Tutorial

Twelve
Systems Dynamics &
Control Components
Analytical Solution of Linear Models

• Previously we have discussed how to obtain the model of systems

• We will discusses how to solve the differential equations analytically

• Starting from the input-output equation

an y (n ) + ... + a2 y + a1 y + a0 y = bmu (m ) + ... + b1u + b0u (t )

• Let F (t ) = bmu (m ) + ... + b1u + b0u (t )

• Therefore, the input-output equation is expressed as

an y (n ) + ... + a2 y + a1 y + a0 y = F (t )

2
Analytical solution of equations

• The solution of the differential equation y(t) can be expressed as

• y(t) = yH(t) + yP(t)

• where

• yH(t): Solution to the Homogeneous differential equation

(Complementary Solution)
• yP(t): Solution to the Non-homogeneous differential equation
(Particular Solution)
Homogeneous Solution
• Assume F(t) =0 → Homogeneous part (Complementary solution)

• Find n roots of c/c’s equation

∴ yH(t) =K1𝑒 𝑟1𝑡 +K2𝑒 𝑟2𝑡 +…..+Kn 𝑒 𝑟𝑛𝑡 where k: constant, r1:first root

Note:
Check special cases in lecture “repeated roots & complex roots”
Non -Homogeneous Solution

• Fa(t) ≠ 0 → Non-Homogeneous part (Particular solution)

F(t) yp(t)
α A
αt+ß At+B
𝑒 αt A 𝑒 αt
cos(wt) or sin(wt) A cos(wt) +B sin(wt)
Example: find & sketch the response of the system
𝑦+4y
ሶ =F(t) ,in 2 cases: a) F(t) = 3 & y(0) = 0
b) F(t) = t & y(0) = 0

• In Both cases:
• Get The homogeneous part “set F(t)=0”:
• 𝑦+4y
ሶ =0
• ry+4y=0 Non-trivial
solution
• y(r+4)=0
• r+4=0 → c/c’s eq. # ∴ r = -4
• y (t) =K𝑒 −4t
H
• Case a:
• F(t) =3 ∴yP(t) = A
• ∴ y(t) = K𝑒 −4t + A
• Differentiate to get 𝑦(t):
ሶ
ሶ = -4K𝑒 −4t
• 𝑦(t)
• Substitute in differential equation:
• 𝑦+4y
ሶ = -4K𝑒 −4t + 4K𝑒 −4t + 4A = 3 ∴A=¾
• ∴ y(t) = K𝑒 −4t +
3
4
• Using initial conditions:
• y(0)= K𝑒 −4(0) 3
+ =K+ = 0
3
4 4
∴ K= -3/4

e-4t
−3 −4t 3
∴ y(t) = 𝑒 + # is a decaying
4 4 exponential
y(t)
3
y(t) =
4

−3 −4t 3
y(t) = 4
𝑒 +4

t
−3 −4t
y(t) = 𝑒
4
• Case b:
• F(t) =t ∴ yP(t) = At + B
• y(t) = K𝑒 −4t + At + B
• Differentiate to get 𝑦(t):
ሶ
ሶ = -4K𝑒 −4t + A
• 𝑦(t)
• Substitute in original differential equation:
• 𝑦+4y
ሶ = -4K𝑒 −4t + A + 4K𝑒 −4t + 4At + 4B = t
 1
• Coefficient of t => A=
4
1
• Coefficient of => A+ 4B = 0 ∴ B= -
t0
16
• ∴ y(t) = K𝑒 −4t 1
+ t-
1
4 16
• Using initial conditions:
• y(0)= K𝑒 −4(0) 1 1
+ (0) − = K−
1
=0
4 16 16
1
∴ K=
16
1 −4𝑡 1 1
∴ y(t) = 𝑒 + t- →#
16 4 16
 1 −4𝑡 1 1
y(t) = 𝑒 + t-
16 4 16
1
y(t) = t
4
1 −4𝑡
y(t) = 𝑒
16

1
y(t) = -
16
find & sketch the response of the system
𝑦+7
ሷ 𝑦+6y=3,
ሶ y(0) = 0, 𝑦(0)=
ሶ -2
• Get The homogeneous part “set F(t)=0”:
• 𝑟 2 +7r+6=0
• (r+6)(r+1)=0 → c/c’s eq.
• r1= -1 ,r2= -6 → c/c’s roots
• y (t) =K 𝑒 −t +K 𝑒 −6t
H 1 2
• Get The particular part:
• F(t) = 3 ∴yp(t) = A
• ∴ y(t) =K1𝑒 −t+K2 𝑒 −6t+A
• Differentiate to get 𝑦(t)ሶ & 𝑦 ̈(t):
• 𝑦ሶ (t) = -K 𝑒 −t -6 K 𝑒 −6t
1 2
• 𝑦(t)=
ሷ K1𝑒 −t +36 K2 𝑒 −6t
• Substitute in original differential equation:
• K1𝑒 −t +36 K2 𝑒 −6t -7K1𝑒 −t - 42K2 𝑒 −6t + 6 K1 𝑒 −t + 6K2 𝑒 −6t +
6A=3
• ∴ A= ½
• ∴ y(t) =K 𝑒 −t+K 𝑒 −6t + 0.5
1 2
• Using initial conditions:
• y (t=0) =K1+K2 +1/2 = 0 → (1)
• 𝑦ሶ (t =0) = -K1 -6K2 = -2 → (2)

• By solving (1) & (2)

• K1 = -1 , K2 =0.5

∴ y(t)= -𝑒 −t +0.5𝑒 −6t +0.5 → #

 1
y(t) =
2

y(t) = 0.5𝑒 −6𝑡 y(t)= -𝑒 −t +0.5𝑒 −6t +0.5

y(t) = -𝑒 −𝑡
 A

B
Solution:

• Fn1(e0) = Fn2(i(t))
𝑑𝑒𝑐 𝑑𝑖
• Remember that: ic= c , eL= L
𝑑𝑡 𝑑𝑡
• KCL @ A
𝑒0−𝑒𝑐 𝑒0
• i(t) = + = 2e0-ec → 1
1 1
• KCL @ B
• e0 - ec= 0.25 𝑒ሶ c → 2
• From 1
• ec= 2e0-i(t) → sub in 2
𝑑
• e0 - 2e0 + i(t) = 0.25 (2e0 - i(t))
𝑑𝑡
𝑑𝑖(𝑡)
• - e0 +0.5 𝑒ሶ 0 = i t + 0.25
𝑑𝑡
• Subs. with i(t) = 1A, e0(0) = 0
• e0(t) =?
• Let e0=y
• y+ 0.5*10−6 𝑦ሶ =1
• 0.5*10−6 r +1=0
• r = - 2*106
• yH(t) =K 𝑒 − 2∗106 𝑡

• yp(t) =A
• y(t)= K 𝑒 − 2∗106 𝑡
+A
• 𝑦ሶ (t) = −2 ∗ 10 K 𝑒
6 − 2∗106 𝑡
• Sub in diff. eq.
•K𝑒 − 2∗106 𝑡
+A +0.5 ∗ 10 (-2 ∗ 10 K 𝑒
−6 6 − 2∗106 𝑡
)=1
• ∴ A=1
• e(0) =y(0)=0
• y(0) =K 𝑒 − 2∗106 (0)
+ A =0 → ∴ k= -1

• e0(t) = 1 - 𝑒 − 2∗106 𝑡
#
First order systems
1
• 𝑦ሶ + y=F(t) → standard form
𝜏

• yH(t) =K

• Note:
• Stable system if yH(t → ∞) = 0
• 𝜏 → time constant, y(𝜏) = 0.63 ∗ ys. s
• y(t) = yH(t) +yp(t)
• If stable system yp= ys.s → steady state response
• Proof:
• y(t)= yH(t) +ys.s
• − 𝑡
=K 𝑒 𝜏 +y s.s
• y(0)= K +ys.s
• ∴ K = y(0) - ys.s
−𝑡
• ∴ y(t) = ys.s + ( y(0) - ys.s) 𝑒 𝜏
[transient response]
Solution:

• Put in standard form:

2 𝐴
• 𝑦ሶ + y =
3 3
3
• ∴ 𝜏=
2

• yH(t) = K
• t → ∞ , yH(t → ∞ ) = 0 ∴ system is stable
 𝐴
• F(t) = ∴ ys.s = constant
3
3
• ∵ y(0) = 0 , 𝜏 =
2
• ∴ y(t) = ys.s - ys.s −2
𝑒 3𝑡
• Differentiate to get 𝑦(t):
ሶ
• 𝑦(𝑡)
ሶ
2
= y 𝑒 3𝑡 − 2

3 s.s
• Substitute in original differential equation:
2 − 2
− 2
• ∴ 3( y 𝑒 3 ) + 2 (y – y 𝑒 3 𝑡 ) = A
𝑡
3 s.s s.s s.s
𝐴
ys.s =
2

∴ y(t) = ys.s - ys.s (1- ) #

For sketching let A = 6

• ∴ ys.s = 3 y(t) = 3

• 𝜏= 1.5 s −2
y(t) = 3 - 3 𝑒 3 𝑡

• y(𝜏) = 0.63 * 3
= 1.8
−23 𝑡
y(t) = - 3 𝑒
Solution
− 𝑡
• ∵ yH(t) =K 𝑒 𝜏 = K 𝑒 rt
1 1
• ∴ r =− or 𝜏 = −
𝜏 𝑟
1
• 𝜏A = = 5 sec. => figure 3
0.2
1
• 𝜏B = = 1 sec. => figure 1
1
1
• 𝜏C = = 2 sec. => figure 2
0.5

Notes:
• for all @(𝜏), the response reaches 63% of steady state “ 0.63*1 = 0.63”
• ∵ unit response , ∴ F(t) = 1
Second Order Response
• 𝑦ሷ + 𝑎1𝑦ሶ + 𝑎0𝑦= F(t)
• 𝑦ሷ + 2δωn𝑦ሶ + ωn2 𝑦= F(t)
• where ωn is the natural frequency and δ is the damping ratio
• Roots of c/c’s equation are given by

• V.imp: Check 2nd order stability conditions in slides

properties

• δ = 0 → un-damped system roots complex with real part =0

“±2j”
• δ2 < 1 → under damped system complex roots “2±2j”
• δ2 = 1 → critically damped system roots are real (repeated)
“2,2”
• δ2 > 1 → over damped system roots real (distinct) “2,5”
 Un-damped

Under damped Min zeta max

oscillation
amplitude

critically damped
Min wn max
oscillation
Over damped speed
time
Solution

A ) − δωn ± 𝜔𝑛 δ2 − 1 = - 0.5 ± j 2
− δωn = - 0.5 → 1
𝜔𝑛2 (1 − δ2) = 4 → 2
Sub in 2 with 𝜔𝑛 = 0.5/δ
0.25
δ 2 (1-δ2) = 4 → δ = 0.2425
ωn = 0.5/δ =2.0616

Do the same with B,C,D

A δ =0.2425 𝜔𝑛 =2.0616 Figure 3

B δ =0.124 𝜔𝑛 =4.0311 Figure 4

C δ =0.2425 𝜔𝑛 =4.1231 Figure 1

D δ =0.4472 𝜔𝑛 =4.4721 Figure 2