Predict Distillation
Tray Efficiency
Markus Duss
Sulzer Chemtech Ltd
Ross Taylor
Clarkson Univ.
D
istillation — the most important unit operation
for separating liquid mixtures — is carried out in
columns equipped with either packing or trays.
In a conventional tray column, vapor rises vertically up O’Connell published a graph showing section efficien-
through the column while the liquid flows downward cies for bubble cap trays, which can be represented by (1):
and across the trays in alternate directions on succes-
sive trays (Figure 1). Column simulation and design
software calculates the number of theoretical stages where μL is the liquid viscosity in mPa-s and α is the rela-
and the section efficiency is used as an efficiency factor tive volatility.
to translate theoretical stages to actual trays. The effi- The relative volatility and the liquid viscosity serve as
ciency of a particular section of a column (ηsection) is the the only input variables in calculating conventional distilla-
number of theoretical stages (nth) found in a section of tion tray efficiencies with the empirical correlation devel-
a column divided by the number of actual trays (ntrays): oped by O’Connell (1). This simple method has survived all
24 www.aiche.org/cep July 2018 CEP Copyright © 2018 American Institute of Chemical Engineers (AIChE)
attempts at being replaced by more rigorous models, despite depends on the flow pattern of the vapor and liquid phases.
lacking a sound theoretical explanation for its success. This Most models for cross-flow trays assume that the vapor
article presents a simple, but fundamentally sound, theoreti- flows vertically in plug flow and the liquid is vertically
cal approach that achieves results equivalent to O’Connell’s well-mixed. With these assumptions, the Murphree point
correlation using the same number of variables — revealing efficiency on a tray is defined as (3):
the hitherto unrecognized relationship between O’Connell’s
empirical correlation and other theoretical models. In addi-
tion, the article introduces a modification to O’Connell’s The Murphree vapor-phase tray efficiency (ηtray) defines
correlation that produces a close-to-perfect match between the fractional approach to equilibrium for a single cross-
the updated correlation and the derived theoretical model. flow tray based on the vapor concentrations on the oper-
ating and equilibrium lines. The following assumptions sim-
The traditional Murphree efficiency plify the mathematics: the liquid travels in plug flow across
approach to tray design the tray, the point efficiency in Eq. 4 is constant across the
The section efficiency is calculated based on fundamen- tray, and the vapor concentration below the tray is uniform.
tal mass-transfer considerations, including the approach to Based on these assumptions, the Murphree tray effi-
equilibrium both locally on the tray (point efficiency) and ciency is (2):
on the tray as a whole (tray efficiency). The conventional
method of designing trays (2) employs Murphree’s defini-
tion of efficiency to quantify the approach to equilibrium.
The overall number of transfer units for the gas phase The section efficiency in a distillation column is related
(NOG) is (3): to the tray efficiency by (2):
where NG and NL represent the number of transfer units in When the equilibrium line is steeper than the operat-
the gas and liquid phases (a measure to quantify mass trans- ing line, as is typically the case for the stripping section of
fer), respectively, and λ is the stripping factor calculated as a column, the stripping factor is greater than one and the
λ = mG/L, where m is the slope of the equilibrium line and section efficiency is less than the tray efficiency. When the
G and L are the gas and liquid flowrates per cross-sectional stripping factor is less than one, the section efficiency is
bubbling area in kmol/m2-s. higher than the tray efficiency, which typically occurs in the
The approach to equilibrium that can be achieved with a rectifying section. When the stripping factor is exactly one
given number of transfer units depends on the concentration the section and tray efficiencies are equal.
profile — the driving force for mass transfer, which, in turn, The more rigorous efficiency model in Eqs. 3–6 requires
a large amount of input data: detailed tray geometry, trans-
Vapor to port properties of the liquid and vapor phases (diffusivities,
Condenser
viscosities, and densities), operating conditions (pressure,
temperature, liquid and gas flowrates), and thermodynamic
Rectifying
Reflux
Section
Liquid Overflows
information (the slope of equilibrium line). This infor-
the Weir mation is necessary for the correlation of mass-transfer
Feed Stage n–1 coefficients, interfacial area, froth height, bubble regimes,
horizontal backmixing, entrainment, and weeping, all of
which are required to calculate the numbers of transfer units
Stripping
Section
Stage n
(and therefore the Murphree point and tray efficiencies) that
Vapor from Stage n+1 are needed for column design.
Reboiler Vapor Rises and
Passes through
Combining Eqs. 3–6 yields:
Holes or Valves
Liquid to in Trays
Sump
Copyright © 2018 American Institute of Chemical Engineers (AIChE) CEP July 2018 www.aiche.org/cep 25
Reactions and Separations
1
= 1/α
60 α = 5; μL = 0.08 – 5 cP
0.8
Typical Range
=α m 50
0.7 mx riu in Rectifying
ilib
Liquid Mole Fraction
o
λx =α u Section
0.6
o
Eq =1 40
L /G
0.5 lope 30
e,S
0.4 Lin 20 ηsection (λ) ≈ ηsection (1/λ)
t ing α= 4
ra L/G = 1 10
0.3
O pe
0.2 0
Typical Range 0 10 20 30 40 50 60 70 80 90 100
0.1 in Stripping
Section O’Connell ηsection
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 p Figure 3. A parity plot shows that section efficiencies determined by
Vapor Mole Fraction O’Connell’s correlation agree well with the results of the theoretical model.
The range studied, with the relative volatility varying between 1 and 5 and
p Figure 2. This McCabe-Thiele diagram depicts a binary system with a the liquid viscosity varying between 0.08 and 5 mPa-s, covers the majority
relative volatility of α = 4. of commercial distillation applications
26 www.aiche.org/cep July 2018 CEP Copyright © 2018 American Institute of Chemical Engineers (AIChE)
range investigated is less than 10%. The section efficiencies basis only when such symmetry exists.
obtained from the traditional model were calculated using The terms NG and NL quantify mass-transfer behavior
the stripping factor for x = 0 and x = 1. Astonishingly, sub- in the vapor and liquid phases, respectively. Since the slope
stituting Eq. 10 into Eq. 7 produces nearly identical results of the equilibrium line influences the distribution of the
whether λ or 1/ λ is used. The reciprocal of the stripping resistances to mass transfer across the two phases, assuming
factor often is referred to as the absorption factor. that NG = NL does not imply that the liquid-side and vapor-
It is important to understand this outcome since this pro- side resistances to mass transfer are equal. The fundamental
vides the justification for using the relative volatility instead relationship of the gas-side and liquid-side mass transfer
of the stripping factor to represent relevant behavior. While coefficients is (3):
the relative volatility is always greater than one, the strip-
ping factor is smaller than one when the mixture contains
high concentrations of the higher-boiling-point compound.
Using the stripping factor based on α for x = 0 and x = 1 where KOG is the overall gas-phase mass-transfer coefficient
can, in many cases, reflect the approximate slope at total in kmol/m2-s and kG and kL are the gas-side and liquid-side
reflux in the striping and rectifying sections, respectively. mass-transfer coefficients, respectively, in kmol/m2-s.
The dependency of the correlation on the liquid viscos- The relative liquid-phase resistance (LPR) in a column
ity is very similar for the two correlations. However, the is defined as (4):
dependency on α or λ shows that O’Connell overestimates
the impact of α on the section efficiency when compared to
the model based on the traditional approach.
Figure 5 shows the dependency of LPR on the stripping
Understanding the relationship factor. The liquid-phase resistance increases with increas-
Combining Eq. 10 with Eq. 7 yields almost identical ing stripping factor. For small values of λ, the liquid-phase
section efficiencies for a given value of the stripping factor
and its reciprocal. Figure 4 plots the calculated section effi-
ciencies for three different ratios of NG:NL. To allow a com- Nomenclature
parison, the stripping factor was varied and NG was adjusted c1 = proportionality constant
in order to achieve the same maximum section efficiency to G = gas throughput per cross-sectional bubbling area,
kmol/m2-s
reveal apparent symmetries. kG = gas-side mass-transfer coefficient, kmol/m2-s
In the case of NG = NL (Figure 4), there is an apparent kL = liquid-side mass-transfer coefficient, kmol/m2-s
symmetry about λ = 1. Such symmetry is not present in the KOG = overall gas-side mass-transfer coefficient,
other two cases. However, the NG = NL curve is not strictly kmol/m2-s
symmetrical about λ = 1, as the efficiencies for values of λ L = liquid throughput per cross-sectional bubbling
and 1/ λ other than one are slightly different. The use of the area, kmol/m2-s
relative volatility in O’Connell’s method has a theoretical LPR = liquid-phase resistance
m = slope of the equilibrium line
0.8 NG = number of transfer units in gas phase
N L = number of transfer units in liquid phase
0.7 NG= 0.785 NOG = overall number of transfer units in gas phase
0.6
nth = number of theoretical stages
ntrays = number of actual trays
0.5 x1 = exponent to be applied to the liquid viscosity
NG=1.5
Greek Letters
ηsection
Copyright © 2018 American Institute of Chemical Engineers (AIChE) CEP July 2018 www.aiche.org/cep 27
Reactions and Separations
In(λ)
0.5 λ, the driving force for mass transfer increases. The
0.4 translation of the tray efficiency to section efficiency is
0.3 the reason for the decrease of the section efficiency.
0.2
0.1
Modifying the O’Connell correlation
0
The derived simplified theoretical model is
0.01 0.1 1 10 100 expected to more accurately predict section efficien-
Stripping Factor, λ cies, albeit with increasing complexity. Therefore, a
p Figure 6. This graph shows the relationship between the stripping factor, various simple modification to O’Connell’s correlation better
efficiencies, and the number of transfer units, assuming NG = NL = 1. Reducing reflects theoretical outcomes.
liquid resistance reduces section efficiencies, the opposite of what is expected, as The relative volatility affects the section efficiency
Murphree tray efficiency increases under the same conditions.
in the O’Connell correlation more than it does in
0.8 the theoretical model. Therefore, the exponent on α
ηsection = 0.503 μL-0.226 α-0.08
should be adjusted independently from the exponent
0.7
on the liquid viscosity to better match the theoretical
0.6 prediction.
0.5 Figure 7 shows the relevant range for adjusting
the exponent on α in the O’Connell correlation. The
ηsection
0.4
liquid viscosity was assumed to be 0.2 mPa-s while α
0.3 was varied between 1 and 5. The exponent on α was
regressed and the best-match curve revealed the value
0.2
α = 1/ λ α= λ to be –0.08. The black line represents the theoretical
0.1 dependency of the section efficiency on the stripping
0
factor. The red line is the outcome using an exponent
0.1 0.2 1 5 10 of –0.08 on α for the modified O’Connell method.
Stripping Factor, λ The modified O’Connell correlation (Eq. 2) there-
p Figure 7. The modified O’Connell correlation (red) achieves a good agreement
fore becomes:
with the traditional model (black) for α < 5 when the exponent for α is –0.08. The
liquid viscosity was assumed to be 0.2 mPa-s while α was varied between 1 and 5.
28 www.aiche.org/cep July 2018 CEP Copyright © 2018 American Institute of Chemical Engineers (AIChE)
The new exponent for α is now used to regress the pro- proposed modification. Data points for capacities of more
portionality constant (c1). Equation 10 is thus rewritten: than 98% or less than 25% were excluded from the analy-
sis because the majority of those data points were associ-
ated with reduced efficiency due to jet flood or weeping,
As the parity plot in Figure 8 demonstrates, the modified respectively.
O’Connell correlation (Eq. 13) achieves an almost-perfect Figure 9 indicates that using the exponent of –0.08 for
fit with the combined Eqs. 7 and 14. the relative volatility improves the prediction of section
Based on theoretical considerations, the stripping factor efficiency. Using the stripping factor improves the accuracy
(λ) or its reciprocal (1/ λ) is preferred over the relative vola- further, but, more importantly, it allows Eq. 15 to be used
tility (α). This is because when α is used, it is not possible for absorption and stripping applications as well as for
to distinguish between total and partial reflux conditions; distillation.
the stripping factor helps to account for this. Therefore, the
modified O’Connell correlation can be written as: Eq. 2 Eq.13 Eq. 15
Average Relative Deviation = –17% –6% –2%
Average Absolute Deviation = 21% 16% 15%
where σ = λ if λ > 1 and σ = 1/λ if λ < 1. 140
x O’Connell, Eq.2 +20%
Modified O’Connell, Eq.13
Validation 120 Modified O’Connell, Eq.15
When the O’Connell correlation was modified, the –20%
proportionality constant was retained at 0.503. Since this 100
Predicted Section Efficiency
60
50
Literature Cited
40 1. O’Connell, H. E., “Plate Efficiency of Fractionating Columns
and Absorbers,” Transactions of the American Institute of Chemi-
30
cal Engineers, 42, pp. 741–775 (1946).
20 2. Lewis, W. K., “Rectification of Binary Mixtures,” Industrial and
Engineering Chemistry, 28 (4), pp. 399–402 (1936).
10 3. Lockett, M. J., “Distillation Tray Fundamentals,” Cambridge
Univ. Press, New York, NY (1986).
0
4. Chen, G. X., and K. T. Chuang, “Liquid-Phase Resistance to
0 10 20 30 40 50 60 70 80 90 100 Mass Transfer on Distillation Trays,” Industrial and Engineering
Modified O’Connell ηsection Chemistry Research, 34 (9), pp. 3078–3082 (1995).
5. Chan, H., “Tray Efficiencies for Multicomponent Distillation
p Figure 8. A parity plot of section efficiencies calculated by the modified Columns,” PhD Dissertation, the Univ. of Texas at Austin
O’Connell correlation and the theoretical model shows almost-perfect (May 1983).
agreement.
Article continues on next page
Copyright © 2018 American Institute of Chemical Engineers (AIChE) CEP July 2018 www.aiche.org/cep 29
Reactions and Separations
30 www.aiche.org/cep July 2018 CEP Copyright © 2018 American Institute of Chemical Engineers (AIChE)