Scribd
Upload
46 views9 pages

CHE3007s - Process Control - Lecture 6 - Linearization

Uploaded by

khmern002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
46 views9 pages

CHE3007s - Process Control - Lecture 6 - Linearization

Uploaded by

khmern002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

CHE3007S Block3

Lecture 6
Linearization of Non-linear Systems
Overview
• What is a non-linear system?
• How can we linearize a non-linear system?
Additional Reading / Learning
Sections of textbook by Seborg et al. to study:
• Chapter 3 – 3rd Edition
• Chapter 4 – 4th Edition

Glossary:
• Linearization, Taylor Series, Non-linear systems
What is a non-linear system?
A differential equation is considered linear if:

𝑑𝑛 𝑦 𝑑2𝑦 𝑑𝑦
𝑎𝑛 𝑡 + ⋯ + 𝑎2 𝑡 + 𝑎1 𝑡 + 𝑎0 (𝑡)𝑦 = 𝑓(𝑥)
𝑑𝑡 𝑛 𝑑𝑥 2 𝑑𝑡

• 𝑎𝑛 𝑡 is a pure function of t
• not dependent on y

𝑑𝑛 𝑦
• is only of power 1
𝑑𝑡 𝑛
• Remember 𝑦 is the 0th derivative of 𝑦
What is a non-linear system?
Is it linear?
𝑑2 𝑦 𝑑𝑦
• +5 + 4𝑦 = 0
𝑑𝑡 2 𝑑𝑡

𝑑2 𝑦 𝑑𝑦
• + 5 𝑑𝑡 + 4𝑦 2 = 0 (𝑦 2 is not of power 1)
𝑑𝑡 2

𝑑2 𝑦 𝑑𝑦
• sin(𝑡) 𝑑𝑡 2 + 𝑒 −5𝑡 𝑑𝑡 + 𝑦 = 0

𝑑2 𝑦 𝑑𝑦
• sin(𝑡) 𝑑𝑡 2 + 𝑒 −5𝑦𝑡 𝑑𝑡 + 𝑦 = 0 (𝑒 −5𝑦𝑡 is not a pure function of t)

𝑑2 𝑦 𝑑𝑦 𝑑2 𝑦
• + 5 𝑑𝑡 + 4𝑦 = 0 ( is not of power 1)
𝑑𝑡 2 𝑑𝑡 2
What is a non-linear system?
This also extends to multiple variables. Is it linear?
𝑑2 𝑦 𝑑𝑥
• +5 + 4𝑦 = 0
𝑑𝑡 2 𝑑𝑡

𝑑2 𝑦 𝑑𝑥
• + 5 𝑑𝑡 + 4𝑥𝑦 = 0 (𝑥𝑦 is not linear)
𝑑𝑡 2

𝑑2 𝑦 𝑑𝑥
• sin(𝑡) 𝑑𝑡 2 + 𝑒 −5𝑡 𝑑𝑡 + 𝑥 = 0

𝑑2 𝑦 𝑑𝑦
• sin(𝑡) 𝑑𝑡 2 + 𝑒 −5𝑥𝑡 𝑑𝑡 + 𝑦 = 0 (𝑒 −5𝑥𝑡 is not a pure function of t)

𝑑𝑥 𝑑𝑦 𝑑𝑥
• + 5 𝑑𝑡 + 4𝑦 = 0 ( is not of power 1)
𝑑𝑡 𝑑𝑡
How can we linearize?
• Linearization of non-linear process models provides a linear approximation of the
process model at points closest to the point of linearization (steady state
conditions).

• Taylor series expansion of the first order is used for this purpose:
𝑑𝑦 𝜕𝑓 𝜕𝑓
= 𝑓 𝑥, 𝑦 ≈ 𝑓 𝑥1 , 𝑦1 + ቤ 𝑥 − 𝑥1 + ቤ 𝑦 − 𝑦1 Near 𝑥1 , 𝑦1
𝑑𝑡 𝜕𝑥 𝑥 𝜕𝑦 𝑥
1 ,𝑦1 1 ,𝑦1

In terms of deviation variables from steady state


𝑑𝑦 ′ 𝜕𝑓 𝜕𝑓
= 𝑓 𝑥′, 𝑦′ ≈ ′ ቤ 𝑥′ + ቤ 𝑦′
𝑑𝑡 𝜕𝑥 𝑥ҧ ,𝑦ത 𝜕𝑦′ 𝑥ҧ ,𝑦ത
Buffer tank with valve
Total Mass balance: w in

𝑑ℎ 𝑤𝑖𝑛 − 𝑤𝑜𝑢𝑡
=
𝑑𝑡 𝜌𝐴

𝑑ℎ 𝑞𝑖𝑛 𝐶𝑣 ℎ
= − wout
𝑑𝑡 𝐴 𝐴
V
In terms of deviation variables:
Assumptions:
• Constant Density
𝑑ℎ′ 𝑞′𝑖𝑛 𝐶𝑣 ℎ′
= − • The valve regulates the flow such
𝑑𝑡 𝐴 𝐴
that the outlet volumetric flowrate
is:
ℎ′ = ℎ − ℎത 𝑞′𝑖𝑛 = 𝑞𝑖𝑛 − 𝑞ത𝑖𝑛 𝑞𝑜𝑢𝑡 = 𝐶𝑣 ℎ
Buffer tank with valve
𝑑ℎ′ 𝑞′𝑖𝑛 𝐶𝑣 ℎ′ w in
= − = 𝑓(𝑞 ′ 𝑖𝑛 , ℎ′ )
𝑑𝑡 𝐴 𝐴

With partial derivatives:

𝜕𝑓 1 𝜕𝑓 𝐶𝑣
อ = ቤ = wout
𝜕𝑞 ′ 𝑖𝑛 𝐴 𝜕ℎ′ 𝑞 2𝐴 ℎത
𝑞𝑖𝑛 ,ℎത 𝑖𝑛 ,ℎത
V
The linearized approximation is then: Assumptions:
′
• Constant Density
𝑑ℎ′ 𝑞 𝑖𝑛 𝐶𝑣 ℎ′
≈ − = 𝛼𝑞′ 𝑖𝑛 + 𝛽ℎ′ • The valve regulates the flow such
𝑑𝑡 𝐴 2𝐴 ℎത that the outlet volumetric flowrate
is:
1 𝐶𝑣 𝑞𝑜𝑢𝑡 = 𝐶𝑣 ℎ
𝛼= 𝛽=
𝐴 2𝐴 ℎത

You might also like