Distillation dead time

Several control techniques have been developed to compensate for large dead-times in processes and have recently been reviewed by Gopalratnam, et al. (4). Among the most effective of these techniques and the one which appears to be most readily applicable to continuous emulsion polymerization is the analytical predictor method of dead-time compensation (DTC) originally proposed by Moore ( 5). The analytical predictor has been demonstrated by Doss and Moore (6) for a stirred tank heating system and by Meyer, et al. (7) for distillation column control in the only experimental applications presently in the literature. Implementation of the analytical predictor method to monomer conversion control in a train of continuous emulsion polymerization reactors is the subject of this paper. 

The dynamic behavior of processes (pipe-vessel combinations, heat exchangers, transport pipelines, furnaces, boilers, pumps, compressors, turbines, and distillation columns) can be described using simplified models composed of process gains, dead times, and process dynamics. 

Distillation separates the components of a mixture on the basis of their boiling points and on the difference in the compositions of the liquids and their vapors. The product purity of a distillation process is maintained by the manipulation of the material and energy balances. Difficulties in maintaining that purity arise because of dead times, nonlinearities, and variable interactions. 

These lags are cumulative as the liquid passes each tray on its way down the column. Thus, a 30-tray column could be approximated by 30 first-order exponential lags in series having approximately the same time constant. The effect of increasing the number of lags in series is to increase the apparent dead time and increase the response curve slope. Thus, the liquid traffic within the distillation process is often approximated by a second-order lag plus dead time (right side of Figure 2.82). 

SIXTHLY The Heat warms the cold Earth that while cold was half-dead. Thereof says SOCRATES When Heat penetrates, it makes subtle all earthly things, that are of service to the matter, but come to no final form while it is acting on the matter. The Philosophers conclude on the mentioned Heats in brief words, saying Distill seven times and you have separated the destructible moisture and it takes place as in one distillation. 

The order of the postulated model is a very important factor. For a well-known process such as a stirred-tank heater, it is not a problem. On the other hand, it is not obvious what order of dynamics we should assume for a fluid catalytic cracker (see Example 4.15). Also, it is not obvious what type of low-order model we should use to approximate the high-order models of even simple distillation columns (see Example 4.16). As a general starting point one could employ first- or second-order models with or without dead time. There exist a surprisingly large number of processes which could be effectively described by such low-order models. 

Could you have dead time between the overhead vapor and the distillate product If yes, why ... 

Scheme 2. This scheme indirectly adjusts the material balance through the two level control loops. This arrangement has the advantage of reducing the ratio of effective dead time to total lag time within the composition loop. It has the disadvantage of allowing greater interaction between the material and energy balances because internal reflux is not held constant. This scheme should be considered when the reflux is smaller than other flows in the column and when the reflux to distillate ratio is 0.8 or less. It should also be considered for applications where reducing the ratio of effective dead time to total lag time in the composition loop is a significant and necessary consideration.

The composition of a ])roduct leaving a 50-tray distillation colunin e.xhihits a dead time of lO min following a change in reflux flow, ruder propor-tional-plus-resi t control, estimate (a) the i)eriod of oscillation, (h) the reset time, ( ) the dynamic gain of the process. 

The dynamic response of composition to a change in distillate flow exhibits considerable dead time, as is expected in a multicapacity process. But the presence of an additional feature is indicated by step-response tests. Figure 11.13 illustrates results which are typically encountered. The response is the sort which would be seen in a transmission line with... 

Nonetheless, the first dead time dominates the control loop. It may be 5 to 30 min in duration, depending principally on the distance of the loop between the distillate valve and the analyzer. (The volume of the accumulator contributes significantly to the response.) The closed loop can then be expected to oscillate at some period between 20 min and 2 hr. 

When the number of reactors, sections in a plug-flow reactor or plates of a distillation column is large, for example, becomes 20 or 30, the response approaches that of a process dead time for which the transfer function is ... 

Quality measurement is a totally difterent issue than flow, level, temperature or pressure measurement. The primary reason is that there are so maity different quality characteristics that can be measured and they are vastly different with respect to the type of measurement and measuring principles. Just to name a few examples of quality measurements viscosity, odor, color, taste, pH, octane number, particle size distribution, etc. In traditional chemical engineering one type of measurement that is often encountered is composition analysis by means of a gas chromatograph. Because these devices are expensive they are usually used to analyze multiple streams. To give an example, it may be required to analyze the feed flow composition, top and bottom composition of a distillation tower. If one analysis takes a few minutes (residence retention time), and three streams are analyzed it is obvious that composition measurement caimot be realized without a delay or dead time. This is characteristic for numerous quality measurements they often have a measurement delay associated with them, which can vary from a few minutes to tens of minutes. In control this will pose a problem and a dead time compensation technique may be required to achieve adequate control loop performance. 

If Ktr/ttr is positive and close to unity, a change in boilup causes no change in level for a period of time equal to ttr. The control system seems aflliaed with dead time, but in reality, as shown by equation (13.39), it is not. Thistlethwaite has carried out a more extensive analysis of inverse response in distillation columns. 

A final and important consideration to keep in mind is the dead time that may be present in the column. In Chapter 3, dead time was described as being generated by a series of lags (material or energy capacitances). It is easy to see bow a distillation column with its multiple stages can generate dead times. The control scheme on a distillation column should be set up to minimize tbe dead times with respect to the process lags and disturbances. 

If the feed contains multiple components, fixing the temperature and pressure of a stage in the distillation column may not fix the composition. Therefore, a steady-state model may be used to compare advantages of using an online composition analyser rather than a temperature controller. Factors to consider are yield loss, energy consumption and dead time [11]. 

A key consideration for plant operability and controllability is variable interaction. We have learned that dead time is one of our enemies, as it always makes tight control more difficult to achieve. Variable interaction places similar restrictions on the way we can control a process and can significantly reduce the overall control system performance. Three common sources of variable interaction are the nature of the process (i.e. distillation), the combination of multiple unit operations and heat integration. Each of these points can be highly advantageous in the steady state, but they can also create operability and controllability problems that may not be evident without considering the process dynamics at the design phase. 

Reconfigure the controllers on the stabilizer from Workshop 7 to provide two-point composition control. Assume that each product is equally important and that the control objectives are 0.01 per cent propane in the bottoms and 0.1 per cent isopentane in the distillate. Also assume that you have two perfect analysers (i.e. no dead time, no error) available so that the two compositions can be controlled directly. 

If you did not have perfect analysers available or did not want to introduce dead time or error that would be present with real analysers, is there a combination of easily measured temperatures that you could successfully use to infer the distillate and bottoms compositions for the whole range of feed variance given in Table W7.1 ... 

---