A1 FA20 Sol

University of Houston – College of Technology
ELET 4304 Control Systems – FA20
Assignment A1 5 pts Due TH 9/10
Follow Syllabus Rubric for Credit
Submit a single PDF solution file via Blackboard

A series RC circuit is controlled with a constant voltage source 𝑢(𝑡). The differential equation governing the behavior of the
capacitor voltage 𝑣𝑐 (𝑡) is obtained by writing a single KVL equation
𝑑𝑣𝑐 (𝑡) 𝒅𝒗𝒄 (𝒕) 𝟏 𝟏

𝑢(𝑡) = 𝑅𝑖(𝑡) + 𝑣𝑐 (𝑡) = 𝑅𝐶 + 𝑣𝑐 (𝑡) ⟹ + 𝒗𝒄 (𝒕) = 𝒖(𝒕)
𝑑𝑡 𝒅𝒕 𝑹𝑪 𝑹𝑪
Using the 𝑝-operator notation, we can write the Plant Transfer Function
𝑣𝑐 (𝑡) 1/𝑅𝐶 1
= 𝐺𝑝 (𝑝) = = ; 𝑓𝑜𝑟 𝑅 = 1𝑀Ω, 𝐶 = 1𝜇𝐹
𝑢(𝑡) 𝑝 + 1/𝑅𝐶 𝑝 + 1
The RC circuit as a SISO system is placed in a standard negative, unity-feedback control system shown in the block diagram
below using the standard notation:
𝑟(𝑡) = 𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑖𝑛𝑝𝑢𝑡; 𝑐(𝑡) = 𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑙𝑒𝑑 𝑜𝑢𝑡𝑝𝑢𝑡 = 𝑣𝑐 (𝑡); 𝑒(𝑡) = 𝑒𝑟𝑟𝑜𝑟 𝑠𝑖𝑔𝑛𝑎𝑙 = 𝑟(𝑡) − 𝑐(𝑡)
𝑢(𝑡)
The Controller block is in series (cascade) and shown as 𝐺𝑐 (𝑝) = . In this problem we will examine the effect of the simplest
𝑒(𝑡)
controller, a constant gain so that 𝑢(𝑡) = 𝐾𝑒(𝑡), or equivalently 𝐺𝑐 (𝑝) = 𝐾.
𝑟(𝑡) + 𝑒(𝑡) 𝑢(𝑡) 𝑐(𝑡)

𝐺𝑐 (𝑝) 𝐺𝑝 (𝑝)
1. The case of NO FEEDBACK. The feedback is disconnected, hence 𝑒(𝑡) = 𝑟(𝑡), and the applied control signal is simply
𝑢(𝑡) = 𝐾𝑟(𝑡). This is called an OPEN-LOOP System.
a. Find the capacitor voltage 𝑣𝑐 (𝑡), 𝑡 > 0 when the reference is a step input 𝑟(𝑡) = 10𝑢𝑠 (𝑡) and initially the capacitor
voltage is 𝑣𝑐 (0) = 20 𝑉.
b. Use Matlab to plot the capacitor voltage for a range of time of at least 5 time constants.
c. Comment on two performance characteristics: the 5% settling time defined as the time it takes for the output to
be within 5% of its final value; and the steady-state error defined by the final value of 𝑒(𝑡), namely 𝑒𝑠𝑠 = lim 𝑒(𝑡)
𝑡→∞
𝑣𝑐 (𝑡) 1
From the Open-Loop TF: = 𝐺𝑝 (𝑝) = ⇒ Solve (𝑝 + 1)𝑣𝑐 (𝑡) = 𝑢(𝑡) = 10𝐾, 𝑡 > 0; 𝑎𝑛𝑑 𝑤𝑖𝑡ℎ 𝑣𝑐 (0) = 20
𝑢(𝑡) 𝑝+1
Here we leave 𝑲 unspecified for now. This is the controller gain to be designed.
The characteristic root is 𝑠 = −1;
Then, the total solution is: 𝑣𝑐 (𝑡) = 𝐾1 𝑒 −𝑡 + 𝑣𝑐𝑝 (𝑡)
Since the forcing function is just a constant, we know the form of the particular solution is 𝑣𝑐𝑝 (𝑡) = 𝐾2
Plug in the DFQ: (𝑝 + 1)𝐾2 = 10𝐾 ⇒ 0 + 𝐾2 = 10𝐾 ⇒ 𝐾2 = 10𝐾

Then, the total solution becomes: 𝑣𝑐 (𝑡) = 𝐾1 𝑒 −𝑡 + 10𝐾 Use the IC: 𝑣𝑐 (0) = 20 = 𝐾1 + 10𝐾 ⇒ 𝐾1 = 20 − 10𝐾
Then, the total solution is: 𝒗𝒄 (𝒕) = (𝟐𝟎 − 𝟏𝟎𝑲)𝒆−𝒕 + 𝟏𝟎𝑲, 𝒕 > 𝟎
The time constant is 1 sec independent of the controller gain. Note what happens when 𝐾 = 2
The following Matlab routing graphs the capacitor voltage for a couple of values of controller gain 𝐾 = 3, 5
>>T=0:0.01:5;
>> Vc=-10*exp(-T)+30;
>> plot(T,Vc),grid
>> hold
Current plot held
>> Vc=-30*exp(-T)+50;
>> plot(T,Vc)
>>grid
Note that 5 time constants is sufficient to show how 𝑣𝑐 (𝑡) reaches its
steady-state value of 𝑣𝑐 (∞) = 10𝐾
Since this is OPEN-LOOP there is no error signal properly speaking.

One can comment that the reference is 10, and 𝑣𝑐 (𝑡) ends up at 𝑣𝑐 (∞) = 10𝐾, which is equal to the reference only when 𝐾 = 1.
The case of FEEDBACK. The feedback is reconnected, hence 𝑒(𝑡) = 𝑟(𝑡) − 𝑐(𝑡). We will show in class how to find CLOSED-LOOP
transfer functions from a block diagram. Here is suffices to do the following: substitute 𝑢(𝑡) = 𝐾(𝑟(𝑡) − 𝑣𝑐 (𝑡)) into the differential
equation and find the new CLOSED-LOOP differential equation becomes
𝑑𝑣𝑐 (𝑡) 𝒅𝒗𝒄 (𝒕)

+ 𝑣𝑐 (𝑡) = 𝐾(𝑟(𝑡) − 𝑣𝑐 (𝑡)) ⟹ + (𝟏 + 𝑲)𝒗𝒄 (𝒕) = 𝑲𝒓(𝒕)
𝑑𝑡 𝒅𝒕
1
𝑣𝑐 (𝑡) 𝐾 𝐾
𝑝+1
Or using the formula developed in class: = 1 =
𝑟(𝑡) 1+𝐾 𝑝+1+𝐾
𝑝+1
Solve (𝑝 + 1 + 𝐾)𝑣𝑐 (𝑡) = 𝐾𝑟(𝑡) = 10𝐾, 𝑡 > 0; 𝑎𝑛𝑑 𝑤𝑖𝑡ℎ 𝑣𝑐 (0) = 20
Again, we leave 𝑲 unspecified for now. This is the controller gain to be designed.
The characteristic root is 𝑠 = −1 − 𝐾; First thing to notice: the characteristic root in OPEN-LOOP was 𝑠 = −1
Then, the total solution is: 𝑣𝑐 (𝑡) = 𝐾1 𝑒 −(1+𝐾)𝑡 + 𝑣𝑐𝑝 (𝑡) ; The forcing function is just a constant, then 𝑣𝑐𝑝 (𝑡) = 𝐾2
10𝐾
Plug in the DFQ: (𝑝 + 1 + 𝐾)𝐾2 = 10𝐾 ⇒ 0 + (1 + 𝐾)𝐾2 = 10𝐾 ⇒ 𝐾2 =
1+𝐾
10𝐾 10𝐾 10𝐾

Then, the total solution becomes: 𝑣𝑐 (𝑡) = 𝐾1 𝑒 −(1+𝐾)𝑡 + Use the IC: 𝑣𝑐 (0) = 20 = 𝐾1 + ⇒ 𝐾1 = 20 −
1+𝐾 1+𝐾 1+𝐾
𝟏𝟎𝑲 𝟏𝟎𝑲
Then, the total solution is: 𝒗𝒄 (𝒕) = (𝟐𝟎 − ) 𝒆−(𝟏+𝑲)𝒕 + ,𝒕 > 𝟎
𝟏+𝑲 𝟏+𝑲
1
The time constant is 𝑠𝑒𝑐 dependent on the controller gain. Compared to the OPEN-LOOP case we can already see a potential
1+𝐾
improvement in the speed of response because the time constant is smaller.
10𝐾
In terms of the steady-state error performance: in the limit, 𝑣𝑐 (∞) = which approaches the reference value of 10 as the
1+𝐾
controller gain gets larger. However, in practice there are limits as to how high the controller gain can be because of saturation
problems.
The Matlab commands and graph follows.
>>clear; clf
>>T=0:0.01:5; K=3;
>> Vc=(20-10*K/(1+K))*exp(-(1+K)*T)+10*K/(1+K);
>> plot(T,Vc),grid
>> hold
Current plot held
>> T=0:0.01:5; K=5;
>> Vc=(20-10*K/(1+K))*exp(-(1+K)*T)+10*K/(1+K);
>> plot(T,Vc)
>> grid
Note that 5 time constants is sufficient to show how 𝑣𝑐 (𝑡) reaches its
10𝐾
steady-state value of 𝑣𝑐 (∞) = The response is faster for larger
1+𝐾
controller gain 𝐾 as expected because of the smaller time constant.

A1 FA20 Sol

Uploaded by

Copyright:

Available Formats

A1 FA20 Sol

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A1 FA20 Sol

Uploaded by

Copyright:

Available Formats

University of Houston – College of Technology

ELET 4304 Control Systems – FA20

Assignment A1 5 pts Due TH 9/10

Follow Syllabus Rubric for Credit

Submit a single PDF solution file via Blackboard

𝑑𝑣𝑐 (𝑡) 𝒅𝒗𝒄 (𝒕) 𝟏 𝟏

𝑟(𝑡) + 𝑒(𝑡) 𝑢(𝑡) 𝑐(𝑡)

The characteristic root is 𝑠 = −1;

Then, the total solution is: 𝑣𝑐 (𝑡) = 𝐾1 𝑒 −𝑡 + 𝑣𝑐𝑝 (𝑡)

Plug in the DFQ: (𝑝 + 1)𝐾2 = 10𝐾 ⇒ 0 + 𝐾2 = 10𝐾 ⇒ 𝐾2 = 10𝐾

Since this is OPEN-LOOP there is no error signal properly speaking.

𝑑𝑣𝑐 (𝑡) 𝒅𝒗𝒄 (𝒕)

Solve (𝑝 + 1 + 𝐾)𝑣𝑐 (𝑡) = 𝐾𝑟(𝑡) = 10𝐾, 𝑡 > 0; 𝑎𝑛𝑑 𝑤𝑖𝑡ℎ 𝑣𝑐 (0) = 20

10𝐾 10𝐾 10𝐾

You might also like