PT37

Loop Tuning: Eliminating the

Guesswork of Tuning PID
Controllers

Robert C Rice • Vice President, Engineering

PID tuning can be time consuming and achieving the desired results can be difficult. This presentation explores a
repeatable method proven to be effective when tuning either simple or complex PID control loops.

Control Station www.controlstation.com

 Agenda

Motivation to Process Optimization

What is a PID Controller

6 Steps to Controller Tuning

Special Considerations

“Live” Tuning Example

Control Station www.controlstation.com

 “Loose” Control Costs Money
Tuning + Optimization
A well controlled process has less variability in
the measured Process Variable (PV), so the
process can be operated close to the
maximum profit constraint.

Control Station www.controlstation.com

 PID Control Basics
Basic Rockwell PID Forms
Dependent Form Independent Form
𝐾𝐾𝑐𝑐 𝑑𝑑𝑑𝑑(𝑡𝑡) 𝑑𝑑𝑑𝑑(𝑡𝑡)
𝐶𝐶𝐶𝐶 = 𝐶𝐶𝐶𝐶𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 + 𝐾𝐾𝑐𝑐 � 𝑒𝑒(𝑡𝑡) + � 𝑒𝑒 𝑡𝑡 𝑑𝑑𝑑𝑑 + 𝐾𝐾𝑐𝑐 � 𝑇𝑇𝑑𝑑 𝐶𝐶𝐶𝐶 = 𝐶𝐶𝐶𝐶𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 + 𝐾𝐾P � 𝑒𝑒(𝑡𝑡) + 𝐾𝐾I � 𝑒𝑒 𝑡𝑡 𝑑𝑑𝑑𝑑 + 𝐾𝐾D
Ti 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃
𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼

where: where:
𝐾𝐾𝑐𝑐 = controller gain 𝐾𝐾P = controller gain
Ti = controller reset time 𝐾𝐾I = integral gain
Td = controller derivative time 𝐾𝐾D = derivative gain

PID (Ladder) Uses the Positional PID Form while the PIDE uses the Velocity PID Form. The
PIDE is better suited for advanced PID applications using cascade, adaptive or other advanced PID logic

Control Station www.controlstation.com

 PID Control Basics
Function of the P, I and D Term
• Proportional term considers how far PV is from SP at any instant in time
and adds or subtracts from CObias accordingly (recall e(t) = SP – PV)
• Integral term addresses how long and how far PV has been from SP by
continually summing e(t) over time
• Derivative term considers how fast e(t) is changing at any instant using the
rate of change or slope of the error curve

rapidly changing e(t) = large derivative = large impact on CO

Derivative doesn’t consider if e(t) is positive, negative or how much time has
passed, just how fast e(t) is changing

Control Station www.controlstation.com

 PID Control Basics
PI Control – Parameter Influence

Control Station www.controlstation.com

 6 Steps to Controller Tuning
LOOP-PRO Tuning Recipe: 6-Step Process

Control Station www.controlstation.com

 Step 1 - Find
Find Controller & Define Controller Objective
Safety – What is the worst thing that can happen is this control loop fails? When designing your control
objective, the safety considerations are paramount to all others

Impact – Where does this control loop fit in the overall process diagram. Where do the process
disturbances come from? When this process changes, who gets impacted by it?

Management – If management desires a certain type of performance, it is your job to match that
objective or explain in a logical manner why that type of control is not possible.

Profit – What are the primary economics factors associated with this loop?

Longevity – How often is this loop the ‘culprit’ when diagnosing process performance? Keep the control
strategy simple, the more complicated the strategy, the more likely it will fail.

Equipment – Process equipment is expensive, its components are expensive, and as such, should be
included when formulating your control objective.

Control Station www.controlstation.com

 Step 1 - Find
The user can browse the various Topics and Folders within the OPC Server
address space to scan and eventually select the controller of their choice.

• Directory Tree lists Topics and Folders the user • Any controllers available found by the Scan are listed
can choose from. here in the “Found Controllers” Tab.
• “Deep Scan” Button initiates the scan process • This “Scanning Status Report” details the amount of the
from the currently selected location on the Tree. server space that has been successfully scanned.

Control Station www.controlstation.com

 Step 2 - Bump
Bump the Process
• The data must show a sharp, sudden CO
change of at least 3%.
• Changes of at least 3% ensure that the
valve or damper will actually move;
smaller changes may be inhibited by
stiction or dead-band.
• When a closed loop test is performed,
the current controller tuning values must
be aggressive enough that the SP bump
causes a sharp, sudden CO change.
• If the CO change is ramped, generating
an accurate model is not possible.

Control Station www.controlstation.com

 Step 2 - Bump
Bump the Process
Step Test Pulse Test Doublet Test

• Controller output is • Two step tests performed • Two pulse tests performed
stepped from one in rapid succession. in rapid succession and in
constant value to another. • The controller output is opposite directions.
• It results in the measured stepped up and, as soon • The second pulse is
variable moving from one as the measured variable implemented as soon as
steady state to a new shows a clear response, the process has shown a
steady state. the controller output is clear response to the first
returned to its original pulse.
value.

Control Station www.controlstation.com

 Step 2 - Bump
Bump the Process
• The data must be sampled fast enough to
accurately capture the process dynamics; if
this is not done, the data may show an
inaccurate trend, known as aliasing.
• Flow loop where data was collected at a rate
of 30 seconds.
• Plant’s historian only shows the starting and
ending points associated with the increases
to controller output.
• This data cannot be used to generate an
appropriate model of the process since it is
not possible to determine how fast or with
how much delay the process responded to
the CO steps.

Control Station www.controlstation.com

 Step 2 - Bump
Data Sample Rate – Rules of Thumb
• Data should be collected at a minimum
of 10 times faster than the rate of the
process time constant (𝜏𝜏𝑃𝑃 ).
• The following guidelines may be used
when determining the appropriate data
sample rate for PID loop tuning:

Control Station www.controlstation.com

 Step 2 - Bump
Collect Loop Data Tab: Perform a Bump/Step Test
• The bump test should cause an Output
change that drives the Measured Variable
far enough and fast enough such that the
response dominates the random noise in
the Measured Variable signal.
• Change Set Point to a degree 4-5 x’s the
noise band of the process.

Bump tests can be performed in manual mode by bumping the Controller Output, or in
automatic mode by bumping the Set Point.

Control Station www.controlstation.com

 Step 3 - Model
Fit a Model
Non Self-Regulating
Self-Regulating
(Integrating)

If all inputs are held Process will only

constant, the reach a steady-state
process will seek a at its ‘balancing’
steady-state point
• Ex. Heat Exchanger • Ex. Surge Tank

Control Station www.controlstation.com

 Step 3 - Model
Fit a Model

FOPDT Self-Regulating FOPDT Non Self-Regulating

𝑑𝑑𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑𝑑𝑑 ∗
𝜏𝜏𝑃𝑃 + 𝑃𝑃𝑃𝑃 = 𝐾𝐾𝑃𝑃 � 𝐶𝐶𝐶𝐶 𝑡𝑡 − 𝜃𝜃𝑃𝑃 = 𝐾𝐾𝑃𝑃 � 𝐶𝐶𝐶𝐶 𝑡𝑡 − 𝜃𝜃𝑃𝑃
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑃𝑃𝑃𝑃 ∗ 𝑃𝑃𝑃𝑃
𝐾𝐾𝑃𝑃 → Process Gain 𝐾𝐾𝑃𝑃 → Integrator Gain
𝐶𝐶𝐶𝐶 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡�𝐶𝐶𝐶𝐶
𝜏𝜏𝑃𝑃 → Time Constant [time] 𝜃𝜃𝑃𝑃 → Deadtime [time]
𝜃𝜃𝑃𝑃 → Deadtime [time]
All models are wrong, some are useful -George Box

Control Station www.controlstation.com

 Step 3 - Model
Tunings are Only as Good as the Model

Control Station www.controlstation.com

 Step 3 - Model
Generate Model
• Select the “Show SP” box in order to
track how the process
responded/model fits in response to
the Set Point Change.
• Yellow Line displays a graphical
representation of the currently
generated model fit by NSS.
• Each model built has a corresponding
line of data on the bottom end of this
tab.
• This “Composite Model” section is an
average of all of the model parameters
listed in the model table below.
Control Station www.controlstation.com
 Step 4 - Tune
PID Tuning Correlations for the Dependent Ideal Non-Interacting PID
𝐾𝐾𝐶𝐶 𝑑𝑑𝑑𝑑𝑑𝑑 𝑡𝑡
𝐶𝐶𝐶𝐶 = 𝐶𝐶𝐶𝐶𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 + 𝐾𝐾𝐶𝐶 � 𝑒𝑒 𝑡𝑡 + � 𝑒𝑒 𝑡𝑡 𝑑𝑑𝑑𝑑 − 𝐾𝐾𝐶𝐶 � T𝐷𝐷
T𝐼𝐼 𝑑𝑑𝑑𝑑
First compute, τ𝐶𝐶 , the closed loop time constant
Choose your performance using these rules:
• aggressive: τ𝐶𝐶 is the larger of 0.1 τ𝑝𝑝 or 0.8 𝜃𝜃𝑝𝑝
• moderate: τ𝐶𝐶 is the larger of 1 τ𝑝𝑝 or 8 𝜃𝜃𝑝𝑝
• conservative: τ𝐶𝐶 is the larger of 10 τ𝑝𝑝 or 80 𝜃𝜃𝑝𝑝
P I D
1 τ𝑝𝑝 τ𝑝𝑝
PI
𝐾𝐾𝑃𝑃 𝜃𝜃𝑝𝑝 + τ𝐶𝐶
1 τ𝑝𝑝 + 0.5𝜃𝜃𝑝𝑝 τ𝑝𝑝 + 0.5𝜃𝜃𝑝𝑝 τ𝑝𝑝 𝜃𝜃𝑝𝑝
PID
𝐾𝐾𝑃𝑃 τ𝐶𝐶 + 0.5𝜃𝜃𝑝𝑝 2τp + 𝜃𝜃𝑝𝑝

Control Station www.controlstation.com

 Step 4 - Tune
Impact of More Aggressive Control

Tighter Disturbance Rejection More Overshoot on Set-Point Tracking

Disturbance Rejection Set-Point Tracking
25.3 25.6

25.25 25.5

25.2 25.4

decreasing 𝜏𝜏𝐶𝐶
25.15 25.3
decreasing 𝜏𝜏𝐶𝐶
25.1 25.2

25.05 25.1

25 25
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35

Set-Point Moderate Moderately Aggressive Aggressive Set-Point Moderate Moderately Aggressively Aggressively

Control Station www.controlstation.com

 Step 4 - Tune
Ways of Quantifying Disturbance Rejection
Disturbance Rejection
25.3

25.25

25.2

25.15

25.1

Arrest Time
𝜏𝜏𝐶𝐶 Integrating
Control Loops Recovery Time
25.05
𝜏𝜏𝐶𝐶 Self-Regulating Control Loops

25
0 5 10 15 20 25 30 35

Control Station www.controlstation.com

 Step 4 - Tune
Basic Rules for Closed Loop Time Constant

Flow Loops
τ𝐶𝐶  3 to 5 times the Open Loop Time Constant, τ𝑝𝑝
Pressure Loops
τ𝐶𝐶  2 to 4 times the Open Loop Time Constant, τ𝑝𝑝
Temperature Loops
τ𝐶𝐶  1 to 3 times the Open Loop Time Constant, τ𝑝𝑝

Control Station www.controlstation.com

 Step 4 - Tune
Calculate Tuning Parameters
• This section displays information about
the currently connected PID Block.
• Composite Model parameters are
adjustable by the user to generate the
“Recommended Tuning Parameters”.
• Adjust the controller response by
adjusting the “Closed Loop Time
Constant Slider Bar” aggressively or
conservatively.
• Interpret and use the
robustness/stability plot and the PID
response statistics.

Control Station www.controlstation.com

 Step 5 - Test
Implement and Test Results
• Testing PID Controllers Typically Involve:
• Adjust Set-Point to ensure adequate
tracking
• Did the Process Variable Overshoot?
• Did the Controller Output Move too
much?
• Introduce a Load Change or Disturbance
• Did the Process Variable Recover quick
enough?
Updated Tuning Parameters MUST be tested
NOTE: PID Controllers work off of controller error (SP-PV), if there is no error, there is nothing for the
PID Controller to do. You MUST introduce controller error, and force the controller to respond before
you know if your tuning changes improved the system.

Control Station www.controlstation.com

 Step 5 - Test
Implement and Test Results
• After implementing any change, the
user must test the new tunings by
doing a Bump/ Step Test.
• Bump the process by manipulating the
Set Point.
• In the case you are not satisfied with
the new tunings, you can Click the
“Send Original Tunings” to maintain the
current values.

Control Station www.controlstation.com

 Step 6 - Document
Document the Results
Who:
Who is accountable for the changes?
What:
Which loop has been tuned, what were
the As Found and Recommended Tuning
Values?
When:
When was the Loop Adjusted?
Why:
Why was this particular loop tuned?

Control Station www.controlstation.com

 Special Considerations for Bump
Tests

Control Station www.controlstation.com

 Tuning Flow Controllers
The flow controller is one of the simplest control loops
• Control valve is manipulated by a
controller to maintain flow at the
desired flow rate
• In general, Flow Controllers are
Tuned Slower in CLOSED Loop than
in OPEN Loop.
• This helps improve controller
robustness

Control Station www.controlstation.com

 Tuning Flow Controllers
What is Valve Stiction
• Static friction counteracts opposing
external forces below a certain threshold
• In control valves, this often describes the
stem of the valve becoming “sticky”
when small changes are attempted.
• The result is that the force required to
get the stem to move is more than is
required to get to the desired stem
position, causing the movement to be
jumpy.

Control Station www.controlstation.com

 Tuning Flow Controllers
What is Valve Stiction
• To determine how much stiction is
present in a valve:
• Place the controller in manual mode
• Incrementally increase the controller
output until the flow rate responds
• If stiction is present in the valve, the
valve won’t actually move until the
CO is increased by the amount of
stiction that is present in the valve.

Control Station www.controlstation.com

 Tuning Flow Controllers
Bump test should be larger than valve stiction

Control Station www.controlstation.com

 Tuning Level Controllers
What is the Control Objective?
Controller Objective: Minimize Flow
Variability in the controlled flow out
without exceeding alarm conditions
Scenario
Tank Volume = 100 Liters
Largest Wild-Stream Flow
Fluctuation 2 Liters / Minute
10 Liters of “surge-space” available

Arrest Time < 5 Minutes

Control Station www.controlstation.com 32

 Tuning Level Controllers
For Integrating Process (i.e.: Level Controller)
Arrest Time = 𝜏𝜏𝐶𝐶

Control Station www.controlstation.com

 Tuning Level Controllers
Open Loop Integrating Process

Control Station www.controlstation.com

 Tuning Level Controllers
Closed Loop Integrating Process Tests

Control Station www.controlstation.com 35

 Tuning Cascaded Controllers
Closed Loop Cascade Bump Test

Inner Loop Outer Loop

Control Station www.controlstation.com 36
 LOOP-PRO
TUNER
DEMO

Control Station www.controlstation.com

 Recommendations

• Follow the 6-Step Tuning Recipe

Choose Design Level of Operation
Bump the Process
Fit a Model
Use Tuning Correlations
Test Results
Document the Process
• Move output enough to push past valve stiction
• Level Controllers can be easily tuned in manual or automatic mode
• Adjusting the Inner PID Performance impacts Outer PID Performance,
make sure to repeat the outer bump after adjusting inner loop settings.

Control Station www.controlstation.com