Authors: Brian McQuillan, Crystal Miranda, Brandon Quigley, and John Zhang
Reactors are the central focus of many chemical plants. Many parameters must be controlled in a reactor for proper operation. Temperature is of great importance because it affects reaction rates and equilibrium relationships. A major challenge for temperature control is handling the nonlinear nature of temperature inside most reactors. Therefore, it is important to design an effective control architecture in order to ensure optimal operation of the reactor.
This article discusses the common control architectures and topologies in CSTRs. The control architectures are designed based on whether the reactor is endothermic or exothermic. The same concepts introduced in the wiki can be applied to other reactors as well. However, only CSTRs will be discussed for simplicity.
8.2 Common Topologies
Here we will introduce a few of the most common control topologies which will be examined in applications to endothermic and exothermic CSTRs below. Please refer to the specific wiki pages for each type for detail.
8.2.1 Feedback and Feed-Forward
Feedback and feed-forward refer to the direction in which the sensor information is transfered to an actuator valve. Feedback control dictates that sensor information is "fed back" to a previous part of the process. For example, the reading from a level sensor of a filling tank can be "fed back" to the valve controlling the input to the tank. Feed-forward control means the sensor information is used to control something downstream from where the reading was taken. For example, the measured flow rate of water going into an evaporator can be used to control the heating coil inside of the evaporator. For detail about these two control topologies, please see the respective pages: Feedback Control Feed-Forward Control.
8.2.2 Ratio Control
Ratio Control is used when the ratio between two measured process variables has an optimal value. In the context of two input streams with optimal flow ratio going into a reactor, one stream is designated as the control stream, and one stream is designated as the wild stream. The wild stream fluctuates, and a valve on the control stream is opened or closed to maintain the ratio between the two stream flows. For detail about ratio control, see Ratio Control.
8.2.3 Cascade Control
Cascade control simply means that instead of one control loop found in simple control topologies where a sensor's measurement directly controls an actuator valve, multiple loops are used so that sensors' measurements can control set points for other controllers. For example, the temperature sensor measurement of process fluid exiting a reactor can be used to modify the set point of the flow controller of steam feeding the heating jacket, which then sets the steam valve. This multiple-loop system eliminates some problems caused by variable-pressure feeds, for instance. For detail about cascade control, see Cascade Control.
8.3 Disturbances to CSTRs
There a few very common disturbances that CSTR may be subjected to. When designing a control architecture for a CSTR, you must invertigate the possibility of all of these disturbances, determine the magnitude of each possible disturbance, and address how each will be handled.
- Changes in feed properties
- flow rate
- Changes in enthalpy of heat exchange medium
- Change in heat transfer properties (ex: fouling)
8.4 Disturbances to PFRs
Plug Flow Reactors (PFRs) behave differently than CSTRs and will have different properties to consider when designing a control architecture for them. The biggest difference is that the temperature, flows, and compositions will all vary along the reactor. There is danger of exceeding a design limitation in temperature or flow in certain parts of the PFR, so control of the reactor is important along its length. To design a control architecture for a PFR, the following disturbances and changes to the system must be addressed:
- Temperature control in multiple places along reactor (hot spots can easily occur in PFRs)
- Flow control in multiple places along the reactor
- Inlet/outlet pressures
- Feed property disturbances:
- flow rate
- Changes in enthalpy of heat exchange medium
- Change in heat transfer properties (ex: fouling)
8.5 Endothermic Reactors
DISCLAIMER: All reactors shown posess two controllers. We realize that in the real-world a reactor would require more than two temperature controllers but we are concerned primarily with the common placement of controllers outside of the reactor in this wiki. Thank you and continue reading.
Endothermic reactors tend to be easier to control than exothermic ones because they are much less prone to runaway. There are two commonly used methods to control an endothermic CSTR. These two methods are differentiated by the variables they manipulate. In the first method, the steam pressure in the reactor jacket is in the manipulated variable, whereas in the second method, the steam flowrate is the manipulated variable. In endothermic reactors, as well as exothermic ones as we will later see, feedback is the most common type of control system used. This is done to ensure that changes are being based on what is actually happening in the reactor and not what is predicted to happen. This is important in the endothermic case to ensure that the proper amount of heat is being added to the reactor to obtain a desired conversion. Feedback is more useful for control since the amount of heat needed can change quickly based on the amount or concentration of reactants being added. Generally the temperature control of the reactor is independent of the reactant feed, so the system needs a way to adjust to changes within reactant feed, and this is accomplished by using feedback control in the streams that control the temperature (i.e. the steam feed stream).
8.5.1 Controlled by Steam Pressure
Heat duty can be more effectively removed with the use of steam pressure as the manipulated variable because changes in the heat duty of the reactor will cause an immediate change in the steam pressure. By using the steam pressure, it also linearizes the temperature control system which is not the case when the steam flowrate is the manipulated variable. Using the steam pressure as the manipulated variable does not provide a direct measurement of the heat load, the amount of heat needed by the reactor. If a direct measurement is required, a steam flow indicator can be installed which would allow for the direct measurement of the applied heat load. An added benefit, which those of you who have worked with the temperature controller in ChE 460 might have noticed, is that often times the pressure of steam being provided by the utilities plant is variable. By using the steam pressure as the manipulated variable the control system will automatically adjust to this type of disturbance from the source. If a large change in steam pressure were to occur, the pressure controller would measure this change, and the control scheme would intake the change in steam pressure and translate this into an appropriate change in the valve setting.
Image Modified From: Riggs, James B., Karim, Nazmul M.. Chemical and Bio-Process Control. Third Edition Chapter 18. Ferret Publishing.
As can be seen in the figure above, when steam pressure is used as the manipulated variable the control system is run in a feedback mode. One of two things can control the amount of steam fed into the jacket. The first is the temperature of the product, and the second is the steam pressure in the jacket. As mentioned above, a change in the heat duty required by the reactor will quickly change the steam pressure which is why it is commonly used over the steam flowrate.
Notice that there are two controllers that are responsible for adjusting the steam valve. The pressure controller is sensitive to changes in the heat duty required by the reactor and is used to adjust the steam according to the needs of the temperature controller on the product stream. This is a prime example of cascade control. The temperature of the product stream would output a setpoint to the pressure controller for the amount of steam needed to attain the desired temperature setpoint. The pressure controller would then communicate to the valve what needs to be done in order to achieve this temperature setpoint based on the steam pressure.
8.5.2 Controlled by Steam Flowrate
The second method for controlling a endothermic CSTR is by manipulating the steam flow rate. Using flow rate as the manipulated variable makes the control system prone to changes in heat load and changes in the enthalpy of supplied steam. These changes require direct action of the temperature control system. One advantage of using flowrate as the manipulated variable is that heat load is directly measured, and thus, conversion is directly measured.
As can be seen in the figure above the system works in a feedback mode when the steam flowrate is used as the manipulated variable. The temperature of the product stream is the primary factor in adjusting the amount of steam fed into the system, and therefore this system is less responsive to changes in the amount of heat duty required for the process.
Note here that there are again two controllers used to adjust the steam valve. This setup is similar to the previous case, but now the flow controller is the "slave" controller to the temperature controller. The cascade control scheme is again at work!
8.6 Exothermic Reactors
Exothermic reactors are harder to control because safety is dependent on heat removal. There are two common control architectures for an exothermic CSTR based on the temperature of the coolant. The first method uses the outlet temperature of the coolant as the manipulated variable, while the second method uses the inlet temperature of the coolant. Just like the endothermic reactors, feedback control is very commonly used when controlling the temperature of exothermic reactors. The reasoning for feedback control in the exothermic case is slightly different but similar to that of endothermic reactors. In endothermic reactors we wanted to ensure that the proper amount of heat was being supplied to our reactor, and if the amount needed changed; we needed a way to adjust our temperature accordingly. In the exothermic case we need to ensure that we are removing the proper amount of heat, not only to ensure an optimal reaction temperature, but also to prevent a runaway reaction. Therefore, we need to know what is actually going on within the reactor instead of forecasting possible temperature changes within it. Cascade control is also used. In this case the temperature of the product stream outputs a setpoint to the temperature controller whether it be located on the outlet or inlet coolant feed. The "slave" controller can then adjust the valve position based on the temperature of the coolant water.
8.6.1 Controlled by Outlet Coolant Temperature
The figure above shows the control architecture for CSTR reactor temperature controlled by outlet coolant temperature in a feedback mechanism. There is a temperature control on both the product stream and outlet coolant temperature. Both of the temperature controls are used to control the valve on the inlet coolant stream. The advantage of this setup is that it responds faster to fouling on heat-transfer surfaces than the setup with the temperature controller on the inlet. This is because in order for the inlet temperature to adjust to fouling, the fouling must first affect the temperature of the product stream. Therefore the fouling has less direct effect on the controller, but in this case fouling will immediately affect the temperature of the exiting cooling water. However, this setup has the disadvantage of responding slower to changes in the inlet coolant temperature.
As can be seen in the above diagram you will notice that again cascade control is utilized in this particular system. The temperature sensor on the product stream provides information on whether the stream needs to be cooled more or less and outputs a setpoint to the temperature controller on the recycle stream. This controller can then take into account the temperature of the recycled coolant water and make an adjustment to the amount of fresh coolant water that is added to the system.
8.6.2 Controlled by Inlet Coolant Temperature
As seen in the figure above, the control configuration for CSTR reactor temperature is controlled by inlet coolant temperature in a feedforward mechansim. There are temperature controls on both the product stream and inlet coolant temperature; both of the controls are used to control the valve on the inlet coolant stream. One advantage of this setup is that it responds faster to changes in inlet coolant temperature. A disadvantage is that it responds slower to fouling on heat-transfer surfaces.
8.6.3 More on Exothermic Reactors
How do you decide which exothermic reactor control architecture to use? It depends on whether a faster response to fouling on heat-transfer surfaces or a faster response to changes in inlet coolant temperature is more important.
In addition to stabilizing the temperature within an exothermic reactor, it is often a goal to maximize production rate. This can be achieved by placing a valve on the feed stream. As the temperature control adjusts the cold water valve, the cold water valve position is then communicated to the feed valve and adjusts it.
Cascade control is again utilized. In this case the secondary or "slave" controller is located in a different location, which is the only difference between the two scenarios.
8.7 Worked out Example 1
As part of a community outreach program, your company is sponsoring a haunted house. The haunted house is being constructed on site in an old, dilapidated warehouse located on the plant property. You have been put in charge of designing a scene in which a witch is creating a witch’s brew. It has been proposed that you use a reaction between “Witch’s Brew Deluxe”, a commercial compound used in the haunted house industry, dry ice, and water.
Dry Ice + “Witch’s Brew Deluxe” + Water --> Bubbles + Mist
The manufacturer of “Witch’s Brew Deluxe” supplies optimal conditions for real life cauldron imitation at 30ºC water temperature. Temperatures in excess of this result in a surplus of cauldron bubbles and mist which impair the visual experience. Temperatures below this do not allow for large enough reaction rates and do not produce any steam or bubbles.
The reaction is highly endothermic, so a supply of steam has been diverted from the main plant to supply heat to the reactor. However, due to the great distance this steam is traveling, the supply steam pressure is highly variable. An uncovered, jacketed CSTR is available for use as a cauldron. A mixture of water and “Witch’s Brew Deluxe” is constantly fed into the reactor, and solid dry ice is manually feed by the operator.
What control scheme is most suitable for temperature control and why? Specifically, what should you use as the manipulated variable?
Because the steam supply pressure will vary greatly, a steam pressure control architecture is preferred. By monitoring the jacket pressure, changes in steam supply pressure will immediately be observed and corrected for.
8.8 Worked out Example 2
You have just recently taken a job offer from a haunted engineering firm run by a group of witches located in Salem, Massachusetts. As the new guy on the job they would like your input on which variable they should treat as the manipulated variable for the reactor temperature controller. Their reactor is an exothermic CSTR. The following reaction is what takes place within this CSTR.
3 frog's legs + 2 cow hooves --> 4 moles of magic elixir
The magic elixir is generally used as a wart remover but also shows promise as an antacid.
The head witch at the plant tells you that significant fouling occurs on the heat transfer surfaces because the city only allows them access to dirty water with high mineral content because they are witches and generally considered substandard citizens. She (the head witch) also points out that they have recently developed a spell to maintain the inlet temperature and pressure of their coolant water.
Which variable should be the manipulated one and why?'
The correct manipulated variable for the reactor temperature controller would be the outlet coolant temperature. This allows for a quicker response to fouled heat-transfer surfaces. Since the head witch told you that inlet temperature and pressure are constant, the configuration's weakness to changes in these two parameters can be overlooked.
8.9 Multiple Choice Question 1
Which type of REACTION is generally much easier to control?
b) Distillation of alcohol
c) Cold Fusion
8.10 Multiple Choice Question 2
By using the __________ as the manipulated variable, it linearizes the temperature control system which is not the case when the steam flowrate is the manipulated variable.
a) feed flow rate
b) steam pressure
c) product composition
d) rate of mixing
- Riggs, James B. and Karim, M. Nazmul. Chemical and Bio-Process Control. Chapter 18: Control Case Studies. Third Edition. Ferret Publishing.
- Svrcek, William Y., Mahoney, Donald P., and Young, Brent R. A Real-Time Approach to Process Control. Chapter 7: Common Control Loops. Second Edition. John Wiley and Sons, Ltd.