Posted by Dan O'Connor on June 08, 1999 at 11:35:48:
In Reply to: Re: Pressure-Compensating TCs at Regulatory Level ? posted by Jim Gunderman on June 08, 1999 at 00:19:16:
Don, Jim
Doug Robertson, Tod Peterson, and I once built a pressure compensated TIC at the regulatory level on a DIB tower (light ends unit). The reason we did it was because of the interaction between the tower pressure and the temperature. The overhead pressure was controlled by flooding the condenser, requiring a large gain for acceptable control. The TIC also required a large gain to reject disturbances (the feed was on level control from two upstream towers), and there seemed to be a fair amount of interaction. Pressure compensating the TIC significantly reduced the interaction.
One factor which might make the difference whether the regulatory level approach works might be the steepness of the composition (temperature) profile in the tower. In the DIB, the relative volatilities are pretty low, and the temperature profile is pretty flat i.e., most of the temp. change was because the pressure was changing. In a demethanizer or deethanizer in an olefins plant, the composition profile can be pretty steep. I've cut the reboiler on a demethanizer, and watched as the temperature profile in the tower dropped down through the tower. It is possible (in towers that have multiple ti's in the bottom tower section - why would anyone put extra/any ti's in there?), to watch the temperature profile shift through each tower section. When the heat is first cut, one of the upper ti's will start to drop, while the others stay essentially constant, then the next ti down will become active, while the others are changing by only small amounts, then the next ti, and so on.
I once saw the extremes of this in a methanol/water column. This tower had a top temperature, a ti in the upper tray section above the feed, a ti in the lower tray section (below the feed), and a ti in the bottom. The tower had the reflux on flow control, and the reboiler on steam control. The lower tray section ti and bottom tray ti were essentially the temperature of saturated water at the tower pressure. The upper tray section reflects a temperature of mostly water, and only the top tower ti was methanol temperature. Upon cutting a little steam, nothing happened, except the bottom level began to increase. After a little while, the upper tray temperature began to drop, and this temperature dropped to the point of the temperature of saturated methanol at the tower pressure. The level in the bottom of the tower continued to rise, and nothing else seemed to happen - until the composition profile reached the bottom tray ti. The composition profile was very steep, an a TIC with a high gain was required to catch and hold the bottom tray temperature. For this extreme case, I can't see what pressure compensating the temperature in the TIC really means - because the composition profile is so steep relative to the pressure effects, that it would probably be a waste of time. I did put a delta t in as a constraint between the bottoms temperature (basically pure water), and the tray TIC, to make certain that the TIC stayed on control. Obviously if the composition profile weren't kept in the area of temperature measurement, the TIC would not work.