A control system hierarchy

Control systems are ubiquitous in industrial, medical, military,

automotive, utility and many other applications - and sometimes
appear in our daily lives. However, they come in many types, of
which PID control is just one, albeit the most sophisticated.
Accordingly, we start by explaining the control system hierarchy that
exists today, and showing how PID control fits into this. Then, we
focus in on PID controllers and how they operate.

Open loop control – the most basic type. One example is a garden
hose, connected to a tap. Although the tap can be operated by a user
to regulate water flow, the hose has no feedback mechanism to
‘close the loop’ for automatic regulation.

On/off or ‘Bang-Bang’ control - this is the simplest form of closed

loop control, and is exemplified by a domestic heating system with a
room thermostat. The thermostat compares the room temperature
measured value (MV) with its setpoint (SP) as adjusted by its user.
The system uses the MV feedback to close the loop. If the
temperature is too low, it turns on the central heating to heat the
room until it reaches SP temperature. If the temperature exceeds the
SP, the heating will be turned off, and, possibly, air conditioning will
be started for active cooling.

Practical thermostats have a hysteresis, or dead-band. This could

typically be 4°C, so for a setpoint of 20°C, the thermostat turns on at
18°C and below, and off at 22°C and above. This exposes the
limitations of on/off control; without hysteresis, the thermostat would
switch continuously, which may cause excessive wear and tear, and
possibly intrusive audible noise. Conversely, if the hysteresis gap is
too wide, the room temperature will be sub-optimal for much of the
time.

Additionally, in some industrial systems, a sudden and full load turn-

on could lead to issues with inrush currents.
Proportional control is a continuous control method that offers more
nuanced, robust and better quality performance than the relatively
crude on/off approach. However, designing and optimizing a control
loop for a particular application is not without some challenges. It is
usually desirable if not essential to achieve the fastest possible
response time and greatest steady-state accuracy, without creating a
risk of instability. Once set up, the loop must maintain its optimum
performance even during unpredictable and changing process
conditions.

Below, we look at how proportional control systems operate, and how

their performance can be assessed. Then we introduce the concepts
of the Integral (I) and Derivative (D) functions, and explain their
essential role in optimizing closed loop control performance for PID
controllers. We see how the functions can be tuned to best balance
response time and accuracy against stability under all conditions.

Today, most proportional and PID controllers are implemented either

using embedded microprocessors or microcontrollers, or as software
running on a PLC or larger supervisory control and data acquisition
(SCADA) system. However, there are exceptions. Analogue PID
controllers are still used for high bandwidth and low noise
applications, and pneumatic systems are still available .

Note that, irrespective of the PID controller’s technology, the input

and output variables are usually analogue.

In this configuration, the process variable is the system parameter

that needs to be controlled, such as temperature, pressure or flow
rate. The sensor measures this variable and feeds back a
corresponding signal, which is usually electrical in a modern process
control system; typically, 4 – 20mA. This is compared with the
setpoint value, which could be adjusted by a user as in the
thermostat example. However, in an industrial system, it may be set
by another process, or by a programmable logic controller (PLC)
control program.
The output from this comparison is the Error signal, which is used by
the controller to calculate its output to the process. The controller’s
proportional gain (Kp) determines the ratio of output response to the
error signal.

In any case, the controller’s output is used to drive an actuator –

which could be a heater in an industrial process for example.
Accordingly, as the process heats towards the setpoint, the error
signal is reduced as its temperature approaches its desired level.
However, proportional control alone cannot provide perfectly
accurate control; it cannot, by itself, completely eliminate the error
signal, as the output becomes negligible as the error signal
approaches zero. Additionally, the actuator output is often not the
only influence on the system.

For instance, in a temperature chamber there might be a source of

cool air that sometimes blows into the chamber and changes the
temperature. This is called a disturbance. Additionally, the system’s
response to a control output may change over time or in response to
some variable. For example, a chamber partially filled with fluid will
exhibit a much faster response to heater output when nearly empty
than it will when nearly full of fluid. This creates a nonlinear system in
which the control parameters that produce a desired response at one
operating point might not produce a satisfactory response at another
operating point.

Deadtime, caused by factors like delays arising from fluid flowing

through pipes, can be another issue. Accordingly, control systems
should be designed to minimize the effects of disturbances on the
process variable, nonlinear processes, and deadtime.

RENEWABLE ENERGY
Unlocking the benefits of micro hydro power

The control design process

The control design process begins by defining the performance
requirements. Control system performance is often measured by
applying a step function to the set point input, and then measuring
the response of the process variable. This is commonly quantified by
measuring defined waveform characteristics.

Rise time is the amount of time the system takes to go from 10 % to

90 % of the steady-state, or final, value. Percent overshoot is the
amount that the process variable overshoots the final value,
expressed as a percentage of the final value. Settling time is the time
required for the process variable to settle to within a certain
percentage (commonly 5 %) of the final value. Steady state error is
the final difference between the process variable and set point. Note
that the exact definition of these quantities will vary in industry and
academia.

What is a PID controller?

We will now look at the effects of varying the Proportional component
of the controller’s output, and of introducing Integral (I) and Derivative
(V) terms.

If we have just a Proportional (P) controller, we can start by setting

the Kp gain to a modest value, such as 10 (K is dimensionless). After
applying the step input, the output will be stable, but will approach the
desired value slowly. If Kp is increased to, say, 100, both the steady
state error and the rise time will be reduced – so far, so good.
However, if we increase Kp to 200, although the rise time and steady
state error are further reduced, overshoot starts to appear.

From here on, further increasing Kp merely increases overshoot, with

no further reductions in rise time or steady state error. And overshoot
may eventually become underdamped oscillation, meaning that the
system will be unstable.

The industry’s solution to this has been to add Integral (I) and
Derivative (V) components to the controller’s Proportional (P) output,
to build a PID controller.

PCIM EUROPE 2020

SiC power module for high-temperature automotive application
The Integral component sums the error term over time. The result is
that even a small error term will cause the integral component to
increase slowly. The integral response will continually increase over
time unless the error is zero, so the effect is to drive the steady state
error to zero. A phenomenon called integral windup results when
integral action saturates a controller without the controller driving the
error signal toward zero.

Although integral action can eliminate the steady state error, it can
strongly contribute to controller output overshoot and possible
instability. Nevertheless, PI control can be suitable for some
processes, where system speed is not important. However, where
instability is a potential concern, the solution is to implement a PID
controller by adding a Derivative component.

Derivative components act on the rate of change of the error signal.

The more error changes or the longer the derivative time, the larger
the derivative factor becomes. The effect of this is to counteract the
overshoot caused by P and I. When the error is large, the P and the I
will push the controller output. This controller response makes error
change quickly, which in turn causes the derivative to counteract the
P and the I more aggressively. A properly used derivative allows for
more aggressive proportional and integral factors. Larger derivative
time makes the derivative more aggressively dampen P and I.

Watch also this video to get PID Controller explained!

Tuning a PID controller

Establishing the optimum combination of P,I and D parameters for a
specific control loop is done by tuning – and three approaches to
tuning are possible: manual, heuristic (‘Rule of thumb’), and
automated.

Manual PID tuning is done by setting the reset (integral) time to its
maximum value and the rate (Derivative) to zero, and increasing the
gain until the loop oscillates at a constant amplitude. (When the
response to an error correction occurs quickly a larger gain can be
used. If response is slow a relatively small gain is desirable). Then
set the gain of the PID controller to half of that value and adjust the
reset time so it corrects for any offset within an acceptable period.
Finally, increase the rate of the PID loop until overshoot is minimized
POWER ELECTRONICS
A brief history of power electronics and why it's important

Zeigler and Nichols’ two heuristic methods of tuning a PID

controller were first published in 1942. These work by applying a step
change to the system and observing the resulting response. The first
method entails measuring the lag or delay in response and then the
time taken to reach the new output value. The second depends on
establishing the period of a steady-state oscillation. In both methods,
these values are then entered into a table to derive the values for
gain, reset time and rate for the PID controller.

Most PID controllers sold today incorporate auto-tuning functions.

Operating details vary between manufacturers, but all follow rules
similar to those described above. Essentially, the PID controller
“learns” how the process responds to a disturbance or change in set
point, and calculates appropriate PID settings.

OMEGA’s Platinum series of temperature controllers are examples of

devices with auto-tuning capability.

What are PID tuning parameters? Find the answers in this video!

Designing PID based control systems – from your computer

NI’s LabVIEW toolset features a wide variety of Virtual Instruments
(VIs) that greatly help in the design of a PID based control system .
PID VIs allow development of Proportional (P); proportional-integral
(PI); proportional-derivative (PD); and proportional-integral-derivative
(PID) algorithms.

Conclusion
In this article, we have seen how PID controllers can be used to
optimize control loops for response time, accuracy, and stability
under all process conditions – and how the I and D terms provide a
level of controllability not possible with the Proportional term alone.

FOLLOW US ON LINKEDIN