Module 6: Control Systems

Measurements and Controls of Energy Systems | Dr. Mihir N. Velani | Adani University
End-Term Revision Notes | Academic Use Only

⚡ Topics: Open/Closed Loop · Block Diagrams · Transfer Functions · Test Signals · 1st & 2nd Order Systems
· Poles & Zeros · Stability · Time Domain Specs · Steady-State Error · Routh-Hurwitz Criterion

PART 1: Open-Loop & Closed-Loop Control Systems

1. What is a Control System?
• A system that manages, commands, directs, or regulates the behavior of other devices.
• Goal: Maintain output at desired value despite disturbances.
• Key terms: Input (Setpoint/Reference), Output, Error, Control Action, Disturbance
Error e(t) = Setpoint (SP) − Process Variable (PV)

2. Open-Loop vs Closed-Loop Control System

Feature Open-Loop Closed-Loop

Feedback ❌ No ✅ Yes

Accuracy Low High

Cost Low Higher

Stability Always stable May be unstable if not tuned

Error correction Cannot correct Actively corrects errors

Examples Toaster, Coffee maker, Washing AC, Cruise control, Thermostat

machine (fixed timer)

⚡ Negative feedback preferred: Output is subtracted from input → reduces error → stable, controlled output.

3. Types of Closed-Loop Systems

• Positive Feedback: Feedback added to input — amplifies signal (e.g., microphone squeal,
oscillators). Risky.
• Negative Feedback: Feedback subtracted from input — reduces error (e.g., thermostat, human
body temp). PREFERRED.

4. Block Diagram Elements

• Block: Represents transfer function G(s). Output = Input × G(s)
• Summing Junction: Adds or subtracts signals (marked +/−)
• Branch Point / Take-off Point: Signal splits to multiple destinations
 Transfer Function: G(s) = C(s) / R(s) [Output / Input in Laplace
domain]
⚡ Why Laplace? Converts difficult differential equations into simple algebraic equations!

5. Block Diagram Algebra — 11 Rules

Rule # Operation Result / Formula

1 Blocks in Cascade (series) G_total = G₁ × G₂

2 Parallel blocks G_total = G₁ + G₂

3 Branch point moved RIGHT of block Add 1/G in the branch path

4 Branch point moved LEFT of block Multiply G in branch path

5 Summing point moved RIGHT of Multiply new input by G

block
6 Summing point moved LEFT of Divide new input by G (multiply by 1/G)
block

7 Interchange summing points Order of summing points can be swapped

8 Split summing point Break one summing point into two

9 Combine summing points Merge two summing points into one

10 ★ Eliminate NEGATIVE feedback loop C/R = G / (1 + GH)

11 Eliminate POSITIVE feedback loop C/R = G / (1 − GH)

⚡ ★ Most important: Closed-Loop TF = G(s) / [1 + G(s)·H(s)]. For unity feedback H=1: T(s) = G(s)/[1+G(s)]
PART 2: Time Response Analysis
6. Time Response Overview
• Output C(t) of closed-loop system as function of time.
• In s-domain: C(s)/R(s) = G(s) / [1+G(s)H(s)] = M(s)
Response Type Definition Characteristics

Transient Response Temporary behavior just after Rapid change, oscillations,

input/disturbance is applied energy adjustment

Steady-State Response Behavior after transient dies out — Final settled value, may have
system is stable error from desired

7. Test Signals — Complete Reference

Signal Definition r(t) R(s) [Laplace] Real Example

Step Sudden jump to A, for t≥0 A/s Switch ON heater

constant value

Unit Step Step signal with A=1 1, for t≥0 1/s Turn ON light

Ramp Increases linearly αt, for t≥0 A/s² Climbing a slope

with time

Unit Ramp Ramp with α=1 t, for t≥0 1/s² Gradual fuel increase

Parabolic Increases At²/2, t≥0 A/s³ Car accelerating

quadratically (faster
and faster)

Unit Parabolic with A=1 t²/2 1/s³ Rapidly growing load

Parabolic

Impulse δ(t) Applied δ(t) 1 Lightning strike,

instantaneously at t=0 Voltage spike
only

⚡ Memory trick: Step → Ramp → Parabolic = each is the integral of the previous. Laplace: 1/s → 1/s² → 1/s³

Boiler Scenario → Signal Mapping

Real-World Situation Signal Type to Use

Pump switched ON instantly in a tank Step

Slowly increasing heater power knob Ramp

Load demand increases faster in a power grid Parabolic

Sudden pressure spike in a pipe Impulse

8. Laplace Transform Quick Reference Table

f(t) — Time Domain F(s) — Laplace Domain

δ(t) — Impulse 1

1(t) — Unit Step 1/s

t — Ramp 1/s²

tⁿ n! / sⁿ⁺¹

e⁻ᵃᵗ 1 / (s+a)

te⁻ᵃᵗ 1 / (s+a)²

sin(ωt) ω / (s²+ω²)

cos(ωt) s / (s²+ω²)

e⁻ᵃᵗ sin(ωt) ω / [(s+a)²+ω²]

e⁻ᵃᵗ cos(ωt) (s+a) / [(s+a)²+ω²]

PART 3: System Order, 1st & 2nd Order Systems
9. Order of a System
• Order = highest power of 's' in denominator of transfer function
• = number of energy storage elements in the system
G(s) = 1/(s+2) → 1st Order (one capacitor or inductor)
G(s) = 5/(s²+s+2) → 2nd Order (LC circuit, spring-mass)
G(s) = (s+1)/(s³+2s²+3) → 3rd Order

Converting Time Domain → Transfer Function (Example)

📐 Given: d²y/dt² + 3 dy/dt + 2y = x(t) → Find G(s)
1. Apply Laplace: dy/dt → sY(s), d²y/dt² → s²Y(s), y → Y(s)
2. Substitute: s²Y(s) + 3s·Y(s) + 2Y(s) = X(s)
3. Factor output: Y(s)[s² + 3s + 2] = X(s)
4. Transfer Function: G(s) = Y(s)/X(s) = 1/(s²+3s+2) → 2nd Order
⚡ Assume zero initial conditions — TF represents system behavior due to input only.

10. Standard Form: 1st Order System

• For single energy element systems (RC circuit, water tank):
G(s) = 1 / (τs + 1) where τ = Time constant (RC for electrical
systems)
• Smaller τ → faster response | Larger τ → sluggish, slow response
📐 Find τ for G(s) = 1/(2s+1)
Compare with 1/(τs+1) → τ = 2 seconds (system reaches ~63% of final value in 2s)

11. Standard Form: 2nd Order System

• For two energy storage element systems (RLC circuit, spring-mass-damper):
G(s) = ωₙ² / (s² + 2ζωₙs + ωₙ²)
ωₙ = √(k/m) = undamped natural frequency (rad/s) → how fast system
wants to move
ζ = c/(2√km) = damping ratio → how much
oscillation is reduced

ζ value Type Behavior Practical Use

ζ>1 Overdamped Slow exponential rise, no oscillation Heavy braking systems

— too much damping

ζ=1 Critically Damped Fastest response WITHOUT Precise positioning

oscillation — perfect balance
0<ζ<1★ Underdamped Oscillates, overshoots before settling AC systems, motors
— most common

ζ=0 Undamped Sustained oscillations forever — no Ideal LC circuits

energy loss

ₙFind ωₙand ζ for G(s) = 16/(s²+4s+16)

5. Compare with standard form: ωₙ² = 16 → ωₙ = 4 rad/s
6. 2ζωₙ = 4 → ζ = 4/(2×4) = 0.5 → Underdamped (0 < 0.5 < 1)
PART 4: Poles, Zeros & Stability
12. Poles and Zeros
Concept Definition How to Find Physical Meaning

Poles (×) Values of s making Set D(s) = 0, solve for s Natural behavior — decay,
denominator = 0 growth, oscillation

Zeros (○) Values of s making Set N(s) = 0, solve for s Frequencies system blocks
numerator = 0 or suppresses

Finding Poles and Zeros — Worked Examples

📐 Example A: G(s) = 1/(s²+3s+2)
7. Set denominator = 0: s²+3s+2 = 0 → (s+1)(s+2) = 0
8. Poles: s = -1, s = -2
9. Numerator is constant (=1) → No finite zeros
📐 Example B: G(s) = (s+5)/(s²+6s+5)
10. Zeros: s+5 = 0 → Zero at s = -5
11. Poles: s²+6s+5 = (s+5)(s+1) = 0 → Poles at s = -5, s = -1
⚡ The pole and zero both at s = -5 cancel each other. Effective system has only one pole at s = -1.

13. Stability Classification Based on Poles

Pole Location s-Plane Position Time Response Stability

All poles: negative real Left Half Plane Decays to zero — e⁻ᵃᵗ ✅ STABLE
part (LHP)

Any pole: positive real Right Half Plane Grows to infinity — eᵃᵗ ❌ UNSTABLE
part (RHP)
Poles on imaginary axis Imaginary axis Sustained oscillations ⚠️Marginally Stable
(s = ±jω)

Complex LHP: s = -σ ± Left Half Plane Decaying oscillations ✅ STABLE

jω

Complex RHP: s = +σ ± Right Half Plane Growing oscillations ❌ UNSTABLE

jω

S-Plane: Real axis (σ) → determines growth/decay (Left = dies, Right =

explodes)
Imaginary axis (jω) → determines oscillation (non-zero =
oscillates)

Dominant Poles & Higher Order Systems

• Real systems are often 3rd order or higher, but fast poles die out quickly.
 • The slowest pole (closest to imaginary axis) dominates the response.
G(s) = 1/[(s+1)(s+10)(s+50)] → pole at -1 dominates, behaves like
1st order
📐 Full Stability Analysis: G(s) = (s+3)/(s²+2s+5)
12. Poles: s²+2s+5=0 → s = (-2±√(4-20))/2 = -1 ± j2 (complex conjugate pair)
13. Zero: s+3=0 → Zero at s = -3
14. Stability: Real part of poles = -1 (negative) → STABLE
15. Oscillation: Complex poles → YES, decaying oscillations
16. Speed: Real part = -1 → Moderate decay
17. Zero effect: Zero at -3 → Faster initial rise, slightly more overshoot
PART 5: Time Domain Specifications (2nd Order)
14. Key Time Domain Formulas ★ MEMORIZE THESE ★
ωd = ωₙ√(1-ζ²) [Damped natural frequency]
tp = π/ωd [Peak Time]
Mp = e^(-πζ/√(1-ζ²)) ×100% [Maximum % Overshoot]
ts = 4/(ζωₙ) [Settling Time, 2% criterion]

Specification Symbol Definition

Delay Time td Time for response to reach 50% of final value (first instance)

Rise Time tr Time to rise from 0% to 100% (underdamped) or 10%-90%

(overdamped)

Peak Time ★ tp Time to reach the first peak (maximum value)

Peak Overshoot ★ Mp (%) Normalized difference between peak and final value, as %

Settling Time ★ ts Time to reach and STAY within ±2% to 5% of final value

15. ★ FULL NUMERICAL EXAMPLE — Most Important for Exam

ₙGiven: G(s) = 25/(s²+4s+25). Find: ζ, ωₙ, tp, Mp, ts
18. Compare with ωₙ²/(s²+2ζωₙs+ωₙ²):
ωₙ² = 25 → ωₙ = 5 rad/s
19. Find ζ: 2ζωₙ = 4 → ζ = 4/(2×5) = 4/10 = 0.4
Since 0 < ζ < 1 → System is UNDERDAMPED (will oscillate)
20. Find damped frequency ωd:
ωd = ωₙ√(1-ζ²) = 5×√(1-0.16) = 5×√0.84 = 5×0.9165 = 4.58 rad/s
21. Peak Time tp:
tp = π/ωd = 3.1416/4.58 = 0.685 seconds
22. Peak Overshoot Mp:
Mp = e^(-πζ/√(1-ζ²)) × 100%
= e^(-π×0.4/√(1-0.16)) × 100%
= e^(-1.2566/0.9165) × 100%
= e^(-1.371) × 100% = 0.254 × 100% = 25.4% ≈ 25%
23. Settling Time ts:
ts = 4/(ζωₙ) = 4/(0.4×5) = 4/2 = 2 seconds

Parameter Value Interpretation

ωₙ 5 rad/s Natural oscillation speed

ζ 0.4 Underdamped — will oscillate

ωd 4.58 rad/s Actual oscillation frequency

tp (Peak Time) 0.685 s Time to reach first peak

Mp (Overshoot) 25% Output goes 25% above final value at peak

ts (Settling) 2s Stays within ±2% after 2 seconds

PART 6: Steady-State Error Analysis
16. Why Steady-State Error?
• Even a stable system may not reach the exact desired value.
• Steady-state error = permanent offset remaining after transient dies out.
e(t) = r(t) - y(t) = Setpoint - Output
ess = lim[t→∞] e(t) = lim[s→0] s·E(s) [Final Value Theorem]
For unity feedback: E(s) = R(s) / [1+G(s)]
Therefore: ess = lim[s→0] s·R(s) / [1+G(s)]
⚡ t→∞ in time domain ≡ s→0 in Laplace domain (Final Value Theorem)

17. Error Constants & Steady-State Errors

Input Signal Error Constant How to Calculate Error Formula

Step R(s)=1/s Kp (Position) Kp = lim[s→0] G(s) ess = 1/(1+Kp)

Ramp R(s)=1/s² Kv (Velocity) Kv = lim[s→0] s·G(s) ess = 1/Kv

Parabolic R(s)=1/s³ Ka (Acceleration) Ka = lim[s→0] s²·G(s) ess = 1/Ka

18. ★ Error Calculation Numerical

📐 G(s) = 1/(s+1), Unity Feedback. Find ess for Step, Ramp, and Parabolic inputs.
Step Input R(s) = 1/s:
24. Kp = lim[s→0] G(s) = lim[s→0] 1/(s+1) = 1/1 = 1
25. ess = 1/(1+Kp) = 1/(1+1) = 1/2 = 0.5 → Finite error — system reaches 50% accuracy
Ramp Input R(s) = 1/s²:
26. Kv = lim[s→0] s·G(s) = lim[s→0] s/(s+1) = 0/1 = 0
27. ess = 1/Kv = 1/0 = ∞ → Cannot track ramp input — error grows without bound
Parabolic Input R(s) = 1/s³:
28. Ka = lim[s→0] s²·G(s) = lim[s→0] s²/(s+1) = 0 → ess = ∞

Input Kp / Kv / Ka ess Conclusion

Step Kp = 1 0.5 Finite error — tracks with 50% offset

Ramp Kv = 0 ∞ Cannot track ramp — error unbounded

Parabolic Ka = 0 ∞ Cannot track parabolic at all

⚡ A Type-0 system (no integrators in G(s)) can only track step input with finite error.

19. Effects of Feedback

Effect How Negative Feedback Helps

Reduces gain CL gain G/(1+GH) < OL gain G → controlled, not excessive output

Reduces error Larger G → ess = 1/(1+G) gets smaller → better accuracy

Improves stability Pushes poles toward left-half plane

Rejects disturbances Feedback compensates for external disturbances automatically

Reduces sensitivity System less dependent on exact parameter values of G(s)

Faster response Corrective control action speeds up reaching the setpoint

PART 7: Routh-Hurwitz Stability Criterion
20. What is Routh-Hurwitz?
• Mathematical method to determine stability WITHOUT finding actual pole locations.
• Works by analyzing the characteristic polynomial (denominator = 0).
For: aₙsⁿ + aₙ₋₁sⁿ⁻¹ + ... + a₁s + a₀ = 0
• Construct a Routh Array → Check first column for sign changes.

First Column Condition Stability Verdict

All elements positive — NO sign change ✅ STABLE — All poles in LHP

One or more sign changes ❌ UNSTABLE — # changes = # poles in RHP

Zero in first column (but not all zeros) ⚠️Special case — replace with ε (small +ve
number)

Entire row is zeros ⚠️Marginally stable / repeated roots on imaginary

axis

21. How to Build Routh Array

Row s^n : aₙ aₙ₋₂ aₙ₋₄ ...
Row s^n-1 : aₙ₋₁ aₙ₋₃ aₙ₋₅ ...
Row s^n-2 : b₁ b₂ b₃ ...
where b₁ = (aₙ₋₁·aₙ₋₂ - aₙ·aₙ₋₃) / aₙ₋₁
b₂ = (aₙ₋₁·aₙ₋₄ - aₙ·aₙ₋₅) / aₙ₋₁
⚡ If any element in first column = 0, replace with ε (epsilon → 0⁺) to allow calculation to continue.

22. ★ Routh-Hurwitz Numerical Examples

📐 Example 1: s³ + 2s² + 3s + 4 = 0 (Check Stability)
29. All coefficients: 1, 2, 3, 4 — All present & all positive ✓
30. Build Routh Array:
Power Col 1 Col 2 Calculation

s³ 1 3 Alternating from polynomial

s² 2 4 Alternating from polynomial

s¹ b₁ = (2×3 - 1×4)/2 = (6-4)/2 0 Cross-multiply formula

=1

s⁰ 4 — Last coefficient
31. First column: 1, 2, 1, 4 → All POSITIVE, NO sign change
Result: ✅ STABLE — No roots in Right Half Plane
📐 Example 2: s³ + 2s² − s + 2 = 0 (Negative coefficient!)
32. Coefficients: 1, 2, -1, 2 — Mixed signs detected → Likely unstable, but verify:
Power Col 1 Col 2 Calculation

s³ 1 -1 Alternating from polynomial

s² 2 2 Alternating from polynomial

s¹ (2×(-1) - 1×2)/2 = -4/2 = -2 0 Cross-multiply formula

s⁰ 2 — Last coefficient
33. First column: 1, 2, -2, 2
34. Sign changes: 2→-2 (1st change), -2→2 (2nd change) → 2 sign changes
Result: ❌ UNSTABLE — 2 roots in Right Half Plane

📐 Example 3: s³ + 2s² + s + 2 = 0 (Zero in first column — use ε)

35. Coefficients: 1, 2, 1, 2 — All positive ✓
Power Col 1 Col 2 Calculation

s³ 1 1 Alternating from polynomial

s² 2 2 Alternating from polynomial

s¹ (2×1 - 1×2)/2 = 0 → USE ε 0 Zero! Replace with ε (small positive)

s⁰ 2 — Last coefficient
36. First column: 1, 2, ε, 2 → Since ε > 0, all signs are POSITIVE → No sign change
Result: ✅ STABLE — Poles are on imaginary axis (marginally stable)
⚡ WHY ε? Division by zero is not allowed in the Routh table. ε is a mathematical trick — replace 0 with a tiny
positive number so calculations can continue. It is not a real value.
QUICK REVISION CARD — Last-Minute Review
Essential Formulas to Memorize
Formula Description

T(s) = G(s) / [1 + G(s)H(s)] Closed-loop transfer function (negative

feedback)

G(s) = ωₙ² / (s² + 2ζωₙs + ωₙ²) Standard 2nd order system

ωd = ωₙ√(1-ζ²) Damped natural frequency

tp = π/ωd Peak time

Mp = e^(-πζ/√(1-ζ²)) × 100% Maximum percent overshoot

ts = 4/(ζωₙ) Settling time (2% criterion)

ess = 1/(1+Kp) for step Steady-state error, step input

ess = 1/Kv for ramp Steady-state error, ramp input

ess = 1/Ka for parabolic Steady-state error, parabolic input

Damping Ratio ζ — Quick Decide

ζ Type Response

ζ=0 Undamped Oscillates forever

0<ζ<1 Underdamped Oscillates & settles

ζ=1 Critically damped Fastest, no oscillation

ζ>1 Overdamped Slow, no oscillation

Stability — Quick Decide

All poles in LHP? Stability

Yes — all real parts negative ✅ STABLE

Any real part positive ❌ UNSTABLE

Poles on imaginary axis only ⚠️Marginally Stable

Routh: No sign change in col 1 ✅ STABLE

Routh: N sign changes in col 1 ❌ UNSTABLE (N poles in RHP)

Best of luck for your exam! — All content from Dr. Mihir N. Velani, Module 6, Adani University 2026. For academic use only.