker truck. M(carbon).