Colorado School of Mines CHEN403 1st Order Systems

Linear Open Loop Systems

Linear Open Loop Systems ................................................................................................................................... 1

1st Order Systems .................................................................................................................................................... 1
Example 1st Order Systems — Mercury Thermometer ..................................................................... 2
Example 1st Order Systems — Mass Storage in Tank ......................................................................... 2
Response of 1st Order Systems .......................................................................................................................... 3
Determination of Coefficients from Data ....................................................................................................... 7
Linearization .............................................................................................................................................................. 9

1st Order Systems

Output modeled with a 1st order ODE:

dy
a1  a0 y  b  f t  .
dt

If a0  0 , then:

a1 dy b dy
 y  f t   p  y  K p  f t 
a0 dt a0 dt

where:  p is the time constant.

K p is the steady state gain, static gain, or gain.

For deviation variables, where y  0   f  0   0 , the Laplace transform will be:

ys Kp
  s  1 y  s   K f  s 
p p  Gs 
f s

p s  1

This transfer function is referred to as 1st order lag, linear lag, or exponential transfer lag.

But what if a0  0 ? Then:

dy dy b dy
a1  bf  t    f t    K p  f t  .
dt dt a1 dt

For deviation variables, the Laplace transform will be:

John Jechura (jjechura@mines.edu) -1- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

ys K p
sy  s   K p  f  s   G  s    .
f s s

This transfer function is referred to as purely capacitive or pure integrator.

Kp K p
f s ys f s ys
p s  1 s
1st Order lag Pure Integrator

Example 1st Order Systems — Mercury Thermometer

Last time we developed the following equation for the reading from a mercury
thermometer:

mCˆ p dT mCˆ p dT
 Ta  T   T  Ta
hA dt hA dt

So this is a 1st order lag system with:

mCˆ p
p 
hA
Kp  1

Example 1st Order Systems — Mass Storage in Tank

Mass storage in a tank is a 1st order system, but we don’t know which type until we say
something about how the flow out of the tank is controlled.

F0, 0

h1, A1
F1, 1

For constant density & constant cross-sectional area:

d  1 Ah dh
 0F0  1F1  A  F0  F1 .
dt dt
John Jechura (jjechura@mines.edu) -2- © Copyright 2017
April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

For flow through a valve where we can linearize to F1  Cv h  h / R , then:

dh h dh
A  F0   AR  h  RF0 .
dt R dt

So this is a 1st order lag system with:

p  AR
Kp  R

However, if the flowrate out is controlled separately from the level in the tank, e.g., with a
pump, then:

dh dh F0  F1
A  F0  F1   .
dt dt A

So this is pure integrator system with:

f  t   F0  F1
1
K p  .
A

Response of 1st Order Systems

Look at response to 4 typical driving functions.

Impulse disturbance. f  t       0  f  s    . So, if 1st order lag:

K p
Kp p Kp
ys  Gs f s   y t   
 t / p
 e
p s  1 1 p
s
p

If pure integrator:

K p'
ys  Gs f s    y t   K p'
s

John Jechura (jjechura@mines.edu) -3- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

Unit step change. f  t     H  t   f  s    / s . So, if 1st order lag:

Kp  1 p 
ys  Gs f s   K p  
p s  1 s  s  s  1 
 p 
 
 
1 1
 K p   
s 1 
 s 
 p 

 y t   K p 1  e
 t / p
.
Notice that K p is the fraction of the value of the input disturbance that will show up on the
output signal. Also notice that the slope is:

dy K p t / p dy K p
 e  
dt p dt t 0 p

If the system would maintain its initial rate of change, then it would achieve it ultimate
value in one time constant, i.e., when t  p . In reality, the final value is reached in an
exponential decay manner — in reality, it takes about 4 p to reach the ultimate value
(when y  0.98K p ).

John Jechura (jjechura@mines.edu) -4- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

1.0

0.9

0.8

0.7

0.6

y/Kp 0.5

0.4

0.3

0.2

0.1

0.0
0 1 2 3 4 5 6 7
t/p

If pure integrator:

K p  K p
ys  Gs f s   2  y t   K p t
s s s

This shows the integrating nature of this type of 1st order process.

Ramp. f  t   mt  f  s   m / s 2 . So, if 1st order lag:

m Kp  1 p 2p 
ys  Gs f s   mK p  2   
p s  1 s 2 s s p s  1 


 y t   mK p t  p  pe
 t / p
.
At large times, then:

y  t   mK p  t  p 

Again, K p is the fraction of the value of the input disturbance that will show up on the
output signal. Now  p represents a time offset — how far behind the output signal lags
behind the input signal.

John Jechura (jjechura@mines.edu) -5- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

Sinusoidal Response. f  t    sin t  f  s    /( s 2  2 ) . So, if 1st order lag:

 K p  2p
Kp 1  p s 
ys  Gs f s     
p s  1 s 2  2 1  2p 2  p s  1 s 2  2 
K p  t / p 1 
y t   2 2 p
e  sin t   p cos t 
1  p    
p K p t / p K p p K p
y t   e  sin t  cos t
1  p
2 2
1  p
2 2
1  2p2

There is a trigonometric identity:

p cos   q sin   r sin     

where: r  p2  q2
p
tan  
q

So for this problem:

2 2
  p K p   K p  K p 1  2p 2 K p
r      
1    1   
2 2 2 2
1  p
2 2
1  2p 2
 p   p 
 p K p

1  2p 2
tan     p    tan 1   p 
K p
1  2p 2

and:

p K p K p
y t   sin t  tan 1  p  
 t / p
e 
1  2
p
2
1  
2
p
2

At large times, then:

K p
y t   sin t  tan 1  p  
1   2
p
2

John Jechura (jjechura@mines.edu) -6- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

Again, K p represents part of the fraction of the value of the input disturbance that will
show up on the output signal.  p also plays are a part in the gain, but, more importantly,
represents a time offset that manifests itself as a phase angle lag (lag because the angel is
subtracted from the input signal). At most, even with very large  p values, the output can
lag by no more than 90°.

Determination of Coefficients from Data

How could we determine the parameters in the transfer function for a 1st order process?
We can perturb the process in a controlled manner & look at the transient results.

Impulse disturbance. f  t       0  gives:

Kp
y t   
 t / p
e
p

Can put into “linear form” by taking the logarithm:

 K p   1 
ln  y   ln   t
     
 p   p

Using linear regression of ln  y  vs. t data, the slope will be 1/ p and the intercept will
be ln  K p / p  .

Unit step change. f  t     H  t  gives:


y t   K p 1  e
 t / p

with the initial slope of:

dy K p

dt t 0 p

One would get the gain from the ultimate value, K p  y /  ; however, this requires that
you take data long enough to be confident of the ultimate value, y . You could then get the
initial slope & determine the time constant:

John Jechura (jjechura@mines.edu) -7- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

K p
p 
 dy / dt  t 0

However, then numerical determination of the derivative from data is inherently difficult.
You could also put this equation into “linear form:”

 y  1
ln  1      t
 K p   
  p

and linear regression can again be used. The difficulty now is that the form of the
independent variable, ln 1  y / K p  , includes one of the variables to be determined, K p .
There are two ways this could be handled:

1. Adjust the value of K p until the intercept from the linear regression is zero.
2. Do the regression by forcing the intercept to be zero & adjust K p to maximize the
regression coefficient, r 2 .

The following table shows the response of a tank when the flow rate is increased from 2
ft³/min to 3 ft³/min. Using deviation variables, the value of  will be   3  2  1 ft³/min .

min ft min ft
0 0.0 10 1.8
1 0.2 20 3.2
2 0.4 30 4.3
3 0.6 40 5.2
4 0.8 50 5.5
5 0.9 60 6.0
6 1.1 70 6.7
7 1.4 80 7.0
8 1.5 90 7.1
9 1.6

Doing linear regression with the linear form and allowing an intercept, a value of
K p  7.92 min/ft² gives an intercept of zero; the corresponding slope gives p  39.0 min .
Doing linear regression with the linear form and not allowing an intercept, a value of
K p  8.05 min/ft² gives an intercept of zero; the corresponding slope gives p  41.0 min .
Note that the values were generated with random perturbations added to a linear model
with K p  8 min/ft² and p  40 min . The following chart shows the

John Jechura (jjechura@mines.edu) -8- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

8.0

7.0

6.0

5.0

h (ft) 4.0

3.0

2.0

1.0

0.0
0 10 20 30 40 50 60 70 80 90
t (min)

Linearization
Not all dynamic systems are described by linear ODEs — in truth, probably none of the
systems are really linear. To use transfer functions & Laplace transforms, however, we
must linearize the system of equations. We have already looked at an example of this with
flow through a valve. If the flow through a valve is considered to be proportional to h ,
then the material balance around a tank containing a fluid of constant density is described
by:

dV
 F0  F1  F0  Cv h
dt

and for a constant cross-sectional area in the tank, A , then:

dh
A  F0  Cv h .
dt

We’ve discussed ways to linearize the square root term in this ODE. We can effectively do a
Taylor series expansion by first taking the total differential of the right-hand-side:

   C 
d F0  Cv h  dF0   v  dh
2 h 

John Jechura (jjechura@mines.edu) -9- © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

Now, replace the differentials with deviation variables & evaluate any coefficients at the
initial steady state. So:

 
 F  C h   F    2C h
0 v 0
v
*
 h


and the linearized ODE will be:

dh  C 
A  F0   v  h
dt  2 h 
dh  Cv 
A   h  F0
dt  2 h* 
 2A h*  dh  2 h* 
   h    F0 .
 Cv  dt  Cv 
   

From this, we see that the linearized ODE is simply a 1st order ODE with the following gain
and time constant:

2 h*
Kp 
Cv
2 A h*
p 
Cv

Effectively, we have found that the gain and time constant are not constant, but rather
functions of the liquid level. Whether or not the linearized equation is a good
representation depends upon how far we perturb the system from its steady state value.
For example, given a tank with a cross-sectional area of 5 m² which maintains a level of 16
m when the flow in is 2 m³/min, what happens when we shut off the flow? In the full non-
linear solution:

dh C
 F0  v h
dt A

where:

h  0   h*  16 m
F0*  2 m³/min
F0  t   0

John Jechura (jjechura@mines.edu) - 10 - © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

then:

h 2
dh C dh Cv t  C 
h
  v dt 
A 
hss h
   dt  h   h*  v t 
A0  2A 

The linearized equation is:

dh Cv  h  h 
*
dh C
A   F0*  A  v h  F0*
dt 2 h * dt 2 h*
 Cv  F0*
 As   h  
 2 h*  s
K p F0* 2 h* 2 A h*
h   where K p  and p 
p s  1 s Cv Cv
2F0* h*   
So:  t / 
h  F0*K p 1  e p   h  h* 
Cv
 1  exp  
tCv
 2A h
*
 



The chart below shows the difference in the drainage curve for the two equations. Notice
that the linearized equation stays fairly close to the exact solution for the 1st 8 ft change of
level. Also note that even though the ODE is linearized, it does not predict a straight line
answer — there is still curvature to the final h  t  result.

John Jechura (jjechura@mines.edu) - 11 - © Copyright 2017

April 23, 2017
Colorado School of Mines CHEN403 1st Order Systems

18

16

14

12
Exact Solution
Level (m)

10
Linearized Equation

0
0 10 20 30 40 50 60 70 80 90
Time (min)

John Jechura (jjechura@mines.edu) - 12 - © Copyright 2017

April 23, 2017