Instrumentation

Lecture 8 Paper

Quiz ChatGPT conversation

🎯 1. Big Picture: What is PID Tuning?

PID controllers are the most widely used controllers in industry (over 50%!)

👉 Goal: Make a system behave the way you want by minimizing error between:

Setpoint (target)
Process value (actual output)

💡 Controller tuning = choosing the right parameters so the system:

responds fast ⚡
is stable 🧘
doesn’t overshoot too much 🎯

🧠 2. The PID Equation (Core Idea)

The controller combines three actions:

P (Proportional) → reacts to current error
I (Integral) → reacts to accumulated error over time
D (Derivative) → predicts future error

📌 Variables (from page 2):

( u(t) ): control signal
( ε(t) ): error
( K_c ): gain
( τ_i ): integral scaling
( τ_d ): derivative scaling
( b ): bias (offset)

⚖️ 3. Understanding Gain (SUPER IMPORTANT)

🔽 Too low gain:

System responds slowly
Oscillations die out (damped)

👉 Seen in the graph on page 2:

oscillations shrink over time → stable but sluggish

🔼 Too high gain:

System becomes unstable
Oscillations grow larger and larger

👉 Seen in the graph on page 3:

oscillations increase → system “blows up”

🎯 Perfect tuning:

Oscillations are steady and consistent

This gives: 👉 Ultimate gain (Ku) 👉 Ultimate period (Pu)

These are key for tuning methods.

🔄 4. Open vs Closed Loop Systems

🔁 Closed-loop (feedback):

Output is compared to input
Used in most real systems

🔓 Open-loop:

No feedback
Just observe system response

📌 Important difference:

Ziegler-Nichols → works for both
Cohen-Coon → mainly open-loop

🧪 5. Trial-and-Error Method (Simple but crude)

💡 Idea:

“Try values → see what happens → adjust”

Steps:

Start with P control
Increase (K_c)
Add I and D to refine

📊 Typical values (VERY exam-relevant)

🌊 Flow:

Use P or PI
Avoid D (too noisy!)
(K_c = 0.4–0.65), (T_i = 6s)

🛢️ Level:

PI preferred (reduces offset)
Example:
- (K_c = 2)
- Bias = 50%
PI:
- (K_c = 2–20)
- (T_i = 1–5) min

💨 Pressure:

Depends on medium:

Type	Kc	Ti
Liquid	0.5–2	6–15 s
Gas	2–10	2–10 min

🌡️ Temperature:

Use full PID (slow system!)
(K_c = 2–10), (T_i = 2–10) min, (T_d = 0–5) min

📈 6. Process Reaction Curve Method (Open-loop)

This is VERY important conceptually

🧪 What you do:

Let system reach steady state
Apply a step disturbance
Record response

📊 What you measure:

From curve (page 6 diagram):

⏱️ Dead time (τ_dead) → delay before system reacts
⏳ Time constant (τ) → how fast it responds
📈 Final value (Mu) → steady-state value
🔁 Step size (Xo)

👉 The tangent line at inflection point (shown in diagram page 6) is used to estimate these.

📌 Key ratios:

( R = τ_ / τ )
Used to calculate controller parameters

🎯 Goal:

Convert these values into:

(K_c), (T_i), (T_d)

🔥 7. Ziegler–Nichols Method (MOST IMPORTANT)

Two versions:

🔁 A. Closed-loop Method

🧪 Steps:

Turn OFF I and D
Increase (K_c)
Wait until oscillations are:
- constant amplitude

👉 This gives:

Ultimate gain (Ku)
Ultimate period (Pu)

📊 Visual (page 7):

Increasing gain → oscillations stabilize
Measure oscillation period

🧮 Then:

Plug into formulas:

Controller	Kc	Ti	Td
P	Ku/2	—	—
PI	Ku/2.2	Pu/1.2	—
PID	Ku/1.7	Pu/2	Pu/8

✅ Advantages:

Simple
Uses real system dynamics

❌ Disadvantages:

Can push system into instability
Time-consuming

🔓 B. Open-loop Method (Process Reaction)

🧪 Steps:

Apply step input
Measure:
- (τ_), (τ), (Mu)
Use formulas to compute PID parameters

✅ Advantages:

Faster
Safer (no instability testing)

❌ Disadvantages:

Less accurate
Based on approximations

📉 8. Quarter Decay Ratio (QDR)

Used mainly in open-loop tuning

🎯 Definition:

Amplitude ratio = 4:1

👉 First peak is 4× larger than second peak

💡 Meaning:

System oscillations decay nicely
Good balance between:
- speed ⚡
- stability 🧘

⚖️ 9. Cohen–Coon Method (Alternative)

Mentioned but not deeply detailed here.

📌 Key idea:

Another open-loop tuning method
Uses process reaction curve
Often more accurate than Z-N for some systems

🧠 Final Mental Model (IMPORTANT)

Think of PID tuning like this:

Component	Role
P	Reacts NOW
I	Fixes past errors
D	Predicts future

Method	When to use
Trial & Error	Quick/simple systems
Ziegler-Nichols	No model, need fast tuning
Process Reaction	Safer open-loop systems
Cohen-Coon	More refined open-loop

🚀 Key Takeaways

PID tuning = balancing speed vs stability
Too low gain → slow 🐢
Too high gain → unstable 💥
Ziegler-Nichols is the most important classical method
Process reaction curve is essential for open-loop understanding
Real systems often require experimental tuning

Quiz

Score: 0/44 (0%)

Q0. In PID tuning, what is the main purpose of adjusting controller parameters such as Kc, Ti, and Td?

To eliminate the need for feedback in all control systems

To optimize process behavior and minimize error between the process variable and the set point

To make the controller operate only in open-loop mode

To remove disturbances from the physical system entirely

Q1. Why are classical tuning methods such as Ziegler-Nichols and Cohen-Coon useful when tuning controllers?

They require a complete mathematical model of the system

They are only used for digital controllers

They can be used when a mathematical model of the system is unavailable

They eliminate the need to measure system response

Q2. In the PID equation, what does the error term represent?

The integral of the control signal

The difference between the current process value and the set point

The gain required for oscillation

The steady-state value of the process reaction curve

Q3. What is the role of Kc in the PID algorithm?

It scales the proportional action of the controller

It represents the dead time of the process

It measures the ultimate period of oscillation

It determines the size of the step disturbance

Q4. What happens when the controller gain is too low in a closed-loop system?

Oscillations grow larger with time

The system becomes immediately unstable

The output response is damped and eventually reaches equilibrium

The derivative action becomes dominant

Q5. What happens when controller gain is too high?

The system responds more slowly but remains stable

Oscillations become unstable and grow larger over time

Integral action is automatically removed

The process becomes open-loop

Q6. What is the ultimate gain, Ku?

The gain value that makes the system respond without any overshoot

The experimentally obtained controller gain that gives stable, consistent oscillations

The gain used only in open-loop systems

The gain that eliminates the proportional controller

Q7. What is the ultimate period, Pu, in the Ziegler-Nichols closed-loop method?

The time delay before the system first responds to a disturbance

The time required for one full oscillation at sustained oscillation

The time needed for the system to reach the set point exactly

The time constant of the open-loop process reaction curve

Q8. Which tuning method is based mainly on guess-and-check adjustments?

Cohen-Coon method

Ziegler-Nichols closed-loop method

Trial and error method

Process reaction curve method

Q9. In trial-and-error tuning, which control action is typically treated as the main control action first?

Derivative action

Integral action

Proportional action

Feedforward action

Q10. Why is derivative control usually avoided in flow control applications?

Flow systems never require feedback

Flow dynamics often fluctuate rapidly and contain significant noise

Derivative action cannot be used with proportional action

Derivative control only works in temperature systems

Q11. Why is PI control commonly preferred over P-only control for level control?

PI control removes all dead time from the system

P-only control may produce offset, while integral action helps reduce steady-state error

Derivative action is required for level control

P-only control cannot move a control valve

Q12. Why are PID controllers often used for temperature control?

Temperature sensors usually respond slowly to dynamic temperature changes

Temperature systems are always open-loop

Temperature control does not require integral action

Temperature systems cannot oscillate

Q13. In the process reaction curve method, what is introduced after the system reaches steady state?

A mathematical controller model

A step disturbance

A derivative-only controller

A new feedback sensor

Q14. What does transportation lag or dead time describe in a process reaction curve?

The time required to complete one oscillation

The delay between introducing a disturbance and observing the first output response

The time needed for the controller to calculate Kc

The ratio of the first overshoot to the second overshoot

Q15. What is Mu in the process reaction curve method?

The ultimate gain of the controller

The maximum derivative time

The value the response approaches as the system returns to steady state

The period of sustained oscillation

Q16. How are tau_dead and tau commonly estimated from a process reaction curve?

By setting derivative action to zero and measuring oscillation period

By drawing a tangent line at the point of inflection of the response curve

By measuring the controller bias directly

By applying Cohen-Coon equations without a step test

Q17. What does the quarter decay ratio describe?

A first overshoot amplitude that is four times the second overshoot amplitude

A controller gain that is one quarter of the ultimate gain

A dead time that is four times the process time constant

A derivative time that is one quarter of the integral time

Q18. In the Ziegler-Nichols closed-loop tuning method, why are integral and derivative actions removed first?

To make the system open-loop

To isolate the effect of proportional gain while determining Ku

To eliminate all oscillations permanently

To make Cohen-Coon equations applicable

Q19. Which pair of experimentally determined values is used in the Ziegler-Nichols closed-loop tuning method?

Xo and Mu

tau_dead and R

Ku and Pu

Bias and set point

Q20. What is one major disadvantage of the Ziegler-Nichols closed-loop method?

It cannot be used in systems with feedback

It requires the system to be tested near unstable conditions

It never includes real process dynamics

It only works for derivative-only controllers

Q21. Why can the Ziegler-Nichols closed-loop method provide a useful picture of system behavior?

It ignores process dynamics completely

It includes the dynamics of the whole process during testing

It only measures the controller output mathematically

It avoids using real experimental data

Q22. What is the Ziegler-Nichols open-loop method also called?

Ultimate gain method

Process reaction method

Derivative isolation method

Bias correction method

Q23. What must be measured during a Ziegler-Nichols open-loop step test?

Only Ku and Pu

Dead time, response time or time constant, final response value, and step size

Only controller bias and set point

Only derivative gain and proportional offset

Q24. Which statement best describes the Cohen-Coon method as presented in the document?

It is typically used for open-loop systems as an alternative classical tuning method

It is only used for closed-loop systems with no disturbance

It does not use process response information

It replaces PID control with manual control

Q25. True or false: PID controller tuning is only necessary when a complete mathematical model of the system is available.

True

False

Q26. True or false: In a closed-loop system, feedback is used to compare output with input or set point.

True

False

Q27. True or false: In an open-loop system, the output is compared to the input using feedback.

True

False

Q28. True or false: The Ziegler-Nichols method can be used for both open-loop and closed-loop systems.

True

False

Q29. True or false: Cohen-Coon is typically used for closed-loop tuning only.

True

False

Q30. True or false: If controller gain is too low, oscillations may be damped and eventually settle.

True

False

Q31. True or false: If controller gain is too high, oscillations can become unstable and increase in amplitude over time.

True

False

Q32. True or false: The ultimate gain Ku is found by decreasing Kc until all oscillations disappear.

True

False

Q33. True or false: In trial-and-error tuning, integral and derivative actions are usually used first, while proportional gain is adjusted last.

True

False

Q34. True or false: Derivative control is usually avoided in noisy flow applications.

True

False

Q35. True or false: Temperature control often uses PID because temperature sensors and thermal systems can respond relatively slowly.

True

False

Q36. True or false: The process reaction curve method begins by allowing the system to reach steady state before applying a disturbance.

True

False

Q37. True or false: Dead time is the time between introducing a disturbance and the first sign of change in the output signal.

True

False

Q38. True or false: In the process reaction curve method, tau_dead and tau are estimated using a tangent line at the curve's inflection point.

True

False

Q39. True or false: The quarter decay ratio means that the first overshoot and second overshoot have equal amplitude.

True

False

Q40. True or false: In the Ziegler-Nichols closed-loop method, integral time may be set very high and derivative time set to zero to remove I and D effects.

True

False

Q41. True or false: The Ziegler-Nichols closed-loop method uses Ku and Pu to calculate Kc, Ti, and Td.

True

False

Q42. True or false: A disadvantage of Ziegler-Nichols open-loop tuning is that approximated Kc, Ti, and Td values may not be accurate for all systems.

True

False

Q43. True or false: The Ziegler-Nichols open-loop method is described as quick, robust, popular, and less disruptive than the closed-loop approach.

True

False