PIDConnectControlActuator

Title: Choosing Signal, Control Loops, and Plantwide Control Schemes

Authors: John D'Arcy, Matthew Hagen, Adam Holewinski, and Alwin Ng

Date Presented: September 21, 2006 /Date Revised:


 * First round reviews for this page
 * Rebuttal for this page

Introduction
Many parameters in a chemical process must be continually monitored and adjusted. Their values are measured by sensors that connect to controllers, which, in turn, interpret the signals and relay the information to actuators such as valves, motors, and electric heaters. Controllers tell actuators how quickly and to what extent to bring about a change. For example, an actuator must be informed how far to open or close a valve in order to maintain a tank pressure (see Figure 1). Each control loop (sensor to contoller to actuator) must be designed with the overall process in mind. The following section contains descriptions of the types of signals that can be used to tie the components of a control loop together and then gives a basic methodology for choosing the best way to implement controls over an entire process. Processes considered are assumed to be running at steady state with moderate disturbances, because dynamic systems with large flucuations, such as process startup, are generally regulated manually until steady state is achieved.



Types of signals
Controllers can communicate with actuators through three main types of signals: electric, pneumatic, and digital.

Notation
•Pneumatic signals:

•Electric signals:

•Digital signals:

Signal Choice
Signal choice is influenced by a variety of factors including system environment, cost, and power needs. Pneumatic and electric signals are both analog (of varying amplitude) while digital signals are binary and must be interpreted by a microprocessor.

Electrical signals use inexpensive wire to communicate with the system controls. However, the electrical signal requires a high voltage in order to achieve its necessary power. Because of this, its use must be planned carefully with regard to the system environment. Potentially dangerous points of ignition can easily be created by mistake. You might choose an electrical signal with a motor. A brief interruption in the signal would not be enough to affect the motor since it has enough inertia to continue to spin.

Pneumatic signals require the use of pneumatic piping, as well as an air compressor in order to communicate with the actuator. Running this piping is much more expensive than installing the wire necessary for electrical signals, and additional equipment such as compressors add to the cost. An important consideration when choosing between pneumatic and electric is that when there is an interruption, electrical signals immediately shut off, while pneumatic signals reduce gradually. For example, if you had an alarm that would be reset by the interruption of a signal (like your alarm clock) you would much rather have a pneumatic signal, which would not relay the interruption.

Digital signals are discrete time signals, meaning the values are measured at fixed intervals and characterized by binary numbers (zeros and ones), rather than the variable intensity that characterizes analog signals. They are carried by wire, just like electrical signals, but random variation is essentially non-existent due to the on/off (one/zero) nature of the signal. One might choose a digital signal in the case of.

It is also possible to use multiple signals within the same system. Transducers (converters) take an input signal of one type and output a different type of signal. For example, a transducer may receive a pneumatic signal of a certain pressure and convert it to a proportional signal in electric current.

An important concept to consider when using analog signal is “live zero”. Actuator signals are measured in terms of percentages. Consider a pneumatic signal with measurement range from 0 – 12 psig. 0 psig corresponds to 0 percent measurement and 12 psig corresponds to 100 percent measurement. However, a 0 psig signal can also correspond to a malfunctioning system. “Live zero” scaling can be used to fix this potential problem. Instead of a 0 – 12 psig range, a 3 – 15 psig range is often used in industry, where a 3 psig signal corresponds to 0 percent measurement and 15 psig corresponds to 100 percent measurement. The potential problem is eliminated because if we receive a 0 psig signal, this corresponds to a negative percentage measurement, which clearly indicates system malfunction. Similarly for an electric signal, instead of using a signal that ranges from 0 – 16 mA, industry uses the standard range of 4 – 20 mA to transmit signals from instruments and sensors in the field to a controller.



Plantwide control rules
In order to make the practice of plantwide control and design safe, easily controlled, environmentally friendly, resistent to disturbances, and economical, rules and guidelines have evolved in industry throughout time. These guidelines can seem very rudimentary; however, they can be easily overlooked. They give insight into the methods of assigning controllers to actuators, as well as some other important considerations for process control. The guidelines are shown below.

Material Balances
•	Make sure all components that enter the system are accounted for in material balances

In chemical processes products, reactants, and inerts are all present. The inerts that enter must be separated from the products safely and in an efficient manner. This can be accomplished through distillation, or with the introduction of a flow-controlled purge stream. Excess reactants that are not stoichiometrically accounted for can gradually accumulate in the system as a result of inefficient separation. Likewise products that are not removed can accumulate, and in the case of reversible reactions they can drive the reaction backwards and decrease the yield. Using a carefully calculated material balance that accounts for all inputs and outputs is key to the successful operation of a steady state, controllable process.

Independent Adjustments
•	Material Balance control and Product Quality control should be independently adjustable

In order to adjust process output without altering product purity or vice versa, material balance and product quality must have two independent control loops. The time constants should differ by an order of magnitude. (Time constants represent the speed of a system’s response to a changing input). This enables the plant to increase or decrease product output without affecting desired purity. Conversely purity can be adjusted to a desired setting without changing output levels.

Independent adjustments can be quite complicated. Consider a situation in which you desire to double the output of a CSTR. Simply doubling the input stream flows can affect reaction yield, leaving more unreacted reactant in the effluent and decreasing purity. A control scheme must be able to automatically compensate for this effect in later separations to maintain purity. Likewise if, say residence time is increased to raise purity, the flow rates must adjust accordingly so that output levels do not suffer.

Constant Recycle
•	Every recycle loop should be set to a constant flow rate

Recycle loops reintroduce material to the system. With inconsistent material return, unpredicted outcomes can arise. When controllers raise or lower a flowrate, recycle stream rates vary accordingly, and this can create a positive feedback loop, in which an increased vessel level can cause the control system to increase exit flow, which in turn would increase the recycle rate and raise the tank level further (i.e. “the snowball effect.”) Negative feedback would drain the tank in a similar manner. Limiting the flow rate in the recycle streams obviates the problem. To maintain this constant flowrate the actuator regulating the recycle stream should operate only at one setting.

Phase of Feed
•	Feed to a distillation column should always be in the opposite phase as the desired product stream

Product purity can be better controlled in the distillate if the feed is a liquid, and in the bottoms if the feed is a vapor. This is because it is easier to utilize a phase change to leave impurities behind than to comletely extract the impurities from the desired phase. For example, if a liquid feed enters, the more volatile component will tend to separate out as a vapor. If this is the desired product, it will be fairly pure. On the contrary, if the desired product is in the bottoms, all the impurity must be removed. This may not occur before it reaches the exit.

Controlling Stream
•	In any vessel that is level controlled, the largest stream entering or exiting the unit should be flow controlled.

Dramatic changes in a vessel’s level can be implemented quickly by controlling the most substantial flow. Levels can be raised by increasing input or decreasing output from any connected stream. Kinetics and stoichiometry may influence which streams can be manipulated, but a well controlled system will adapt to a change in one flow by altering other relevant flows in response. The important thing is that this change is brought about as quickly as possible, which is most easily accomplished by altering the largest variable flow rate and allowing the others to adapt in response. The actuator controlling the largest stream should be the most accessible and easiest to vary.

Proportional Controls
•	Proportional P controllers should be used for liquid levels. Proportional-Integral PI controllers should be used for all other control loops, unless they a require very tight control and have large signal to noise ratios. In this case one may resort to a proportional-integral-derivative PID controller.

Algorithm for Plantwide Process Control


Synthesizing control rules and other process objectives can often be a daunting task. To ease the difficulties associated with this, an algorithm for implementing plant wide control has been developed. The steps are depicted in Figure 2 (left) and detailed below.

1)	Determine goals for the control system

One must obviously know the objective of a design before any design steps can be planned. Before doing any design work, clearly define all constraints and objectives. These include the desired yields of reactions, purities after separations, safety restrictions, environmental concerns, and limitations on equipment or material.

2)	Find degrees of freedom in system

The degrees of freedom represent the number of adjustable variables. This should be less than or equal to the number of control valves in the process. Each valve can be linked to a controller to manage a corresponding process parameter. Extra valves can later be used for optimization of the regulated process.

3)	Develop an energy management strategy

Heat that builds up in exothermic processes must be dissipated correctly. Whenever possible, heat should be recycled into other units that require thermal energy. If heat cannot be used in other units, then it must be removed from the process by a coolant stream or some other means to prevent it from accumulating and/or returning back to the reactor through recycle streams. Excess heat building up in an exothermic system has potential for thermal runaways. One possible way to control this is to install a temperature sensor in the recycle stream and an actuator on a cooling stream. If the temperature heading back into the reactor is too high, a valve can open to cool down the inlet stream.

4)	Find the best variable to control production rate

There are many variables inside of a reactor that can affect production rate; for example, temperature, pressure, and reactant concentrations. The best way to control the production rate is to pick the dominant variable and use actuators to set control valves to keep the variable constant. Some guidelines for picking the best dominant variable include selecting one that is: •Directly and instantaneously capable of increasing production rate •Least affected by outside disturbances •Not interfering with other design constraints •Not capable of significantly changing processes down the line

5)	Choose control valves that will most efficiently influence quality given the constraints

Once you have selected the right variable to control, it is important to select the right valve to control that variable. Some factors to consider when selecting the right valve include: small response times, dead times, and large steady state gains. For example, a plug valve can be used when controlling temperature and taking into consideration safety constraints. The feature that we are relying upon is its quick shutoff. An actuator can regulate a plug valve to immediately cut off a heat source going into an overheated reactor. A more detailed list of valves can be found on the page.

6)	Set recycle streams to constant flow rate and choose variables that control inventory

Constant recycle flow makes steady state more easily attainable and prevents drastic changes in operation which could, for example, lead to flooding or turndown in a column.

7)	Verify material balances

Tracking all components in the system and accounting for their removal is critical to avoiding accumulation and maintaining steady state.

8)	Set up control loops on each unit operation

Each unit operation should be capable of controlling its operating parameters independently of the rest of the process. Though it is important that the units function well as a whole, it is equally important that when a unit fails, the rest of the process is equipped to handle the situation. For example, if a reactor is shut down, a subsequent distillation column will respond to the lack of input by going to a fail safe mode rather than being damaged by dry operation due to its dependence on the fragile balance of every other unit working simultaneously.

9)	Use any left over degrees of freedom for optimization

In a system in which there are two independent variables controlling one desired outcome, the control system can be used to minimize the use of the more costly variable. For example, a reactor may require a high temperature, which can be achieved in two possible ways. The reaction vessel may contain a heating coil, the use of which may be costly. The second option may involve running a recycled gas stream in a heat exchanger counter currently to the feed stream, thus heating the stream up while using minimal additional energy costs. The heating coil would only be necessary in drastically upset conditions where manipulation of the actuator on the recycled gas stream alone is insufficient to achieve the desired temperature.

In addition, one should note that in any open-ended problem, there are many possible solutions. Two different engineers may follow an identical algorithm and end up with two different control schemes. Many workable configurations may exist, and no single solution will be "correct." An alternative algorithm is available in the Standard Structures & Locations page.

Example 1: Deciding between Electric, Digital, and Pneumatic signals
For the following situations you are to decide whether to use an electric, digital, or pneumatic signal for the actuator in the control system.

a.) Exhaust air is being vented from a hot kiln, the air cannot exceed 300 ° C or it might damage the equipment that is around it. For control, a fan is inserted to attempt to cool the air when it is over 300 ° C. This fan will only run at full speed or it will not run at all.

b.) A thermostat is located in a small room with very limited space. The thermostat is regulating the temperature in the room because there exists sensitive materials which need to stay within a specified temperature range

c.) For a hospital wing the oxygen supply for the patients is held in a tank under the floor. The flow out of the tank is split to the individual rooms and can be adjusted based on need. When no oxygen is needed in a room there should be no flow.

ANSWERS

a.) A digital signal should be used. This is because with a digital signal only on, and off are capable of being represented. In this case the fan is either on when the temperature is reached, or off when it is cool enough. A digital signal will accomplish this.

b.) An electrical signal should be used. This is because the limited space would not allow for the piping required by the pneumatic system. The digital signal could not be capable of instantaneous adjustments to the temperature because it is designed to be on or off.

c.) A pneumatic signal should be used. An electrical signal requires a high voltage for power and would be dangerous in this area. The digital signal would not allow for fine tuning of the flow rate out of the storage tank because of its binary nature, it could only open or close the valve.

Example 2: Assigning Control Loops
The following system utilizes the Haber process, which converts nitrogen and hydrogen to ammonia. (N2 + 3H2 -> 2NH3). The nitrogen comes from distillation of air at about -200°C, and the hydrogen comes from cracking of hydrocarbons in an adjacent process unit. Nitrogen leaves the top of the first distillation column and is met in the mixer by more cold nitrogen being recycled from the distillation of the product in the second column. The reaction occurs over an iron catalyst at about 350°C and 200 atm to produce the ammonia product. No hydrogen exits the reactor because nitrogen is in large excess. The incoming nitrogen and the exiting ammonia are run through a countercurrent heat exchanger. The cooled effluent enters a second distillation column where the ammonia is separated from the unreacted nitrogen. Ideally the heat exchanger will completely heat and cool the streams. However, steam can be supplied to the reactor if the temperature of the stream is not hot enough. Likewise, a refrigerant (stream 3) can be supplied to the cold stream to cool it down before entering the second distillation column if necessary. However, these processes are more costly. Actuators and sensors have been placed in appropriate locations. On the diagram show where the control loops should be placed to manipulate the actuators.



While several control schemes could potentially work, the solution diagrammed below should suffice. For simplicity, controllers were not drawn in, but assume the dotted lines represent the signal and an appropriate controller. Temperature is regulated by thermocouples that control the flow of steam or coolant to their respective units. Levels in the columns are controlled by connecting the sensors to outlet valve contollers, and the reactor pressure is maintained by connecting the pressure sensor to its outlet valve. Pressure sensors on the columns connect to release valves that open into the distillate receivers in order to maintain column pressure. Flow controllers on recycle streams connect to adjacent valves. These should be set to constant rates. Flow controller 3 is used to set the output level. If the rate is too high, it closes valve 4 and the resulting level rise in the distillate receiver causes the feed to slow down. If the rate is too low, it opens wider, drops the level, which in turn raises the feed rate.