This is by far the most frequent situation encountered when a column is unable to satisfy the allowable tolerance. The following section gives the most com-mon ailments and remedies.
Poor Initial Estimates
Initial estimates are important only to the extent that they provide the initial starting point for the tower algorithm. Generally, poor guesses simply cause your tower to converge more slowly. However, occasionally the effect is more serious. Consider the following:
l Check product estimates using approximate splits. A good estimate for the tower overhead flow rate is to add up all the components in your feed which are expected in the overheads, plus a small amount of your heavy key component. If the tower starts with extremely high errors, check to see that the overhead estimate is smaller than the combined feed rates.
l Poor reflux estimates usually do not cause a problem except in very nar-row boiling point separations. Better estimates are required if you have high column liquid rates relative to vapor rates, or vice versa.
l Towers containing significant amounts of inert gases (for example, H2, N2, and so forth), require better estimates of overhead rates to avoid ini-tial bubble point problems. A nitrogen rejection column is a good
example.
Note: To see the initial estimates, click the View Initial Estimates button on the Monitor page of the column property view.
Input Errors
It is good practice to check all of your input just before running your column to ensure that all your entries, such as the stage temperatures and product flow rates, appear reasonable:
l Check to ensure that your input contains the correct values and units.
Typical mistakes are entering a product flow rate in moles/hr when you really meant to enter it in barrels/day, or a heat duty in BTU/hr instead of E+06 BTU/hr.
l When specifying a distillate liquid rate, make sure you have specified the Distillate rate for the condenser, not the Reflux rate.
l If you change the number of trays in the column, make sure you have updated the feed tray locations, pressure specifications, and locations of other units such as side exchangers on the column.
l If the tower fails immediately, check to see if all of your feeds are known, if a feed was entered on a non-existent tray, or if a composition specification was mistakenly entered for a zero component.
Note: Clicking the Input Summary button on the Monitor page of the column property view displays the column input in the Trace Window.
Incorrect Configuration
For more complex tower configurations, such as crude columns, it is more important that you always review your input carefully before running the tower.
It is easy to overlook a stripping feed stream, side water draw, pump around or side exchanger. Any one of these omissions can have a drastic effect on the column performance. As a result, the problem is not immediately obvious until you have reviewed your input carefully or tried to change some of the spe-cifications.
l Check for trays which have no counter-current vapor-liquid traffic.
Examples of this are having a feed stream on a tray that is either below the top tray of an un-refluxed tower or a tower without a top lean oil feed, or placing a feed stream above the bottom stage of a tower that does not have a bottom reboiler or a stripping feed stream below it. In both cases the trays above or below the feed tray become single phase.
Since they do not represent any equilibrium mass transfer, they should be removed or the feed should be moved. The tower cannot converge
with this configuration.
l The tower fails immediately if any of the sidestrippers do not have a stripping feed stream or a reboiler. If this should occur, a message is generated stating that a reboiler or feed stream is missing in one of the sidestrippers.
l Make sure you have installed a side water draw if you have a steam-stripped hydrocarbon column with free water expected on the top stage.
l Regardless of how you have approached solving crude columns in the past, try to set up the entire crude column with your first run, including all the side strippers, side exchangers, product side draws, and pump arounds attached. Difficulties arise when you try to set up a more sim-plified tower that does not have all the auxiliary units attached to the main column, then assign product specs expected from the final con-figuration.
Impossible Specifications
Impossible specifications are normally indicated by an unchanging heat and spec error during the column iterations even though the equilibrium error is approaching zero. To get around this problem you have to either alter the column configuration or operating pressure or relax/change one of the product specifications.
l You cannot specify a temperature for the condenser if you are also using subcooling.
l If you have zero liquid flows in the top of the tower, either your top stage temperature spec is too high, your condenser duty is too low, or your reflux estimate is too low.
l If your tower shows excessively large liquid flows, either your purity specs are too tight for the given number of trays or your Cooler duties are too high.
l Dry trays almost always indicate a heat balance problem. Check your temperature and duty specifications. There are a number of possible solutions: fix tray traffic and let duty vary; increase steam rates;
decrease product makes; check feed temperature and quality; check feed location.
l A zero product rate could be the result of an incorrect product spec, too much heat in the column which eliminates internal reflux, or the absence of a heat source under a total draw tray to produce needed vapor.
Conflicting Specifications
This problem is typically the most difficult to detect and correct. Since it is rel-atively common, it deserves considerable attention.
l You cannot fix all the product flow rates on a tower.
l Avoid fixing the overhead temperature, liquid and vapor flow rates because this combination offers only a very narrow convergence
envelope.
l You cannot have subcooling with a partial condenser.
l A cut point specification is similar to a flow rate spec; you cannot specify all flows and leave one unspecified and then specify the cut point on that missing flow.
l Only two of the three optional specifications on a pump around can be fixed. For example, duty and return temperature, duty and pump around rate, and so forth.
l Fixing column internal liquid and vapor flows, as well as duties can present conflicts since they directly affect each other.
l The bottom temperature spec for a non-reboiled tower must be less than that of the bottom stage feed.
l The top temperature for a reboiled absorber must be greater than that of the top stage feed unless the feed goes through a valve.
l The overhead vapor rate for a reboiled absorber must be greater than the vapor portion of the top feed.
Heat and Spec Error Oscillates
While less common, this situation can also occur. It is often caused by poor ini-tial estimates. Check for:
l Water condensation or a situation where water alternately condenses and vaporizes.
l A combination of specifications that do not allow for a given component to exit the column, causing the component to cycle in the column.
l Extremely narrow boiling point separations can be difficult since a small step change can result in total vaporization. First, change the spe-cifications so that the products are not pure components. After con-vergence, reset the specifications and restart.