Control of Two-Column System

Two temperatures are controlled in the structure used for this two-column system. There is no direct measurement of composition. If the recovery column does its job and prevents B from escaping out the bottom of the column, the flowrate of distillate recycle stream D2 is an indication of the amount of excess B that is leaving the reactive column and entering the separation column. 

The total flow to the top of reactive zone FB,tot is the sum of the fresh feed of B (Fob) and distillate Da- This total stream is flow controlled by manipulating the 

Pressures in both columns are controlled by condenser duties (not shown in Fig. 11.29). 

Reflux-drum levels are controlled by reflux flowrates. 

Reflux ratios are maintained by measuring distillate flowrates and adjusting reflux flowrates. 

---

In Parts I and II we explored the steady-state designs of several ideal hypothetical systems. The following three chapters examine the control of these systems. Chapter 10 considers the four-component quaternary system with the reaction A + B C + D under conditions of neat operation. Chapter 11 looks at control of two-column flowsheets when an excess of one of the reactants is used. Chapter 12 studies the ternary system A + B C, with and without inerts, and the ternary system A B + C. We will illustrate that the chemistry and resulting process structure have important effects on the control structure needed for effective control of reactive distillation columns. 

In this section we use an extractive distillation column as an example to demonstrate how to build a steady-state simulation. This is one column of an overall two-column system for separating isopropanol and water. The detailed design and control of the overall distUlation system will be given in Chapter 10. 

Non-ideal systems that can result in three-phase distillation are discussed in Section 10.3. This phenomenon, where the liqnid splits into two phases, could occur on a number of trays or in the entire colnmn. From a practical standpoint and for better control of the column operation, one of the liquid phases is usually withdrawn when it forms. 

An alternative to operating neat is to operate the reactive column with an excess of one of the reactants. This eliminates the need to perfectly balance the reactant feeds in the reactive column itself, which makes the control of the reactive column easier. However, this mode of operation may have the disadvantage of requiring the recovery and recycle of the reactant that is in excess. The flowsheet typically consists of a two-column system a reactive column and a recovery column. 

In the two-column process, the recovery column acts concepmally as an on-line analyzer a higher recycle flowrate means that more of the excess reactant is leaving the reactive column, so the fresh feed flowrate of that reactant must be decreased. An effective control structure for the two-column system is to flow control the sum of the recycle stream and the appropriate fresh feedstream. When the recycle flowrate increases, the fresh feed flowrate is decreased to keep the total constant. Thus, the scheme changes the fresh feed flowrate to accommodate changes in the component inventory of the reactant. From a steady-state economic perspective, the two alternative processes (one column and two columns) have different capital investments and different operating costs. From a dynamic perspective, the two processes show different dynamic behavior and require different control structures. The economic design differences are quantitatively explored in this chapter. The control of these types of systems is discussed in Chapter 11. 

The control of these two alternative flowsheets is studied in Ghapter 11. The question to be answered is will the improvement in control justify the increase in costs when using a two-column system ... 

However, the flowsheets for the first and second types of chemistry come in two flavors. The flowsheet can consist of either a single reactive column or a two-column system with a reactive distillation column followed by a recovery column and recycle of excess reactant back to the reactive column. The type of flowsheet depends on whether we want to operate the reactive distillation column neat (i.e., no excess of either reactant). The excess reactant flowsheet has higher capital and operating costs, but its control is easier. The control of this system is considered in Chapter 11. 

The two-column system has two pressures and four levels to control. This leaves 12 -6 = 6 control degrees of freedom remaining. Throughput is set by 1 control degree of freedom somewhere in the system. This leaves 5 optional degrees of freedom in the two-column system. In the control structure discussed in this chapter for the two-column system, these 5 optional degrees of freedom are consumed by one total feed, two reflux ratios, and two temperatures (one in each column). 

Now that we have demonstrated that the one-column neat reactive distillation system can be controlled but may require a composition measurement to handle all types of disturbances, we want to see how the two-column system with an excess of reactant handles disturbances. The case considered is the 20% excess of B fed to the reactive coliunn. Figure 11.25 contains the flowsheet of the reactive column/iecovery column system. The two product streams are distillate Di from the reactive column containing product C and bottoms B2 from the recovery column containing product D. The distillate D3 from recovery... 

Figures 11.32-11.38 give the response of the two-column system to the same sequence of feed composition dismrbances applied in the one-column cases. In Figure 11.32 some B impurity is added to fresh feed Fqa (zoa(A) = 0.95, Zoa(B) = 0.05). The larger amount of B coming into the reactive column produces an increase in the bottoms Bi fed to the recovery column. The vapor boilup in recovery column Vs2 increases to twice its normal steady-state value, at which point the control valve is wide open. Control of tray 18 in the recovery column is lost. Despite hitting this constraint, product purities are not adversely affected. In Figure 11.33 the maximum value of Vs2 is increased to 4 times the steady-state value, which removes the constraint. Now both temperatures are controlled, and the B2 product stream becomes very pure in D.

The one-column and two-column systems are both controllable using several types of control structures. The two-temperature control scheme for the neat reactive column handles most disturbances, but it cannot handle one type of feed composition disturbance. The use of an internal composition measurement provides more robust control. 

The two-column system provides stable operation but does not maintain tight control of product composition for some feed composition disturbances. A cascade composition/ temperature structure may be required. 

For the two-temperature control structures studied thus far, one of the fresh feeds is fixed, which provides a direct handle on the production rate. The ratio to the second fresh feed (feed ratio FR) is adjusted to control a tray temperature and subsequently the stoichiometric balance. A second tray temperature is controlled by manipulating the reboiler heat duty. This is called the Q-FR control structure (Fig. 13.8a). Roat et al. ° propose a different two-temperature control stmcture for the control of the MeAc system. Two tray temperatures in the column are maintained using two fresh feedstreams. Production rate changes are achieved by adjusting the vapor boilup (Fig. 13.8c). We call this an F-F control structure, which has been studied by Kaymak and Luyben. A third control stmcture is similar to the F-F control stmcture but, instead of manipulating both feedstreams, one fresh feedstream is used to maintain a tray temperamre and the ratio to the second fresh feed is adjusted to control a second tray temperamre, which we call the F-FR control stmcture shown in Figure 13.8b. The purpose here is to compare the effectiveness of these three alternative control stmctures. A comparison will be made for dynamic responses for production rate changes for the MeAc, BuAc, and AmAc systems. 

Despite clear economic incentives for reactive distillation systems, there are relatively few articles that study the dynamics and control of reactive distillation columns. Al-Arfaj and Luyben give a review of the literature dealing with the closed-loop control of reactive distillation systems. Several control structures for an ideal two-product reactive distillation system and real chemical systems " have been proposed. One important principle in the control of reactive distillation is that we need to control one intermediate composition (or tray temperature) in order to maintain the stoichiometric balance between the two reactant components. ... 

By a combination of the two control handles, one can affect component specification in the product streams, but there is always some maximum possible extent of separation (value of Fenske ratio) in a given system. Nearly all control measures are designed to permit control of the two handles. It should be realized that column operation needs to be kept reasonably smooth, otherwise separation already achieved may in the following five minutes (or six hours) partially be undone by surging in the system. 

The reaction takes place at low temperature (40-60 °C), without any solvent, in two (or more, up to four) well-mixed reactors in series. The pressure is sufficient to maintain the reactants in the liquid phase (no gas phase). Mixing and heat removal are ensured by an external circulation loop. The two components of the catalytic system are injected separately into this reaction loop with precise flow control. The residence time could be between 5 and 10 hours. At the output of the reaction section, the effluent containing the catalyst is chemically neutralized and the catalyst residue is separated from the products by aqueous washing. The catalyst components are not recycled. Unconverted olefin and inert hydrocarbons are separated from the octenes by distillation columns. The catalytic system is sensitive to impurities that can coordinate strongly to the nickel metal center or can react with the alkylaluminium derivative (polyunsaturated hydrocarbons and polar compounds such as water). 

---