Fenske-Underwood-Gilliland Shortcut Method

In the design of industrial distillation columns, engineers face a critical trade-off: rigorous simulation is accurate but computationally expensive, especially for preliminary design and feasibility studies. The Fenske-Underwood-Gilliland (FUG) Shortcut Method provides a powerful and rapid alternative for estimating key column parameters. This method is not a substitute for detailed simulation, but it is an indispensable tool for initial sizing, screening separation feasibility, and establishing reasonable starting points for more rigorous models. It efficiently bridges the gap between simple hand calculations and complex computer-aided design.

Foundational Concepts: Keys and Separation Goals

Before applying the FUG method, you must define the separation problem in terms of key components. In multicomponent distillation, you typically specify the recovery or purity of two components that define the separation's difficulty. The light key (LK) is the heaviest component that you wish to recover primarily in the distillate. The heavy key (HK) is the lightest component that you wish to recover primarily in the bottoms. All components lighter than the LK are "light non-keys" and will overwhelmingly go to the distillate; all components heavier than the HK are "heavy non-keys" and will overwhelmingly go to the bottoms.

For example, in separating a mixture of propane (LK), butane (HK), and pentane (heavy non-key), you might specify that 95% of propane should be recovered in the distillate and 95% of butane in the bottoms. The FUG method uses these key component specifications to approximate the behavior of the entire multicomponent system.

The Fenske Equation: Minimum Theoretical Stages

The first step in the FUG method is determining the minimum number of theoretical stages ($N_{min}$) required to achieve the specified separation. This represents an ideal, but impossible, scenario: operation at total reflux, where all vapor is condensed and returned to the column (no distillate product is withdrawn). The Fenske equation is:

$$N_{min}=\frac{\log \left[{\left(\frac{x_{LK}}{x_{HK}}\right)}_D{\left(\frac{x_{HK}}{x_{LK}}\right)}_B\right]}{\log ({\alpha }_{LK/HK}{)}_{avg}}$$

Where:

• $(x_{LK}/x_{HK}{)}_D$ is the ratio of light key to heavy key mole fractions in the distillate.
• $(x_{HK}/x_{LK}{)}_B$ is the ratio of heavy key to light key mole fractions in the bottoms.
• $({\alpha }_{LK/HK}{)}_{avg}$ is the average relative volatility of the light key relative to the heavy key. A common approximation is the geometric mean of the relative volatility at the top and bottom column conditions: ${\alpha }_{avg}=\sqrt{{\alpha }_{top}\cdot {\alpha }_{bottom}}$.

This equation tells you the absolute lower limit on the number of equilibrium stages needed. If your specified separation requires an impossibly high $N_{min}$, the separation is likely too difficult for conventional distillation.

The Underwood Equations: Minimum Reflux Ratio

The second theoretical limit is the minimum reflux ratio ($R_{min}$). This is the lowest reflux rate at which the specified separation is theoretically possible, but it would require an infinite number of stages. The Underwood method involves a two-step calculation.

First, you solve the following equation for the parameter $\theta$, which lies between the relative volatilities of the light and heavy keys:

$$\sum _{i=1}^n\frac{{\alpha }_i\cdot x_{i,F}}{{\alpha }_i-\theta }=1-q$$

Here, ${\alpha }_i$ is the relative volatility of component i (typically relative to the heavy key), $x_{i,F}$ is its feed mole fraction, and $q$ is the feed condition (1 for saturated liquid, 0 for saturated vapor). This equation is solved iteratively for a root $\theta$ that satisfies ${\alpha }_{HK}<\theta <{\alpha }_{LK}$.

Second, you use this $\theta$ value to calculate the minimum reflux ratio:

$$R_{min}+1=\sum _{i=1}^n\frac{{\alpha }_i\cdot x_{i,D}}{{\alpha }_i-\theta }$$

Here, $x_{i,D}$ are the estimated distillate compositions, often based on the key component specifications. The Underwood equations account for the presence of non-key components, giving a more realistic $R_{min}$ than binary approximations.

The Gilliland Correlation: Linking Actual Design

The Fenske and Underwood equations define the two extremes of column operation: total reflux (minimum stages) and minimum reflux (infinite stages). Real columns operate between these limits. The Gilliland correlation is an empirical relationship that allows you to estimate the actual number of theoretical stages ($N$) required for a chosen actual reflux ratio ($R$).

The correlation is typically presented graphically or via associated empirical equations. It relates two dimensionless parameters:

• Abscissa: $X=\frac{R-R_{min}}{R+1}$
• Ordinate: $Y=\frac{N-N_{min}}{N+1}$

The correlation shows that as you move from minimum reflux ($X=0$) toward higher reflux, the required number of stages drops rapidly from infinity toward $N_{min}$. The most common operational reflux ratio is often in the range of $R=1.2\cdot R_{min}$ to $1.5\cdot R_{min}$. You select an $R$, calculate $X$, use the Gilliland chart or equation to find $Y$, and then solve for the actual number of stages $N$.

Worked Example Outline: For a feed of Benzene (LK), Toluene (HK), and Xylene, with specified recoveries:

1. Use recoveries and feed data to estimate distillate/bottoms compositions.
2. Calculate average ${\alpha }_{Benzene/Toluene}$ from given K-values.
3. Apply Fenske Equation to find $N_{min}$.
4. Apply Underwood Equations (with feed q=1) to find $\theta$ and then $R_{min}$.
5. Choose $R=1.3\cdot R_{min}$.
6. Use the Gilliland correlation (e.g., the Eduljee form: $Y=0.75(1-X^{0.5668})$) to find $N$.

Common Pitfalls

1. Misidentifying Key Components: The most frequent error is incorrectly assigning the light and heavy keys. The LK must be the heaviest component you want in the distillate; the HK must be the lightest component you want in the bottoms. Switching these renders all subsequent calculations meaningless for your intended separation.

1. Incorrect Average Relative Volatility: Using a simple arithmetic mean instead of the geometric mean for ${\alpha }_{avg}$ in the Fenske equation can introduce significant error, especially if the relative volatility changes considerably from the top to the bottom of the column. Always use the geometric mean: ${\alpha }_{avg}=\sqrt{{\alpha }_{top}\cdot {\alpha }_{bottom}}$.

1. Blind Application of Gilliland: The Gilliland correlation is empirical and based on data for conventional columns and relatively ideal systems. Its accuracy diminishes for highly non-ideal systems, columns with multiple feeds, or complex side streams. It gives a reasonable first estimate but should not be treated as a high-precision tool.

1. Ignoring Non-Key Components in Underwood: While the Fenske equation only needs the keys, the Underwood equations must include all components in the summation, especially those with volatilities between the keys. Omitting a significant component like an intermediate boiler will give an incorrect and non-conservative (too low) value for $R_{min}$.

Summary

• The FUG Shortcut Method is a three-step procedure for the rapid preliminary design of multicomponent distillation columns, using the light key (LK) and heavy key (HK) to define the separation.
• The Fenske Equation calculates the minimum number of stages ($N_{min}$) at total reflux, providing the lower bound on column height.
• The Underwood Equations calculate the minimum reflux ratio ($R_{min}$), providing the lower bound on energy requirement (reboiler/condenser duty).
• The Gilliland Correlation empirically relates the actual reflux ratio ($R$) and actual number of stages ($N$) to these two limits, allowing for practical design.
• This method is best used for feasibility studies, initial sizing, and generating starting points for rigorous simulation, not for final design.
• Success depends on correctly identifying key components, using proper average relative volatilities, and including all components in the Underwood calculations.