Good day!
Today, I'm going to rant
a brief on a favourite
(read: "favourite")
subject of mine for this semester.
ENT220/4 Linear Control Systems.
I've posted on Instagram
about how cool the lecturer is.
I've posted on Facebook
about the syllabus.
Now, allow me to elaborate about the subject
a lil bit more.
Beware!
It's only suitable for people
with mild engineering knowledge.
So before we go further
This subject, simply put
is all about
CONTROL.
lol, yeah.
Exactly what the lecturer said
at the first lecture
as in
"If you want to be queen/king of control,
you must learn this subject" impression.
Next, he came with this.
"If you put your grandma into the input,
you'll get your grandma from the output."
That is, by having control of a particular system
(or plant, as specified by throughout the course)
You shall get the output
exactly as it is desired
(or in this case, as it is inserted at the input).
It's clear anyway
that the control we speak of here
is the control in the context of engineering field.
The first chapter's all about the basic terms
systems, sub-systems, plant, control, open & closed loop.
and some revisions on Laplace Transforms (LTs)
so it's nothing much in particular.
Boleh google sendiri kot istilah kat atas tu ea? Hihi
Afterwards, using the reaffirmed knowledge of LTs
we learnt to deduce transfer functions
from block diagrams
and signal flow graphs
also, by using the above
we model a few systems
deducing transfer functions outta them
we model a few systems
deducing transfer functions outta them
be it mechanical, electrial & biomedical
 |
| Simple electrical system => RLC circuit |
 |
| Mechanical system consisting of mass, spring, gears & damper |
 |
| Biomedical system => ECG machine |
Going forward
we're being familiarized to analyze a system
in time domain & frequency domain.
A pinch of terms 101,
system is something that produces output
upon an input is fed into it
while time domain & frequency domains
are just like s-domain in LTs.
might as well say it's similar to currencies used in banking
so time & frequency domains be like American & Australian Dollar
s-domain's like Britain Pound Sterling. Lol.
In usual practices, it is just fine to do analysis
in these two domains.
Time domain analysis (TDA)
is practical for most types of inputs
(step, ramp, parabolic)
while frequency domain analysis (FDA)
is practical when the input is sinusoidal.
We learnt a few more new terms & specifications
unique to TDA & FDA respectively.
For TDA
we were taught to perform it in two situations:
when the system is giving usual response (transient)
and when it is stable (steady-state)
also, new terms
rise & settling time, time delay, maximum time, maximum overshoot
and how damping factor affects a system response
again, we compute the system response
using LTs
in cases of simple, multiple & complex poles
(partial fraction thingy, if anyone recalls)
also, we consider the steady-state error
that difference between the output
and reference input at steady-state
considering the system isn't producing the exact result.
As for FDA
those new terms be like
resonant peak & frequency, bandwidth & cut-off rate.
Else, the focus is more on doing Bode plot & Nyquist plot
so as to estimate better the frequency response
as well as determining system stability
and converting it later for estimation in time response.
Kalau dah datang US kenalah pakai US$ kan,
takkan nak guna MYR pulak. Hew
Last but not least
what's the point of having the control
if you're clueless about the system's stability?
Realizing that (kononnya),
we were exposed to a few techniques
to examine the stability.
namely Routh's Table
and root locus
As a side note,
Routh-Hurwitz's criterion
is always the major consideration here.
After knowing all these,
we are acquired
to design controllers
be it PID or lead-lag.
while PID or Proportional Integral Derivative type
is a combination of PI & PD types,
one features a good steady-state error & cut-off rate improvement
while another features higher stability & bandwidth improvement
thus PID is a compensation for both types
as lead-lag
is merely another name for PI-PD combo.
In addition, Ziegler-Nichols First & Second Tuning Method
are also applied
in order to find parameters for said controller types
Kp, Ti, Td
assuming the plant's on transient response.
Well, that sums up the subject that consumes
partly the past 14 weeks of four-hour lectures.
Sorry for the extremely brief story, I guess.
I'm in the attempt of
lessening the gravity of the matter
while keeping the interest intact in between
nanti ada yang pening. Siapa tau
Kepada sesiapa yang belajar benda ni untuk sem ni
terutamanya geng satu kelas
semoga berjaya!
untuk subjek ni dan subjek-subjek yang sewaktu dengannya.
Best wishes to everyone else whom taking finals too.
Assalamualaikum!
P/S: Using sources from Wiki since I can't link sources from my varsity portal.
For TDA
we were taught to perform it in two situations:
when the system is giving usual response (transient)
and when it is stable (steady-state)
also, new terms
rise & settling time, time delay, maximum time, maximum overshoot
and how damping factor affects a system response
again, we compute the system response
using LTs
in cases of simple, multiple & complex poles
(partial fraction thingy, if anyone recalls)
also, we consider the steady-state error
that difference between the output
and reference input at steady-state
considering the system isn't producing the exact result.
As for FDA
those new terms be like
resonant peak & frequency, bandwidth & cut-off rate.
Else, the focus is more on doing Bode plot & Nyquist plot
 |
| Bode plot in MATLAB environment. |
 |
| An example of Nyquist plot |
as well as determining system stability
and converting it later for estimation in time response.
Kalau dah datang US kenalah pakai US$ kan,
takkan nak guna MYR pulak. Hew
Last but not least
what's the point of having the control
if you're clueless about the system's stability?
Realizing that (kononnya),
we were exposed to a few techniques
to examine the stability.
namely Routh's Table
 |
| one of those examples of root locus techniques |
Routh-Hurwitz's criterion
is always the major consideration here.
After knowing all these,
we are acquired
to design controllers
be it PID or lead-lag.
while PID or Proportional Integral Derivative type
is a combination of PI & PD types,
one features a good steady-state error & cut-off rate improvement
while another features higher stability & bandwidth improvement
thus PID is a compensation for both types
as lead-lag
is merely another name for PI-PD combo.
In addition, Ziegler-Nichols First & Second Tuning Method
are also applied
in order to find parameters for said controller types
Kp, Ti, Td
assuming the plant's on transient response.
Well, that sums up the subject that consumes
partly the past 14 weeks of four-hour lectures.
Sorry for the extremely brief story, I guess.
I'm in the attempt of
lessening the gravity of the matter
while keeping the interest intact in between
nanti ada yang pening. Siapa tau
Kepada sesiapa yang belajar benda ni untuk sem ni
terutamanya geng satu kelas
semoga berjaya!
untuk subjek ni dan subjek-subjek yang sewaktu dengannya.
Best wishes to everyone else whom taking finals too.
Assalamualaikum!
P/S: Using sources from Wiki since I can't link sources from my varsity portal.
