High Purity Columns

Distillation columns that produce products with parts-per-million (ppm) levels of impurities offer some challenging control problems. These columns exhibit very nonlinear responses to changes in manipulated variables and disturbances. The nonlinear effects take two forms nonlinear steady-state gains and nonlinear dynamics. 

Changing an input variable in one direction can produce small changes in product composition, while changing it in the other direction can produce very large changes in product composition. This asymmetric behavior of the steady-state gains can result in sluggish control in one direction and oscillatory control in the other direction. 

Since high-purity products are usually produced only in situations where the separation is relatively easy, most of these columns have fairly large temperature gradients. Therefore it is possible to use tem-perature/composition cascade control systems. The secondary temperature controller serves as a fast loop to detect disturbances quickly and hold the temperature profile in the column. The primary composition 

Dual composition control is not recommended for high-purity columns. It is better to select one end or the other for control and provide sufficient reflux to handle the worst-case conditions such that the purity of the uncontrolled product is always at or above specification. 

---

This can be used in the MESH equations to account for stage nonideality, This vaporization efficiency is applied to the equilibrium constant Ktj and appears as the product E Ky. The vaporization efficiency does solve a computational problem in placing an efficiency in the MESH equations. As shown by Lockett (105), a major disadvantage of the vaporization efficiency is that it does vary with composition. Near the top of a high-purity column, as yLj + x and x(j approach unity, Ejj also approaches unity, and so a vaporization efficiency does not ti uly reflect stage nonidealities. 

The inside-out methods (Sec. 4.2,10) can be used for most columns. The Russell method is simple to implement and does work well for a wide variety of refinery and hydrocarbon columns. The Boston method also works well for a wide range of columns and has been shown to work for superfractionators or tall, high-purity columns. Since the outer loops of the two methods are similar, they can be combined with a choice between the two methods, depending on the type of column, as part of the inside loop. 

Two-point temperature (or composition) control of single, or coupled, very high purity columns without significant time delays. 

The material balance scheme is often favoured on high purity columns. We have seen that the permitted range of cut is much smaller under these circumstances. A scheme that keeps tight control of cut should therefore perform better. 

In theory we should control the temperature at the point at which the product is withdrawn from the column. However there are several reasons why this may not be practical. Firstly, the liquid may not be homogeneously at its bubble point. This is likely to be the situation close to the top tray since reflux is often subcooled. It can also occur at the base of the column if the vapour from the reboiler is superheated. Secondly, in pseudo-binary columns, the relationship between composition and bubble point also depends on the proportion of non-key components in the product The greatest proportion of LLK will occur at the top of the column, and of HHK at the bottom. Thus, if non-key composition varies, these are the regions most prone to inaccuracy. Finally, particularly on high purity columns, the temperature may not be sensitive to changes in composition. 

On high purity columns it is usually better to develop regressed inferentials for log HKj) and log LKb). This helps accommodate the nonlinear relationships common on these columns but also has the advantage that the predicted values for HK and LKb will never be negative. 

Theoretically, the temperature at the end of the column should be controlled in a binary constant-pressure system to maintain constant produa composition. However, the temperature changes at the ends of the column are quite small in moderate- to high-purity columns. Therefore, small changes in pressure or the presence of other lighter or heavier components can aflfect temperature much more than composition of the key component. 

Suppose that the distillation column shown in Fig. 13.2 has been designed to separate a methanol-water mixture that is 50% methanol (MeOH). This high-purity column has a large number of trays and a nominal distillate composition of xd = VV of MeOH. Because a composition analyzer is not available, it is proposed to control xj) indirectly, by measuring and controlling the liquid temperature at one of the following locations ... 

As shown in Fig. 10.6, the vapor from the reactor flows into the bottom of a distillation column, and high-purity dichloroethane is withdrawn as a sidestream several trays from the column top. The design shown in Fig. 10.6 is elegant in that the heat of reaction is conserved to run the separation and no washing of the reactor... 

The Phillips process is a two-stage crystallisation process that uses a pulsed column in the second stage to purify the crystals (79,80). In the pulsed column, countercurrent contact of the high purity PX Hquid with cold crystals results in displacement of impurities. In the first stage, a rotary filter is used. In both stages, scraped surface chillers are used. This process was commercialized in 1957, but no plants in operation as of 1996 use this technology. 

The wet ester is distilled in the dehydration column using high reflux to remove a water phase overhead. The dried bottoms are distilled in the product column to provide high purity acrylate. The bottoms from the product column are stripped to recover values and the final residue incinerated. Alternatively, the bottoms maybe recycled to the ester reactor or to the bleed stripper. 

The technology is based on the rapid equhibrium estabUshed between the reactants and products at high temperatures. The equhibrium shifts to the product side when potassium is removed continuously by distihation through a packed column. This process can produce high purity potassium metal. Appropriate adjustments of conditions give a wide range of potassium—sodium ahoys of specified compositions. 

Propane and light ends are rejected by touting a portion of the compressor discharge to the depropanizer column. The reactor effluent is treated prior to debutanization to remove residual esters by means of acid and alkaline water washes. The deisobutanizer is designed to provide a high purity isobutane stream for recycle to the reactor, a sidecut normal butane stream, and a low vapor pressure alkylate product. 

Avoid attempts to recover simultaneously both high and low boiling nodes in high purity from mixtures of >3 components, particularly in columns that reflux compositions different from the distillate composition, ie, reflux of one phase from a decanter, as such operations may be difficult to control. 

The rosin column spHt is controUed by the fatty acid content specified for rosin. This is usuaUy set at 2% fatty acids. At the high temperature near the bottom of the column and the reboUer, rosin dimerizes to some extent. By taking rosin from the column as a sidestream above the bottom, its rosin dimer content is minimized. Because of its high purity, sidestream rosin product is prone to crystallization. 

The continuous production of high purity methyl or ethyl carbonate from the alcohol and chloroformates has been patented (80). Chloroformate and alcohol are fed continuously into a Raschig ting-packed column in which a temperature gradient of 72—127°C is maintained between base and head of the column HCl is withdrawn at the head, and carbonate (99%) is withdrawn at the base. 

The carbon monoxide purity from the Cosorb process is very high because physically absorbed gases are removed from the solution prior to the low pressure stripping column. Furthermore, there is no potential for oxidation of absorbed carbon monoxide as ia the copper—Hquor process. These two factors lead to the production of very high purity carbon monoxide, 99+ %. Feed impurities exit with the hydrogen-rich tail gas therefore, the purity of this coproduct hydrogen stream depends on the impurity level ia the feed gas. 

The dephlegmator process recovers a substantially higher purity C2+ hydrocarbon product with 50—75% lower methane content than the conventional partial condensation process. The C2+ product from the cryogenic separation process can be compressed and further separated in a de-ethanizer column to provide a high purity C3+ (LPG) product and a mixed ethylene—ethane product with 10—15% methane. Additional refrigeration for the deethanization process can be provided by a package Freon, propane or propylene refrigeration system. 

Whereas there is extensive Hterature on design methods for azeotropic and extractive distillation, much less has been pubUshed on operabiUty and control. It is, however, widely recognized that azeotropic distillation columns are difficult to operate and control because these columns exhibit complex dynamic behavior and parametric sensitivity (2—11). In contrast, extractive distillations do not exhibit such complex behavior and even highly optimized columns are no more difficult to control than ordinary distillation columns producing high purity products (12). 

---