BP EYRC#9 Task1 Analysis

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Think and Answer

e-YRC#9
Team leader name AKSHAY S RAO
College B.M.S. COLLEGE OF ENGINEERING
Email akshayrao.ec17@bmsce.ac.in
Date 26 – 11 - 2019

Please answer all the questions given below. You are allowed to use figures or diagrams
to support your answer. Since these questions test your understanding of the whole
subject, please refrain from directly asking for answers on Piazza.

Section 1 - Simple Pendulum


Q1) Find the eigenvalues of Simple Pendulum at equilibrium point (0,0). Is the system
stable or unstable at this point? (2)
To find Eigen values we have to find |𝐴 − 𝜆𝐼| and solve for 𝜆.
0 1
Matrix A at (0, 0) is given by, 𝐴=[ ]
−𝑔/𝑙 0
−𝜆 1
|𝐴 − 𝜆𝐼| = | [ 𝑔
− −𝜆] |
𝑙
𝑔 𝑔
→ 𝜆2 = → 𝜆 = ±𝑖√
𝑙 𝑙
Code - [eigen_value] = eig(A);
The system is stable at point (0, 0).
Q2) Can the Pendulum be balanced at an arbitrary point such as (2π/3, 0) using the Pole
Placement or LQR controller? Why? Why Not? Justify your answer. (3)
We try to make a closed loop controller with overall eigen values on the left half of the
complex plane irrespective of the eigen values of the open loop system, hence the
system will be stable for any input hence the Pendulum can be balanced at an arbitrary
point using both Pole Placement and LQR controller.
Proof :
Controllability is the ability to drive a state from any initial value to a final value in
finite amount of time by providing a suitable input. If we have a system which is
controllable, then we can place its eigenvalues anywhere in the left half plane by
choosing appropriate gain matrix K.
For simple pendulum,
0 1 0
𝐴 = [ 𝑔 0] 𝐵 = [ 1 ] 𝑅 = [ 𝐵 𝐴𝐵 ]
𝑙 2 𝑚𝑙
1
𝐴𝐵 = [𝑚𝑙2 ]
0
1
0 𝑚𝑙2
So 𝑅 = [ 1 ]
0
(𝑚𝑙2 )

Since the rank of the matrix is two and so is the number of state variables, the system
is controllable and therefore the system can be driven to any arbitrary point.

Section 2 - Mass Spring System


Q3) Derive the equations of Mass Spring system. (3)
Let 𝑚 be the mass of the block, 𝑘 be the spring constant, 𝑥 and 𝑥̇ be the
displacement and velocity respectively.
We know that Langrangian L = K.E – P.E
1
K.E = 𝑚𝑥̇ 2
2
1
P.E = 𝑘𝑥̇ 2
2
1 1
L= 𝑚𝑥̇ 2 - 𝑘𝑥̇ 2
2 2
𝑑 𝜕𝐿 𝜕𝐿
𝑇ℎ𝑒 𝐸𝑢𝑙𝑒𝑟 − 𝐿𝑎𝑔𝑟𝑎𝑛𝑔𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 ( )− = 𝐹 − − (1)
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
𝑑 𝜕𝐿 𝑑
( ) = (𝑚𝑥̇ ) = 𝑚𝑥̈ − − (2)
𝑑𝑡 𝜕𝑥̇ 𝑑𝑡

𝜕𝐿
= −𝑘𝑥 − − (3)
𝜕𝑥

𝑓𝑟𝑜𝑚 2 & 3 (1) → 𝑚𝑥̈ + 𝑘𝑥 = 𝐹


𝑘 𝐹
𝑥̈ = − 𝑥+
𝑚 𝑚

So the states are given by, 𝑥1 = 𝑥 𝑥2 = 𝑥̇


𝑘 𝐹
Derivative of state variables are given by, 𝑥̇ 1 = 𝑥2 𝑥̇ 2 = 𝑥̈ = − 𝑚 𝑥1 + 𝑚

Q4) Is the mass spring system a linear system or non-linear? Justify your answer. (1)
It is linear system since it can be represented in the standard form ẋ=Ax+Bu where x is
the state variable and u is the input. The system is linear for all positions of x and no
small regions is taken for the analysis. This is a linear system because all of its terms
are single, first-degree variables with constant coefficients.
Q5) Can the mass spring system be driven to arbitrary state (0.8, 0) using pole
placement controller? (Assuming 0.8 is the position and 0 is the velocity). (1)
Yes, pole placement controller can be used to drive the mass spring system to an
arbitrary state since the state equations is linear for all values of the state variables. We
try to make a closed loop controller with overall eigenvalues on the left half of the
complex plane irrespective of the eigenvalues of the open loop system.
Proof:
For simple pendulum,
0 1 0
𝐴 = [−𝑘 ] 𝐵 = [ 1] 𝑅 = [ 𝐵 𝐴𝐵 ]
0 𝑚
𝑚
1
𝐴𝐵 = [𝑚]
0
1
0 𝑚
So, 𝑅 = [ 1 ]
0
𝑚

Since the rank of the matrix is two and so is the number of state variables, the system
is controllable and therefore the system can be driven to any arbitrary point.
Section 3 - Simple Pulley
Q6) Under what conditions, will the system remain perfectly at rest? Justify your
answer. (1)
For input = 0 and 𝑚1 = 𝑚2 there will be no net acceleration, so the system will be
at rest.
Q7) How many equilibrium points does the system have? Are they stable or unstable?
Justify your answer. (2)
For the simple pulley to be at equilibrium, T = 0
𝑇
(𝑚1 −𝑚2 )𝑔+
𝑟
𝑥̇ 1 = 𝑥2 = 0 & 𝑥̇ 2 = = 0 → 𝑚1 = 𝑚2
𝑚1 +𝑚2

Only condition for equilibrium is 𝑚1 = 𝑚2 which is irrespective of the position of the


mass.
Its stable equilibrium,
0 1
𝑝𝑟𝑜𝑜𝑓 → 𝐴 𝑚𝑎𝑡𝑟𝑖𝑥 = [ ]
0 0
Its eigenvalues are ±0 which is stable.

Section 4 - Complex Pulley


Q8) Derive the equations of motion for the complex pulley system. (5)
Let 𝑣1 be the velocity of mass 𝑚1
𝑣2 be the velocity of mass 𝑚1
𝑣3 be the velocity of mass 𝑚3
Let 𝑥1 = 𝑥 , 𝑥2 = 𝑥̇ , 𝑥3 = 𝑦, 𝑥4 = 𝑦̇
→ 𝑥̇ 1 = 𝑥2 , 𝑥̇ 3 = 𝑥4
Assuming 𝑚1 < 𝑚2 + 𝑚3 and 𝑚3 > 𝑚2
𝑣1 = −𝑥̇ ; 𝑣2 = 𝑥̇ − 𝑦̇ ; 𝑣3 = 𝑥̇ + 𝑦̇
1 1 1
K.E = 𝑚1 𝑥̇ 2 + 𝑚2 (𝑥̇ − 𝑦̇ )2 + 𝑚3 (𝑥̇ + 𝑦̇ )2
2 2 2

P.E = −𝑚1 𝑔𝑥 − 𝑚2 𝑔(𝐿𝐴 − 𝑥 + 𝑦) − 𝑚3 𝑔(𝐿𝐴 − 𝑥 + 𝐿𝐵 − 𝑦)


We know that Lagrangian, L = K.E – P.E.
1 1 1
𝐿= 𝑚1 𝑥̇ 2 + 𝑚2 (𝑥̇ − 𝑦̇ )2 + 𝑚3 (𝑥̇ + 𝑦̇ )2 + 𝑚1 𝑔𝑥 + 𝑚2 𝑔(𝐿𝐴 − 𝑥 + 𝑦)
2 2 2
+ 𝑚3 𝑔(𝐿𝐴 − 𝑥 + 𝐿𝐵 − 𝑦)

The 𝐸𝑢𝑙𝑒𝑟 − 𝐿𝑎𝑔𝑟𝑎𝑛𝑔𝑒 equations for this system is given by,


𝑑 𝜕𝐿 𝜕𝐿 𝑇 𝑑 𝜕𝐿 𝜕𝐿 𝑇
( ) − 𝜕x = 𝑟𝐴 − − (1) ( ) − 𝜕𝑦 = 𝑟𝐵 − − (2)
𝑑𝑡 𝜕ẋ 𝐴 𝑑𝑡 𝜕𝑦̇ 𝐵

𝜕𝐿
( ) = 𝑚1 𝑥̇ + 𝑚2 (𝑥̇ − 𝑦̇ ) + 𝑚3 (𝑥̇ + 𝑦̇ ) = (𝑚1 + 𝑚2 + 𝑚3 )𝑥̇ + (−𝑚2 + 𝑚3 )𝑦̇
𝜕ẋ
𝑑 𝜕𝐿
( ) = (𝑚1 + 𝑚2 + 𝑚3 )𝑥̈ + (−𝑚2 + 𝑚3 )𝑦̈ − − (3)
𝑑𝑡 𝜕ẋ
𝜕𝐿
= 𝑚1 𝑔 − 𝑚 2 𝑔 − 𝑚3 𝑔 − − (4)
𝜕x
𝜕𝐿
= −𝑚2 (𝑥̇ − 𝑦̇ ) + 𝑚3 (𝑥̇ + 𝑦̇ ) = (−𝑚2 + 𝑚3 )𝑥̇ + (𝑚2 + 𝑚3 )𝑦̇
𝜕ẏ
𝑑 𝜕𝐿
( ) = (−𝑚2 + 𝑚3 )𝑥̈ + (𝑚2 + 𝑚3 )𝑦̈ − − (5)
𝑑𝑡 𝜕𝑦̇
𝜕𝐿
= 𝑚2 𝑔 − 𝑚 3 𝑔 − −(6)
𝜕𝑦
(3), (4) 𝑖𝑛 (1) & (5), (6) 𝑖𝑛 (2) →
𝑇𝐴
(𝑚1 + 𝑚2 + 𝑚3 )𝑥̈ + (−𝑚2 + 𝑚3 )𝑦̈ = (𝑚1 −𝑚2 −𝑚3 )𝑔 + − − (7)
𝑟𝐴

𝑇𝐵
(−𝑚2 + 𝑚3 )𝑥̈ + (𝑚2 + 𝑚3 )𝑦̈ = (𝑚2 −𝑚3 )𝑔 + − − (8)
𝑟𝐵
Solving for 𝑥̈ & 𝑦̈ we get,

𝑇 𝑇
( 𝐴 (𝑚2 + 𝑚3 ) − 𝐵 (−𝑚2 + 𝑚3 ))
(𝑚1 𝑚2 + 𝑚1 𝑚3 − 4𝑚2 𝑚3 )𝑔 𝑟𝐴 𝑟𝐵
𝑥̇ 2 = 𝑥̈ = + − − (9)
𝑚1 𝑚2 + 𝑚1 𝑚3 + 4𝑚2 𝑚3 (𝑚1 𝑚2 + 𝑚1 𝑚3 + 4𝑚2 𝑚3 )

𝑇𝐴 𝑇
(2𝑚1 𝑚3 − 2𝑚1 𝑚2 )𝑔 + (−𝑚2 + 𝑚3 ) − 𝐵 (𝑚1 + 𝑚2 + 𝑚3 )
𝑟𝐴 𝑟𝐵
𝑥̇ 4 = 𝑦̈ =
(𝑚1 𝑚2 + 𝑚1 𝑚3 + 4𝑚2 𝑚3 )
Q9) Derive the A and B matrices for the complex pulley system. Is the system linear or
non linear? (4)
The system is non-linear since it doesn’t obey superposition principle (it involves some
constants which affects linearity).
From the equations derived above 𝑥̇ 1 = 𝑥2 , 𝑥̇ 3 = 𝑥4 , 𝑥̇ 4 & 𝑥̇ 2 𝑟𝑒𝑓𝑒𝑟 (9) 𝑎𝑏𝑜𝑣𝑒
𝜕𝑥̇ 1 𝜕𝑥̇ 1 𝜕𝑥̇ 1 𝜕𝑥̇ 1
𝜕𝑥1 𝜕𝑥2 𝜕𝑥3 𝜕𝑥4
𝜕𝑥̇ 2 𝜕𝑥̇ 2 𝜕𝑥̇ 2 𝜕𝑥̇ 2 0 1 0 0
𝐴=
𝜕𝑥1 𝜕𝑥2 𝜕𝑥3 𝜕𝑥4 𝐴= [
0 0 0 0
]
𝜕𝑥̇ 3 𝜕𝑥̇ 3 𝜕𝑥̇ 3 𝜕𝑥̇ 3 0 0 0 1
0 0 0 0
𝜕𝑥1 𝜕𝑥2 𝜕𝑥3 𝜕𝑥4
𝜕𝑥̇ 4 𝜕𝑥̇ 4 𝜕𝑥̇ 4 𝜕𝑥̇ 4
[𝜕𝑥1 𝜕𝑥2 𝜕𝑥3 𝜕𝑥4 ]

 . .

  x1  x1 
 u1 u2 
 . .   0 0 
  x2  x2   m2  m3 m2  m3 
 u u2  1  r rB 
B=  .1 .  → 𝐵 = 𝑚1 𝑚2 +𝑚1 𝑚3 +4𝑚2 𝑚3
 A

  x3  0 0 
 x3 
   m2  m3 m1  m2  m3 
 u1 u2   rA rB 
 . .

  x4  x4 
 u1 u2 

Q10) Under what conditions, will the system remain perfectly at rest? Justify your
answer. (3)

When 𝑚1 = 2𝑚2 = 2𝑚3

For rest condition,


𝑥̇ , 𝑦̇, 𝑥̈ , 𝑦̈ , 𝑇𝐴 , 𝑇𝐵 = 0
When this happens,
(𝑚1 𝑚2 + 𝑚1 𝑚3 − 4𝑚2 𝑚3 ) = 0 − −(1)
(2𝑚1 𝑚3 − 2𝑚1 𝑚2 ) = 0 − −(2)
𝑓𝑟𝑜𝑚 (2) 𝑚2 = 𝑚3 − − (3)
(2) 𝑖𝑛 1 → 𝑚1 = 2𝑚2
Section 5 - Inverted Cart Pendulum
Q11) Derive the equations of motion for the inverted cart pendulum system. Is this
system linear or non-linear? Why? (7)
This system is non-linear since the equations representing the dynamics of the system
doesn’t obey superposition principle, also from the equation 7 and 8 derived below we
can see derivative of the state variables are sinusoidal functions of ɵ , we can’t
represent these type of equations in 𝐴 = 𝐵𝑥 form .
The states are given by
𝑥1 = 𝑥 𝑥2 = 𝑥̇ 𝑥3 = ɵ 𝑥4 = ɵ̇
Where 𝑥 is position of the cart.
𝑥̇ is velocity of the cart.
ɵ is angular displacement of the pendulum from vertical.
ɵ̇ is the tangential angular velocity of the pendulum.

𝑥̇ 1
𝑥̇
Let [ 2 ] be the derivative of state variable vector, where 𝑥̇ 1 = 𝑥2 𝑥̇ 3 = 𝑥4
𝑥̇ 3
𝑥̇ 4
We know that Lagrangian L = K.E – P.E.
Let M - mass of the cart, m – mass of the pendulum, l = length of the pendulum.
1 1 1 1
Kinetic energy of the system = 𝑀𝑥̇ 2 + 2 𝑚ɵ̇ 2 = 𝑀𝑥̇ 2 + 𝑚(ɵ̇ 𝑥 2 + ɵ̇ 𝑦 2 )
2 2 2
1 1 1
.= 𝑀𝑥̇ 2 + 𝑚(𝑥̇ − 𝑙ɵ̇ cos ɵ)2 + 𝑚(−𝑙ɵ̇ sin ɵ)2
2 2 2

1 1
= (𝑚 + 𝑀)𝑥̇ 2 + 𝑚(−2𝑥̇ 𝑙 ɵ̇ 𝑐𝑜𝑠ɵ + l2 ɵ̇ 2 )
2 2

Potential energy of the system = −𝑚𝑔𝑙 cos ɵ


Hence the Lagrangian is given by
1 1
𝐿 = 2 (𝑚 + 𝑀)𝑥̇ 2 − 𝑚𝑥̇ 𝑙 ɵ̇ 𝑐𝑜𝑠ɵ + 2 l2 ɵ̇ 2 𝑚 + 𝑚𝑔𝑙 cos ɵ --- (1)
Equations that represent the dynamics of the system are given by
𝑑 𝜕𝐿 𝜕𝐿 𝑑 𝜕𝐿 𝜕𝐿
( )−
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
= 𝐹 − − (2) ( )−
𝑑𝑡 𝜕ɵ̇ 𝜕ɵ
=0 − − (3)

(2) −> (𝑚 + 𝑀)𝑥̈ − 𝑚𝑙ɵ̈ 𝑐𝑜𝑠ɵ + lɵ̇ 2 𝑚𝑠𝑖𝑛ɵ = 𝐹 − − (4)

(3) −> −𝑚𝑥̈ cos ɵ + 𝑚𝑙ɵ̈ + 𝑚𝑔 sin ɵ = 0 − − (5)

−𝑔 sin ɵ+ 𝑥̈ cos ɵ
𝑓𝑟𝑜𝑚 (5) ɵ̈ = − − (6)
𝑙

𝐹− 𝑚𝑔 sin ɵ cos ɵ − 𝑚𝑙ɵ̇ 2 𝑠𝑖𝑛ɵ


(6) 𝑖𝑛 (4) → 𝑥̇ 2 = 𝑥̈ = − − (7)
M+m(1−cos2 ɵ)

𝑚𝑙ɵ̇ 2 sin 2ɵ
−(𝑚+𝑀)𝑔 sin ɵ + 𝐹 cos ɵ −
2
(4) 𝑖𝑛 (6) → 𝑥̇ 4 = ɵ̈ = − − (8)
L(M+m(1−cos2 ɵ))

Q12) How many equilibrium points does the inverted cart pendulum system have?
Categorize them as stable or unstable? (3)
There are two equilibrium points 𝑖. (0 0 0 0) 𝑖𝑖. (0 0 π 0).
For the first point matrix 𝐴 − 𝜆𝐼 is given by,
−𝜆 1 0 0
g
0 −𝜆 −m ∗ 0
𝐴 − 𝜆𝐼 = M
0 0 −𝜆 1
g
[ 0 0 −(m + M) −𝜆]
LM
Its eigenvalues are,
0
(𝑀 + 𝑚)𝑔
−𝑖 √
𝐿𝑀

(𝑀 + 𝑚)𝑔
𝑖√
[ 𝐿𝑀 ]
For the second point matrix 𝐴 − 𝜆𝐼 is given by,
−𝜆 1 0 0
g
0 −𝜆 −m ∗ 0
𝐴 − 𝜆𝐼 = M
0 0 −𝜆 1
g
[ 0 0 (m + M) −𝜆]
LM
Its eigenvalues are,
0
(𝑀 + 𝑚)𝑔
−√
𝐿𝑀

(𝑀 + 𝑚)𝑔

[ 𝐿𝑀 ]

As shown above point (0 0 0 0) is stable equilibrium since its eigenvalues are on


LHS of the S-plane, and (0 0 π 0) is an unstable equilibrium point since one its
eigenvalues is on the RHS of the S-plane an also after a small change in the parameters
around that point, the pendulum will not return to that point again unless there is an
external force.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy