Page 338 - Chemical engineering design

313
DESIGN INFORMATION AND DATA
1. The level of design; less accuracy is obviously needed for rough scouting calcula-
tions, made to sort out possible alternative designs, than in the final stages of design;
when money will be committed to purchase equipment, and for construction.
2. The reliability of the design methods; if there is some uncertainty in the techniques
to be used, it is clearly a waste of time to search out highly accurate physical
property data that will add little or nothing to the reliability of the final design.
3. The sensitivity to the particular property: how much will a small error in the property
affect the design calculation. For example, it was shown in Chapter 4 that the
estimation of the optimum pipe diameter is insensitive to viscosity. The sensitivity
of a design method to errors in physical properties, and other data, can be checked
by repeating the calculation using slightly altered values.

It is often sufficient to estimate a value for a property (sometimes even to make an
intelligent guess) if the value has little effect on the final outcome of the design calculation.
For example, in calculating the heat load for a reboiler or vaporiser an accurate value of
the liquid specific heat is seldom needed, as the latent heat load is usually many times
the sensible heat load and a small error in the sensible heat calculation will have little
effect on the design. The designer must, however, exercise caution when deciding to use
less reliable data, and to be sure that they are sufficiently accurate for his purpose. For
example, it would be correct to use an approximate value for density when calculating
the pressure drop in a pipe system where a small error could be tolerated, considering
the other probable uncertainties in the design; but it would be quite unacceptable in the
design of a decanter, where the operation depends on small differences in density.
Consider the accuracy of the equilibrium data required to calculate the number of
equilibrium stages needed for the separation of a mixture of acetone and water by distil-
lation (see Chapter 11, Example 11.2). Several investigators have published vapour-liquid
equilibrium data for this system: Othmer et al. (1952), York and Holmes (1942), Kojima
et al. (1968), Reinders and De Minjer (1947).
If the purity of the acetone product required is less than 95 per cent, inaccuracies in the
v l e plot will have little effect on the estimate of the number of stages required, as the
relative volatility is very high. If a high purity is wanted, say >99 per cent, then reliable
data are needed in this region as the equilibrium line approaches the operating line (a
pinch point occurs). Of the references cited, none gives values in the region above 95 per
cent, and only two give values above 90 per cent; more experimental values are needed
to design with confidence. There is a possibility that the system forms an azeotrope in
this region. An azeotrope does form at higher pressure, Othmer et al. (1952).

8.5. PREDICTION OF PHYSICAL PROPERTIES

Whenever possible, experimentally determined values of physical properties should be
used. If reliable values cannot be found in the literature and if time, or facilities, are not
available for their determination, then in order to proceed with the design the designer must
resort to estimation. Techniques are available for the prediction of most physical properties
with sufficient accuracy for use in process and equipment design. A detailed review of