Heat transfer, reactors jacketed vessels

We assume that the reactor is constructed from fairly exotic material and has heat transfer equipment (jacket or coil) and an agitator. So the capital cost is estimated at 10 times the basic vessel cost. The total capital costs of the reactors in the three different flowsheets are 7,429,000, 4,043,000, and 3,657,000 for the 1-CSTR, 2-CSTR, and 3-CSTR processes, respectively. You can see that the reduction in cost between the 2-CSTR and the 3-CSTR processes is quite small. The cost of a 4-CSTR process could be somewhat higher because of having more vessels, even if each is somewhat smaller. 

In many industrially important situations, it is impossible to maintain geometric, mechanical (kinematic/hydrodynamic and turbulence similarities), and thermal similarities simultaneously. Consider a stirred tank reactor with heat exchange only through a jacket on its external surface. The jacket heat transfer area to vessel volume ratio is proportional to (l/T). Evidently, with scale-up, this ratio decreases, and it is difficult to maintain the same heat transfer area per unit volume as in the small-scale unit. Additional heat transfer area is required to cater to the extra heat load resulting from increase in reactor volume. This area can be provided in the form of a coil inside the reactor or an external heat exchanger circuit. In both cases, the flow patterns are significantly different than the model contactor used in bench-scale studies and kinematic similarity is violated. This is the classic dilemma of a chemical engineer it is impossible to preserve the different types of similarities simultaneously. 

For heat transfer fluids inside reactor jackets or other process vessels with agitation to fluids in vessels (Figure 10-93A), the heat transfer is expressed as... 

A useful application is for tank and vessel heating, with the heater protruding into the vessel. Bayonet heat exchangers are used in place of reactor jackets when the vessel is large and the heat transfer of a large mass of fluid through the wall would be difficult or slow, because the bayonet can have considerably more surface area than the vessel wall for transfer. Table 10-43 compares bayonet, U-tube, and fixed-tubesheet exchangers. ... 

The temperature of the reactor could theoretically be controlled by changing the flow rate or the temperature of the water in the jacket. It will now be shown that the former is impractical. The over-all heat transfer coefficient is given in the major equipment section as around 50 BTU/hr ft2°F or greater. This means that the major resistance to heat transfer is the film on the inside of the reaction vessel. 

The factors that can affect the rate of heat transfer within a reactor are the speed and type of agitation, the type of heat transfer surface (coil or jacket), the nature of the reaction fluids (Newtonian or non-Newtonian), and the geometry of the vessel. Baffles are essential in agitated batch or semi-batch reactors to increase turbulence which affects the heat transfer rate as well as the reaction rates. For Reynolds numbers less than 1000, the presence of baffles may increase the heat transfer rate up to 35% [180]. 

Three different principles govern the design of bench-scale calorimetric units heat flow, heat balance, and power consumption. The RC1 [184], for example, is based on the heat-flow principle, by measuring the temperature difference between the reaction mixture and the heat transfer fluid in the reactor jacket. In order to determine the heat release rate, the heat transfer coefficient and area must be known. The Contalab [185], as originally marketed by Contraves, is based on the heat balance principle, by measuring the difference between the temperature of the heat transfer fluid at the jacket inlet and the outlet. Knowledge of the characteristics of the heat transfer fluid, such as mass flow rates and the specific heat, is required. ThermoMetric instruments, such as the CPA [188], are designed on the power compensation principle (i.e., the supply or removal of heat to or from the reactor vessel to maintain reactor contents at a prescribed temperature is measured). 

Figure 4 shows a semi-batch reactor in which a heat transfer fluid circulates through a jacket surrounding the vessel to assist in controlling the temperature. 

A 2.5 m3 stainless steel stirred tank reactor is to be used for a reaction with a batch volume of 2 m3 performed at 65 °C. The heat transfer coefficient of the reaction mass is determined in a reaction calorimeter by the Wilson plot as y = 1600Wnr2KA The reactor is equipped with an anchor stirrer operated at 45 rpm. Water, used as a coolant, enters the jacket at 13 °C. With a contents volume of 2 m3, the heat exchange area is 4.6 m2. The internal diameter of the reactor is 1.6 m. The stirrer diameter is 1.53 m. A cooling experiment was carried out in the temperature range around 70 °C, with the vessel containing 2000 kg water. The results are represented in Figure 9.16. 

Agitated jacketed vessel the main resistance to heat transfer is located at the wall, where there is practically no resistance to heat transfer inside the reaction mass. Due to agitation, there is no temperature gradient in the reactor contents. Only the film near the wall presents a resistance. The same happens outside the reactor in its jacket, where the external film presents a resistance. The wall itself also presents a resistance. In summary, the resistance against heat transfer is located at the wall. 

The reactor system used for these experiments is a 190 liter, jacketed/ stainless steel vessel equipped with initiator and emulsifier metering system. The reactor is monitored and controlled by a minicomputer. The computer monitors the reactor temperature and pressure/ the jacket water inlet and outlet temperatures and flow rate, and the initiator and emulsifier flow rates. The computer calculates the amount of heat transferred through the jacket from the process measurements and transmits signals to control the reactor temperature and metering pumps. 

The use of a jacket surrounding the reactor vessel is probably the most common method for providing heat transfer because it is relatively inexpensive in terms of equipment capital cost (see Fig. 1.10a). If heating is required, steam is condensed in the jacket or a hot heat transfer fluid stream is fed to the jacket. If cooling is required, a cooling medium is fed to the jacket. For moderate reactor temperatures (between 50 and 80°C), cooling water at 30°C is typically used. For lower temperature reactors, a cold refrigeration stream (brine) is used. 

Figure 2.11 gives a Matlab program that performs these sizing calculations. Results for the base case feedrate, a 50% conversion, and a 320 K reactor temperature yield a coolant exit temperature of 314 K and a log-mean temperature difference of 13.5 K. The heat transfer area of the coil is 81.44 m2, compared to a jacket area of 63.4 m2. The vessel... 

Before we leave this example, let us take a look at the issue of heat transfer. In setting up the simulation, we have specified the reactor temperature (430 K) and volume (100 m3) but have said nothing about how the heat of reaction is removed. The simulation calculates a heat removal rate of 12.46 x 106 W. If the aspect ratio of the vessel is 2, a 100-m3 vessel is 4 m in diameter and 8 m in length, giving a jacket heat transfer area of 100.5 m2. If we select a reasonable 30 K differential temperature between the reactor and the coolant in the jacket, the jacket temperature would be 400 K. Selecting a typical overall heat transfer coefficient of 851 W K-1 m-2 gives a required heat transfer area of 488 m2, which is almost 5 times the available jacket area. Aspen Plus does not consider the issue of area. It simply calculates the required heat transfer rate. 

We have frequently emphasized in this book that the key issue in reactor control is heat transfer area. All cases considered so far in this chapter have assumed jacket cooling, so heat transfer area is limited by the geometry of the vessel. As discussed in Chapter 2, using a cooling coil provides 25% more area, so some improvement in control is expected when a coil is used. 

The ethylbenzene CSTR considered in Chapter 2 (Section 2.8) is used in this section as an example to illustrate how dynamic controllability can be studied using Aspen Dynamics. In the numerical example the 100-m3 reactor operates at 430 K with two feedstreams 0.2 kmol/s of ethylene and 0.4 kmol/s of benzene. The vessel is jacket-cooled with a jacket heat transfer area of 100.5 m2 and a heat transfer rate of 13.46 x 106 W. As we will see in the discussion below, the steady-state simulator Aspen Plus does not consider heat transfer area or heat transfer coefficients, but simply calculates a required UA given the type of heat removal specified. 

To illustrate some of the design and control issues, a vessel size (DR = 2 m, VR = 12.57 m3, jacket heat transfer area Aj = 25.13 m2) and a maximum reactor temperature (7j) ax = 340 K) are selected. The vessel is initially heated with a hot fluid until the reaction begins to generate heat. Then a cold fluid is used. A split-range-heating/ cooling system is used that adds hot or cold water to a circulating-water system, which is assumed to be perfectly mixed at temperature Tj. The setpoint of a reactor temperature controller is ramped up from 300 K to the maximum temperature over some time period. 

Temperature control for laboratory reactors is typically easy because of high heat transfer area-reactor volume ratios, which do not require large driving forces (temperature differences) for heat transfer from the reactor to the jacket. Pilot- and full-scale reactors, however, often have a limited heat transfer capability. A process development engineer will usually have a choice of reactors when moving from the laboratory to the pilot plant. Kinetic and heat of reaction parameters obtained from the laboratory reactor, in conjunction with information on the heat transfer characteristics of each pilot plant vessel, can be used to select the proper pilot plant reactor. 

Similarly, when moving from the pilot plant to manufacturing, a process engineer will either choose an existing vessel or specify the design criteria for a new reactor. A necessary condition for operation with a specified reactor temperature profile is that the required jacket temperature is feasible. We have therefore chosen to focus on heat transfer-related issues in scale-up. Clearly there are other scale-up issues, such as mixing sensitive reactions. See Paul [1] for several examples of mixing scale-up in the pharmaceutical industry. 

Based on initial heat flow calorimetry studies, a process development engineer must choose the appropriate reactor vessels for pilot plant studies. A pilot plant typically has vessels that range from 80 to 5000 L, some constructed of alloy and others that are glass lined. In addition some vessels may have half-pipe coils for heat transfer, while others have jackets with agitation nozzles. A process drawing for a typical glass-lined vessel is shown in Figure 4. In Sections 3.1.4.1 and 3.1.4.2 we review fundamental heat transfer relationships in order to predict overall heat transfer coefficients. In Section 3.1.4.3 we review experimental techniques to estimate heat transfer coefficients in process vessels. 

The major problem in temperature control in bulk and solution batch chain-growth reactions is the large increase in viscosity of the reaction medium with conversion. The viscosity of styrene mixtures at I50°C will have increased about 1000-fold, for example, when 40 wt % of the monomer has polymerized. The heat transfer to a jacket in a vessel varies approximately inversely with the one-third power of the viscosity. (The exact dependence depends also on the nature of the agitator and the speed of fluid flow.) This suggests that the heat transfer efficiency in a jacketed batch reactor can be expected to decrease by about 40% for every 10% increase in polystyrene conversion between 0 and 40%. 

All reactors are jacketed to permit heat removal through the vessel walls. It is frequently necessary to add extra heat removal means as the reaction vessels are scaled up because the heat transfer area of the reactor walls increases with reactor volume to the two-thirds power while the rate of heat generation is proportional to the volume itself. 

---