PID

CONTROLLERS
 Feedback Controllers
Controllers is the “brain” of the control loop.
It decides “what to do” based on the error between the desired set point
value (SP) of controlled variable and its measured value.

𝑒 𝑡 = 𝑦𝑠𝑝 𝑡 − 𝑦𝑚 𝑡

Where: 𝑒 𝑡 = error signal (offset)

𝑦𝑠𝑝 𝑡 = set point
𝑦𝑚 𝑡 = measured value of the controlled variable
(or equivalent signal from the sensor/transmitter)

Objective of feedback controller is to reduce the error signal to zero

PID Controllers
MV = Proportional (P)
controller E SP = Set
output point
+ Integral (I)
+
- CV =
Derivative (D) Controlled
variable
Note: Error = E  SP - CV sensor
Final
Process
element
variable
PROCESS

Three “modes”: Three ways of using the time-varying

behavior of the measured variable
 Disturbance
Variable (DV)

D(s)
Disturbance
TF
GD(s)
Manipulated
Error Variable (MV)
+ E(s) P(s) X(s) Y(s)
GC(s) GV(s) GP(s)
YSP(s) _

Controlled Variable
Set point Controller TF Actuator TF Process
TF
Ym(s) Y(s)
GM(s)
Measure Variable
Sensor / transmitter TF
1. Proportional Control
For proportional control, the controller output is proportional to the
error signal, 𝑝 𝑡 varies linearly with 𝑒 𝑡

𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑒 𝑡
where: 𝑝 𝑡 = controller output
𝑝lj = bias (steady−state) value
𝐾𝑐 = controller gain (usually dimensionless)

• When error (offset) has zero value, the controller output signal
reaches its steady state bias value (𝑝).
lj
• Action of controller: as absolute error increases, the controller
output signal must increase to give large change in manipulated
variable
1. Proportional Control
The key concepts behind proportional control are the following:
1. The controller gain (Kc) can be adjusted to make the controller
output changes as sensitive as desired to deviations between set
point and controlled variable;
2. The sign of Kc can be chosen
𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑦𝑠𝑝 𝑡 − 𝑦𝑚 𝑡
• Direct acting controller (Kc < 0) → choose negative Kc value
to increase the controller output signal as the measured
controlled variable increase above set point.
• Reverse acting controller (Kc > 0) → choose positive Kc
value to increase the controller output signal as the measured
controlled variable decrease below set point.
3. For proportional controllers, bias 𝑝lj can be adjusted, a procedure
referred to as manual reset.
1. Proportional Control

As error changes, 𝑝 𝑡 will change immediately (fast corrective

action in very simple form).
Reverse and direct-acting
proportional controllers.
(a) reverse acting (Kc > 0)
(b) direct acting (Kc < 0)

8
1. Proportional Control
Some controllers have proportional band setting instead of a
controller gain.

100%
𝑃𝐵 =
𝐾𝑐

• Proportional band is defined only for dimensionless Kc.

• As will be shown later, the final error “offset” can be reduced
by increasing the absolute value of Kc (reducing PB).
• Introducing very high values of Kc will lead to oscillatory
response or even unstable situation.
1. Proportional Control
Effects of Kc on the controlled variable response under step
change in disturbance variable:
1. Proportional Control: Transfer Function
In order to derive the transfer function for proportional controller,
define a deviation variable 𝑝′ 𝑡 as

𝑝′ 𝑡 = 𝑝 𝑡 − 𝑝lj

Then equation can be written as

𝑝′ 𝑡 = 𝐾𝑐 𝑒 𝑡
Laplace

𝑃′ 𝑠 = 𝐾𝑐 𝐸 𝑠
The transfer function for proportional-only control:

𝑃′ 𝑠
= 𝐾𝑐
𝐸 𝑠
Disadvantage of P-only controller:
• Tend to produce offset. → error between actual steady state
value and set point.
• Offset can be eliminated by manually resetting the set point
(ysp) or by using controller with integral action

Application:
• P-only control → for process with non-sluggish
dynamic behaviour and when offset is not important.
• Pressure or level control
2. Integral Control
For integral control action, the controller output depends on the
integral of the error signal over time,
1 𝑡 ∗
𝑝 𝑡 = 𝑝lj + න 𝑒 𝑡 ​𝑑 𝑡 ∗
τ𝐼 0
where τ𝐼 : integral time or reset time, has units of time.

• Integral control action is widely used because it provides an important

practical advantage, the elimination of offset.
• Integral-only control action seldom used by itself.
• Integral control action is normally used in conjunction with proportional
control as the proportional-integral (PI) controller:
PI Controllers
Synonyms : “Reset” or “floating” control
Proportional-integral (PI) controller signal output:
1 𝑡 ∗
𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑒 𝑡 + න 𝑒 𝑡 𝑑𝑡 ∗
τ𝐼 0
• Some commercial controllers are calibrated in terms of 1/τ𝐼 (repeats per
minute) rather than τ𝐼 (minutes, or minutes per repeat).
1
• Reset rate →τ𝑅 =
τ𝐼

• Infinite integral time or zero reset → P controller

Advantage: Disadvantages:
• Fast action • Oscillatory or unstable with integral control
• Eliminate offset • Need two parameters to be adjust (tune)
(𝐾𝑐 𝑎𝑛𝑑 τ𝐼 )
PI Controllers
The transfer function for the PI controller

1 𝑡 ∗
𝑝 𝑡 − 𝑝lj = 𝐾𝑐 𝑒 𝑡 + න 𝑒 𝑡 𝑑𝑡 ∗
τ𝐼 0
𝐾𝑐 𝑡 ∗
𝑝′ 𝑡 = 𝐾𝑐 𝑒 𝑡 + න 𝑒 𝑡 𝑑𝑡 ∗
τ𝐼 0 See slide Laplace
transform of Integral
′ 𝑠
𝐾𝑐 𝐸 𝑠
𝑃 = 𝐾𝑐 𝐸 𝑠 +
τ𝐼 𝑠

𝑃′ 𝑠 1
= 𝐾𝑐 1 +
𝐸 𝑠 τ𝐼 𝑠

𝑃′ 𝑠 τ𝐼 𝑠 + 1
= 𝐾𝑐
𝐸 𝑠 τ𝐼 𝑠
 PI Controllers

K𝑐
Integral action =
τ𝐼
Disadvantage of integral controller:
• Tend to produce oscillatory responses of the controlled
variable.
• Oscillatory response can be eliminated by proper tuning
of the controller or by including derivative action.

Application:
• PI control → when offset elimination is important
• Flow, Temp, composition
3. Derivative Control
• The function of derivative control action is to anticipate the
future behavior of the error signal by considering its rate of
change.
• For ideal derivative action
𝑑𝑒 𝑡
𝑝 𝑡 = 𝑝lj + τ𝐷
𝑑𝑡

where τ𝐷 , the derivative time, has units of time.

• Derivative action control is never used alone.

• Always in conjunction with proportional or proportional
integral control
PD controller 𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑒 𝑡 + 𝐾𝑐 τ𝐷
𝑑𝑒 𝑡
• For ideal PD control 𝑑𝑡

• Zero derivative time → P controller

• Derivative mode is used to improve dynamic response of the
controlled variable by decreasing process settling time.
• Derivative mode does not eliminate the offset
• an ideal PD controller has the transfer function:
𝑃′ 𝑠
= 𝐾𝑐 1 + τ𝐷 𝑠
𝐸 𝑠
• An ideal PD controller is physically unrealizable because it cannot
be implemented using analog or digital components.
• The process measurement is noisy if it contains high frequency or
random fluctuation.
• Derivative action will amplify the noise
PD controller : Transfer Function
• The transfer function in “real” PD controller can be
approximated by

𝑃′ 𝑠 τ𝐷 𝑠
= 𝐾𝑐 1 +
𝐸 𝑠 ατ𝐷 𝑠 + 1
derivative filter
where the α constant (0.05 < α < 0.2)
(0.1 being a common choice).

• Above Eq includes a derivative filter that reduces the

sensitivity of the calculations to high-frequency noise in the
measurement.
• This controller is physically realizable
Proportional-Integral-Derivative (PID)
Control
• Now we consider the combination of the proportional,
integral, and derivative control modes as a PID controller.
• A survey indicated that large-scale continuous processes
typically have 500 and 5,000 feedback controllers for
variable such as flow rate and liquid level
• Of these controllers, 97% utilize some form of PID control.

• Many variations of PID control are used in practice.

• Next, we consider the three most common forms.
• Parallel Form of PID Control
• Series Form of PID Control
• Expanded Form of PID Control
 1. Parallel Form of PID Control
The parallel form of the PID control (without a derivative filter)
is given by 1 𝑡 ∗ 𝑑𝑒 𝑡
∗
𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑒 𝑡 + න 𝑒 𝑡 𝑑𝑡 + τ𝐷
τ𝐼 0 𝑑𝑡
The corresponding transfer function is:
𝑃′ 𝑠 1
= 𝐾𝑐 1 + + τ𝐷 𝑠
𝐸 𝑠 τ𝐼 𝑠

Block diagram of the parallel form of PID control (without a derivative filter)
1. Parallel Form of PID Control
The parallel form of the PID controller transfer function (with a
derivative filter) is given by

𝑃′ 𝑠 1 τ𝐷 𝑠
= 𝐾𝑐 1 + +
𝐸 𝑠 τ𝐼 𝑠 ατ𝐷 𝑠 + 1

Block diagram of the parallel form of PID control (with a derivative filter)
2. Series Form of PID Control
The series-form of PID controller without derivative filter

𝑃′ 𝑠 τ𝐼 𝑠 + 1
= 𝐾𝑐 τ𝐷 𝑠 + 1
𝐸 𝑠 τ𝐼 𝑠

E(s) τ𝐼 𝑠 + 1 P’(s)
K𝑐 τ𝐷 𝑠 + 1
τ𝐼 𝑠

The series-form of PID controller with derivative filter

𝑃′ 𝑠 τ𝐼 𝑠 + 1 τ𝐷 𝑠 + 1
= 𝐾𝑐
𝐸 𝑠 τ𝐼 𝑠 ατ𝐷 𝑠 + 1

E(s) τ𝐼 𝑠 + 1 τ𝐷 𝑠 + 1 P’(s)
K𝑐
τ𝐼 𝑠 ατ𝐷 𝑠 + 1
3. Expanded Form of PID Control
• In addition to the well-known series and parallel forms, the expanded
form of PID control in Eq. 8-16 is sometimes used:
• The control parameter are 3 gains: 𝐾𝑐 , 𝐾𝐼 , 𝐾𝐷
𝑡
𝑑𝑒 𝑡
𝑝 𝑡 = 𝑝lj + 𝐾𝑐 𝑒 𝑡 + 𝐾𝐼 න 𝑒 𝑡∗ 𝑑𝑡 ∗ + 𝐾𝐷
0 𝑑𝑡
• The transfer function of expanded form is:

𝑃′ 𝑠 𝐾𝐼
= 𝐾𝑐 + + 𝐾𝐷 𝑠
𝐸 𝑠 𝑠
Controller Comparison
• P → Simplest controller to tune (𝐾𝑐 )
→Offset occur during disturbance or set point change.
• P I → More complicated to tune (𝐾𝑐 and τ𝐼 )
→Better performance than P-only.
→offset elimination
→ Most popular FB controller
• P ID → Most complicated to tune (𝐾𝑐 , τ𝐼 and τ𝐷 )
→Better performance than PI.
→offset elimination
→ Derivative action may be affected by noise
Key characteristic of commercial PID controller
Automatic and Manual Control Modes

• Automatic Mode
Controller output, p(t), depends on e(t), controller
constants, and type of controller used. ( PI vs. PID etc.)
• Manual Mode
Controller output, p(t), is adjusted manually.
Manual Mode is very useful when unusual conditions
exist:
✓ plant start-up
✓ plant shut-down
✓ emergencies
Typical Response of Feedback Control
Systems
Consider response of a controlled system after a sustained disturbance occurs
(e.g., step change in the disturbance variable)

• No control: the process slowly

reaches a new steady state.
• P : speed up the process
y response and reduce the offset
• PI : eliminate offset and the
response more oscillatory
• PID : reduce degree of
oscillation and response time

Figure 8.12. Typical process responses with feedback control.

Typical Response of Feedback Control
Systems

Figure 8.13. Proportional control: effect of controller gain.

• Increasing Kc tends to make the process response faster (less

sluggish)
• Too large Kc results in undesirable degree of oscillation or even
become unstable
• Intermediate value of Kc usually result in the best control
Typical Response of Feedback Control
Systems

y y

Figure 8.14. PI control: (a) effect of reset time (b) effect of controller gain.

• Increasing τ𝐼 tends to make the process response more sluggish

(slower)
• Too large τ𝐼 : the controlled variable will return to the set point very
slowly after a disturbance or set point change occurs.
Typical Response of Feedback Control
Systems

Figure 8.15. PID control: effect of derivative time.

• Increasing τ𝐷 tends to improve process response by reducing the

maximum deviation, response time and degree of oscillation
• Too large τ𝐷 : measurement noise is amplified and process response
more oscillatory
• Intermediate value of τ𝐷 is desirable
Selection of Controller
• Should consider the combination behavior of :
▪ Final control element
▪ Process
▪ Sensor / transmitter
• P-only control: for process with non-sluggish dynamic
behavior and offset is not important (e.g. pressure, level)
• PI control : when offset elimination is important (e.g. Flow,
Temp., Composition)
• PID : for sluggish process (biomass concentration, Temp.,
Composition)
END