Process Control systems: Fundamentals

Goals

This section covers the basic principles of process control.

• Clearly explain the concepts of:

- On-off control

- Modulating control

- Open loop control

- Ratio control.

• List the 10 most common acronyms and basic terminology used in the process control (e.g. PV, MV, OP).

• Describe the differences between a reverse and a direct acting controller.

• Indicate what deadtime is and how it impacts on a process.

On-off control

The oldest strategy for control is to use a switch giving simple on-off control. This is a discontinuous form of control action, and is also referred to as two-position control. The technique is crude, but can be a cheap and effective method of control if a fairly large fluctuation of the process variable (PV) is acceptable.

A perfect on-off controller is 'on' when the measurement is below the setpoint (SP) and the manipulated variable (MV) is at its maximum value. Above the SP, the controller is 'off' and the MV is a minimum.

On-off control is widely used in both industrial and domestic applications. Most people are familiar with the technique as it’s commonly used in home heating systems and domestic water heaters. Consider the control action on a domestic gas-fired boiler. When the temperature is below the setpoint, the fuel is 'on'; when the temperature rises above the setpoint, the fuel is 'off'.

There is usually a dead zone due to mechanical delays in the process. This is often deliberately introduced to reduce the frequency of operation and wear on the components.

The end result of this mode of control is that the temperature will oscillate about the required value.

===

Input signal; Output signal (m) 0 Differential gap Sinusoid mMAX mMIN Time +=+=+=+ 1 Response of a two positional controller to a sinusoidal input Time; Time Ideal curve (no delay) Process delay Off On Deadtime On Temperature Fuel flow Off

+=+=+=+ Graphical example of on-off control

==

Water heater Hot water Cool water Steam Temperature indicator - process disturbance Steam valve - manipulated variable (MV)

+=+=+=+ Concept of feedforward control

==

Modulating control

If the output of a controller can move through a range of values, this is modulating control.

Modulation control takes place within a defined operating range only, that is, it must have upper and lower limits. Modulating control is a smoother form of control than step control. It can be used in both open loop and closed loop control systems.

Open loop control

In open loop control, the control action (controller output signal OP) is not a function of the process variable (PV). The open loop control does not self-correct when the PV drifts, and this may result in large deviations from the optimum value of the PV.

Use of open loop control:

This control is often based on measured disturbances to the inputs to the system. The most common type of open loop control is feedforward control. In this technique the control action is based on the state of a disturbance input without reference to the actual system condition, i.e. the system output has no effect on the control action, and the input variables are manipulated to compensate for the impact of the process disturbances.

Examples of open loop control:

A common domestic application that shows open loop control is a washing machine.

The system is pre-set and operates on a time basis, going through cycles of wash, rinse and spin as programmed. In this case, the control action is the manual operator assessing the size and dirtiness of the load and setting the machine accordingly. The machine does not measure the output signal, which is the cleanliness of the clothes, so the accuracy of the process, or success of the wash, will depend on the calibration of the system. An open loop control system is poorly equipped to handle disturbances which will reduce or destroy its ability to complete the desired task. Any control system operating on a time base is an open loop. Another example of this is traffic signals. It’s difficult to implement open loop control in a pure form in most process control applications, due to the difficulty in accurately measuring disturbances and in foreseeing all possible disturbances to which the process may be subjected. As the models used and input measurements are not perfectly accurate, pure open loop control will accumulate errors and eventually the control will be inadequate.

Introduction to ratio control:

Ratio control, as its name implies, is a form of feedforward control that has the objective of maintaining the ratio of two variables at a specific value. E.g., if it’s required to control the ratio of two process variables XPV and YPV the variable PVR is controlled rather than the individual PVs (XPV and YPV).

Thus: A typical example of this is maintaining the fuel to air ratio into a furnace constant, regardless of maintaining or changing the furnace temperature. This is sometimes known as cross limiting control.

===

The objective is to keep the PV constant despite disturbances. To achieve this, the blocks FF-control and f(control) must change the PV by the same magnitude and timing but in opposite direction to that which the disturbance would have done without control. Then the feedforward control principle of compensating the disturbance is fulfilled.

Objective:

f(control) f(disturb) FF-control Disturbance (Change in feed flow) Feed flow Fuel flow PV T2 T2 =Outlet temperature Feedforward controller f(process)

+=+=+=+ Feedforward block diagram

===

Closed loop control

In closed loop control, the objective of control, the PV, is used to determine the control action.

===

Water heater Hot water Cool water Steam; Steam valve - manipulated variable (MV) Temperature indicator - process variable (PV)

+=+=+=+ Manual feedback control

PV ERR PID Process OP PV MV Adjust process Measure result OP =Manual + (P, I and D of ERR) Process gain= ?PV/?MV Controller gain = ?MV/?E(error) Loopgain (KLOOP)=KC(controller gain) × KP(process gain)

+=+=+=+ Closed loop block diagram

===

This is also known as feedback control and is more commonly used than feedforward control. Closed loop control is designed to achieve and maintain the desired process condition by comparing it with the desired condition, the setpoint value (SP), to get an error value (ERR).

Reverse or direct acting controllers:

As the controller's corrective action is based on the magnitude-in-time of the error (ERR), which is derived from either SP - PV or PV - SP it’s of no concern to the P, I or D functions of the controller which algorithm is used, as the algorithms only change the sign of the error term.

However; if we refer to (water level control), which shows a controller, performing the same function, but in different ways:

• In case one, we manipulate the outlet flow through V2 to control the tank level; this is direct action. Where as the PV increases (tank filling) the OP increases (opening the outlet valve more) to drain the tank faster.

Direct acting = PV OP then ERR = PV SP - ?? ?

• In case two, we control the inlet flow through V1 to control the tank level, this is reverse action. Where as the PV increases (tank filling) the OP decreases (closing the inlet valve more) to reduce the filling rate.

Reverse acting = PV OP then ERR = SP PV - ?? ?

The controller output changes, by the same magnitude and sign, based on the resultant error value and sign.

PID control ERR=SP - PV ERR=PV - SP SP PV= level V2 V1 OP(1) OP(2)

+=+=+=+ Direct and reverse acting controllers

Control modes in closed loop control:

Most closed loop controllers can be controlled with three control modes, either combined or separately. These modes, proportional (P), integral (I) and derivative (D) are discussed in-depth in the next section.

Illustration of the concepts of open and closed loop control:

The diagrams illustrate the concepts of open loop and closed loop controls in a water heating system.

• In the open loop, feedforward example, the steam flow rate is varied according to the temperature of the cool water entering the system. The operator must have the skills to determine what change in the valve position will be sufficient to bring the cool water entering the system to the desired temperature when it leaves the system.

• In the closed loop, feedback example, the steam flow rate is varied according to the temperature of the heated water leaving the system. The operator must determine the difference between that measurement and the desired temperature and change the valve position until this error is eliminated.

• The above example is for manual control but the concept is identical to that used in automatic control, which should allow greater accuracy of control.

Combination of feedback and feedforward control:

The advantages of feedback control are its relative simplicity and its potentially successful operation in the event of unknown disturbances. Feedforward control has the advantage of faster response to a disturbance in the input which may result in significant cost savings in a large-scale operation.

=== Feedforward inlet flow and temp variations

Inlet Fuel

The FC maintains a constant fuel flow, varied by the feedforward control, as a feedforward/ feedback configuration: SP FC PV F F1 T1 T2

+=+=+=+ 8 Block diagram of feedforward and feedback combination

===

In general, the best industrial process control can be achieved through the combination of both open and closed loop controls. If an imperfect feedforward model corrects for 90% of the upset as it occurs and the remaining 10% is corrected by the bias generated by the feedback loop, then the feedforward component is not pushed beyond its abilities, the load on the feedback loop is reduced, and much tighter control can be achieved.

Deadtime processes

In processes involving the movement of mass, deadtime is a significant factor in the process dynamics. It’s a delay in the response of a process after some variable is changed, during which no information is known about the new state of the process. It may also be known as the transportation lag or time delay. Deadtime is the worst enemy of good control and every effort should be made to minimize it. All process response curves are shifted to the right by the presence of deadtime in a process. Once the deadtime has passed, the process starts responding with its characteristic speed, called the process sensitivity.

0.63K K Slope = reaction rate T L = Effective dead time; Time constant (T ) Measurement

+=+=+=+ Process reaction or response curve, showing both deadtime and time constant

Reduction of deadtime:

The aim of good control is to minimize deadtime and to minimize the ratio of deadtime to the time constant. The higher this ratio, the less likely that the control system will work properly.

Deadtime can be reduced by reducing transportation lags, which can be done by increasing the rates of pumping or agitation, reducing the distance between the measuring instrument and the process, etc.

Deadtime effects on P, I and D modes and sample-and-hold algorithms:

If the nature of the process is such that the deadtime of a loop exceeds its time constant then the traditional PID (proportional-integral-derivative) control is unlikely to work, and a sample and hold control is used. This form of control is based on enabling the controller so that it can make periodic adjustments, then effectively switching the output to a hold state and waiting for the process deadtime to elapse before re-enabling the controller output. The algorithms used are identical to the normal process control ones except that they are only enabled for short periods of time.

The only problem is that the controller has far less time to make adjustments, and therefore it needs to do them faster. This means that integral setting must be increased in proportion to the reduction in time when the loop is in automatic.

===

Manual; Manual Dead time; Manual Auto; Auto Time Controller output (OP); Process variable (PV)

+=+=+=+ 10 Sample and hold algorithms are used when the process is dominated by large deadtimes

===

Process responses

The dynamic response of a process can usually be characterized by three parameters: process gain, deadtime and process lag (time constant).

+=+=+=+ Example of a process response related to a step change of the input value define the three constitutional parts of the process response curve.

Response process gain:

The process gain is the ratio of the change in the output (once it has settled to a new steady state) to the change in the input. This is the ratio of the change in the process variable to the change in the manipulated variable. It’s also referred to as the process sensitivity as it describes the degree to which a process responds to an input.

A slow process is one with low gain, where it takes a long time to cause a small change in the MV. An example of this is home heating, where it takes a long time for the heat to accumulate to cause a small increase in the room temperature. A high gain controller should be used for such a process.

A fast process has a high gain, i.e. the MV increases rapidly. This occurs in systems such as a flow process or a pH process near neutrality where only a droplet of reagent will cause a large change in pH. For such a process, a low gain controller is needed.

The three component parts of process gain from the controllers perspective is the product of the gains of the measuring transducer (KS), the process itself (KC) and the gain of what the PV or controller output drives (KV). This becomes: SCV Process gain = KKK ××

Response deadtime:

The deadtime (L) is the delay between the manipulated variable changing and a noticeable change in the process variable. Deadtime exists in most processes because few, if any, real world events are instantaneous. A simple example of this is a hot water system. When the hot tap is switched on there will be a certain time delay as hot water from the heater moves along the pipes to the tap. This is the deadtime.

Response process lag:

The process lag (T ) is caused by the system's inertia and affects the rate at which the process variable responds to a change in the manipulated variable. It’s equivalent to the time constant.

Dead zone

In most practical applications, there is a narrow bandwidth due to mechanical friction or arcing of electrical contacts through which the error must pass before switching will occur. This may be known as the dead zone, differential gap, or neutral zone. The size of the dead zone is generally 0.5-2% of the full range of the PV fluctuation, and it straddles the setpoint. When the PV lies within the dead zone no control action takes place, thus its presence is usually desirable to minimize the cycling of the process. One problem with on-off control is wear and tear of the controlling element. This is reduced as the bandwidth of fluctuation of the process is increased and thus frequency of switching decreased.

Exercise: Single Flow Loop - Flow Control Loop Basic Example This will give practical experience in the concept of closed loop control. It would be appropriate to do this exercise now, in order to become familiar with the concepts of closed loop control as well as the operation of the simulation software.

NEXT: Closed loops--Stability and control modes (part 1)
PREV: Valve sizing

All related articles   Top of Page   Home