THE COPPERBELT UNIVERSITY

School of mines and minerals sciences

CE 560 / MT 580
Process control

Lecture 18
Dynamic behaviour of first-order systems

C. Botha (Mr.)
Contacts: Chemical Engineering Department
Email: [email protected]
Alt. Email: [email protected]

May, 2023
 First-order systems

o A system whose output 𝑦(𝑡) is modelled by a first-order differential equation is known as

a first-order system.
o Therefore, for a linear (or linearised system), then

dy (t )
a1  a0 y (t )  bf (t )...............................................17.1
dt

o Where 𝑓(𝑡) and 𝑦(𝑡) are the input (forcing function) and output variables of the process.
o If 𝑎0 ≠ 0 then 17.1 can be rearranged as follows:

a1 dy (t ) b
 y (t )  f (t )...............................................17.2
a0 dt a0
 First-order systems
o Then the following can be defined;
a1 b
p  and K  ...............................................17.3
a0 a0
o The differential equation (process model) for a first-order system then becomes
dy (t )
p  y (t )  K p f (t ).........................................17.4
dt
o Here 𝜏𝑝 is process time constant and 𝐾𝑝 is the steady-state or static gain of the process.

o Equation 17.4 is the standard form of the process model for a first-order system.
o The two parameters will be defined in later slides.
o 𝑦(𝑡) and 𝑓(𝑡) are usually in terms of deviation variables defined around steady state.
o Thus the initial conditions are;
y (0)  0 and f (0)  0........................................17.5
 First-order systems
o From equation 17.4, it can be deduced that the transfer function of a first-order system
has a standard form
y( s ) Kp
G ( s)   ..............................................17.6
f ( s )  p s  1

o First-order system with above transfer function is sometimes known as first-order lag or
linear lag or exponential transfer lag.
o On the other hand if 𝑎0 = 0, then
dy (t ) b
 f (t )  K p f (t )......................................17.7
dt a1
o And this results in a transfer function
y ( s ) K p
G ( s)   ................................................17.8
f (s) s
 First-order systems

o The process with a transfer function as in eq. 17.8 is known as pure capacitive or pure
integrator.
o First-order systems are characterised by:
1) The capacity to store material (mass), energy or momentum,
2) The resistance associated with the flow of material, energy or momentum.
o Hence, dynamic response of tanks that possess the capacity to store liquids or gases
can be modelled as first-order systems.
o Resistance to flow in this case is associated with valves, pumps and pipes attached to
either the inflow or outflow.
 First-order systems
o Also, the temperature response of solid, liquid or gaseous systems which have the
capacity to store thermal energy (possess thermal capacity) can be modelled as first
order. E.g. thermocouple/ thermometer
o For such systems, the resistance is associated with the flow of thermal energy
through the walls, liquids or gases.
o Therefore, any system with capacity to store mass or energy can be modelled as first-
order.
o Such system in most cases act as a buffer between the inlet and outlet streams, e.g.,
surge tanks.
o First-order systems are the most common type of dynamic systems in a process plant.
o Since majority of process equipment are mixing tanks, heating tanks, reaction vessels
etc. with capacity to store both mass and energy.
 First-order systems with capacity to store mass: Capacitive

o Two liquid level systems, one with a pump attached to the outlet stream, are considered
here again.
Fi Fi

h h
A A Fo

Tank 1 R Tank 2
Fo

Fig.17.0 Typical first-order system (liquid level)

o For tank 1 in Fig. 17.0, the inlet and outlet volumetric flow-rates (volume/time) are 𝐹𝑖 and
𝐹0 respectively.
o In the outlet stream, there is a resistance to flow such as a valve or pipe.
 First-order systems with capacity to store mass

o Assuming that the outflow is linearly related to the hydrostatic head ℎ of the liquid
through the resistance 𝑅 then
Driving force for flow h
F0   ....................................17.9
Resistance to flow R
o The tank possesses the capacity to store mass, therefore, the total mass balance yields
dh h
A  Fi  Fo  Fi  ........................................17.10
dt R
dh
AR  h  RFi ................................................17.11
dt
o Where 𝐴 is the cross-sectional area.
o For steady-state operation
hs  RFi ,s .....................................................17.12
 First-order systems with capacity to store mass

o Equation 17.11 takes the following form

d (h  hs )
AR  (h  hs )  R( Fi  Fi ,s ).............................17.13
dt
o And defining the deviation variables as follows

h  h  hs and Fi  Fi  Fi ,s ........................................17.14

o Yields
dh
RA  h  RFi..................................................17.15
dt
o The transfer function is therefore,
H ( s ) R Kp
G ( s)    ...................................17.16
Fi( s ) ARs  1  p s  1
 First-order systems with capacity to store mass
o Where  p  AR  process time constant
K p  R  steady state or static gain of the process

o The following can be noted

1) The cross-sectional area 𝐴 measures the capacity of the tank to store mass.
2) Since 𝜏𝑝 = 𝐴𝑅 then for the liquid level system, it can be said that

Time constant  Storage capacity  Resistance to flow.......17.17

First-order systems with capacity to store mass: pure capacitive

o Consider tank 2 shown in Figure 17.0 with the following difference:
o The outlet flow-rate is determined by a pump and not by the hydrostatic head ℎ of the
liquid.
 First-order systems with capacity to store mass: pure capacitive

o For such a system, the total mass balance around the tank is:
dh
A  Fi  Fo ..........................................................17.18
dt
o For steady state operation 0  Fi ,s  Fo ..................................................17.19
o In terms of deviation variables,
d  h  hs 
A  Fi  Fi ,s ...................................................17.20
dt
o And
dh
A  Fi.........................................................17.21
dt
o Which yields the following transfer function
H ( s ) 1 A K p
G ( s)    .....................................17.22
Fi,s ( s ) s s
 First-order systems with capacity to store energy

o In the Figure below, liquid is heated by saturated steam which flows through a coil
immersed in the liquid.

T
Q

Tst
Fig.17.2 Typical first-order system with capacity to store energy

o Total energy balance of the system is

dT
V cp  Q  UAi (Tst  T )...............................................17.23
dt
 First-order systems with capacity to store energy

o Where V  Total volume of the liquid in the tank

 , c p  liquid density and specific heat capacity
U  Overall heat transfer coefficient between the steam and liquid
Ai  Total heat transfer area
Tst  Temperature of the saturated steam
o Considering steady state
0  UAi (Tst ,s  Ts )....................................................17.24

o Equation 17.23 in terms of deviation variables then yields

d (T  Ts )
V cp  UAi (Tst  Tst ,s )  (T  Ts )  ....................17.25
dt
 First-order systems with capacity to store energy

o Or dT 
V cp  UAi Tst  T ..........................................17.26
dt

o Where T   T  Ts and Tst  Tst  Tst ,s ..................................17.27

o Taking the Laplace transform yields the following transfer function

V  c p sT ( s )  UAT
i
( s )  UAT   
i st ( s ) or T ( s ) V  c p s  UAi   UAT
 
i st ( s )...17.27( a )

T ( s ) 1 Kp
G ( s)    ................................17.28
Tst ( s ) V  c p  ps  1
s 1
UAi
o Where
V cp
 p  time constant of the process  and K p  steady-state gain  1
UAi
 First-order systems with capacity to store energy
o Equation 17.28 demonstrate that the stirred tank heater system has first-order dynamics.
o The system possesses capacity to store thermal energy.
o Resistance to the flow of thermal energy is characterised by 𝑈.
o Capacity to store thermal energy is measured by 𝑉𝜌𝐶𝑃 .
o And resistance to flow of thermal energy from steam to the liquid is measured by 1/𝑈𝐴𝑖.
o Hence, the time constant for this system is given by:
 1 
 p  V cp     storage capacity  resistance to flow
 UAi 

o The expression for time constant is thus the same as before and is true for a first-order
system.
 Dynamic response of a pure capacitive system

o Transfer function of a pure capacitive process has been shown to be

y( s ) K p
G ( s)   .....................................17.29
f ( s ) s
o The response 𝑦(𝑡) will be examined when 𝑓(𝑡) undergoes a unit step change.
1
f (t )  1 for t  0 and f ( s)  .....................................17.30
s
o Hence, Equation 17.29 yields
K p
y( s )  2
.................................................17.31
s
o Inverting 17.31 yields
y(t )  K pt.................................................17.32

o The plot of 𝑦′(𝑡) is shown in the figure below

 Dynamic response of a pure capacitive system

o It can be noticed that the output grows

linearly with time.
y(t )
o Thus, 𝑦(𝑡) → ∞ as 𝑡 → ∞.
o The response shown in figure 17.3 is
characteristic of a pure capacitive system. K p
o Response behaves as if there is an t
integrator between the input and output Fig.17.3 Unbounded response of a pure capacitive system

o Integrator evaluates the time integral of the input to obtain an output.

o Hence the name pure integrator.
o Pure capacity system cannot balance itself and causes control problems.
 Dynamic response of a pure capacitive system
o For the tank system shown in Figure 17.2 (tank 2);
o Speed of the constant-displacement pump can be adjusted manually to balance the
incoming flow and thus keep the level constant.
o However, any small change in the flow-rate of the inlet stream will make the tank flood
or run dry.
o Thus constant monitoring and pump speed adjustment to balance inlet flow is required.
o A pure capacitive system is then said to be non-self-regulating.
Dynamic response of a capacitive first-order system
o The standard transfer function for such a first-order system was derived as
y( s ) Kp
 .............................................17.33
f ( s )  p s  1
 Dynamic response of a first-order system

o Response of a first-order system to a unit step change in 𝑓′(𝑡) will now be examined.
1
f (t )  1 for t  0 and f ( s)  .....................................17.34
s
o Then,
Kp
y( s )  ......................................................17.35
s ( p s  1)
o Inverting equation 17.35 results in
t  p
y(t )  K p (1  e ).................................................17.36

o And if the step change in 𝑓′(𝑡) has a magnitude 𝐴 then

t  p
y(t )  AK p (1  e ).........................................17.37

o Fig. 17.4 below shows how 𝑦′(𝑡) changes with time.

 Dynamic response of a first-order system

o Several features of the plot in Fig. 17.4

characterises a first-order system. 1.0

o A capacitive first-order system is self- y(t )

regulating, unlike a pure capacitive, that is, it
AK p
reaches a new steady state.
o In terms of the liquid level system of Fig. 17.0 1 2 3 4

(tank 1), when the inlet flow-rate undergoes a t p

step increase, the liquid level will go up. Fig.17.4 Normalised response of a first-
order system to a step change in input

o Increase in liquid level will increase hydrostatic pressure which in turn will increase the
outlet flow-rate.
o Eventually a point will be reached at which outlet flow will exactly be equal to inlet flow.
 Dynamic response of a first-order system

o Thus, liquid level will stop increasing i.e., a new equilibrium liquid level is attained.
o The slope of the response at 𝑡 = 0 i.e. initial rate of change of response is equal to 1/𝜏𝑝 .
d  y(t ) AK p  1 t  p 1
 e  .........................................17.38
dt p p
t 0

o The smaller the time constant, the faster (steeper) the initial response of the system.
o Therefore, time constant is a measure of how fast a system responds to a change in the input.
o The smaller the time constant the faster the response and vice versa.
o When 𝑡 → ∞, 𝑦′ 𝑡 = 𝐴𝐾𝑝 i.e., the ultimate value of the response is simply 𝐴𝐾𝑝.

o And if 𝑡 = 𝜏𝑝 then the response will have reached 63.2% of its ultimate value since,
 Dynamic response of a first-order system

o As y(t ) t p
1 e p  0.632.........................................17.39
AK p

o Since as 𝑡 → ∞, 𝑦′ 𝑡 = 𝐴𝐾𝑃 , this characteristic renders the name steady state or

static gain given the parameter 𝐾𝑝.
o For any step change ∆(𝑖𝑛𝑝𝑢𝑡), resulting change in output is ∆ 𝑜𝑢𝑡𝑝𝑢𝑡 = 𝐾𝑝 ×
∆ 𝑖𝑛𝑝𝑢𝑡 .
(steady state output)
Kp  .................................(17.40)
o Thus, (steady state intput)

o Therefore, 𝐾𝑝 measures the sensitivity of the system.

o The larger the value of 𝐾𝑝, the larger the steady state response for a unit change in
steady-state input.
 Effect of parameters on response of first-order system

o Consider two tanks with cross-sectional area 𝐴1 and 𝐴2 such that 𝐴1 > 𝐴2 and the same
resistance to flow 𝑅.
o Obviously, 𝜏1 > 𝜏2 i.e. the tank with a larger capacity has a large time constant.
o Suppose that the two tanks are subjected to the same unit step change in the inlet flow.
o Liquid level response in each tank is according to the Fig.17.5.
o In this case, the tank with a smaller cross-
h
sectional area has a faster response and reaches A2 ( 2 )
the ultimate value faster.
A1 ( 1 )
o However, both tanks reaches the same ultimate
A1  A2
value.
1 2 3 4 t
Fig.17.5 effect of time constant
 Effect of parameters on response of first-order system

o Suppose now that the tanks have different cross-sectional areas and different resistance such
that A R 1
 2
......................................................17.41
A2 R1

o Therefore,  p1  A1R1   p 2  A2 R2 ......................................................17.42

o Since 𝐴1 > 𝐴2 the 𝑅2 > 𝑅1 or 𝐾𝑝2 > 𝐾𝑝1 and the response is shown in Fig.17.5.
K p 2  R2

o Since the two tanks have the same time constant the K p1  R1
initial rate of response is the same in both cases. h

o However, the tank with a large resistance (𝐾𝑝2 ) lets 1 1


less liquid to flow out and hence the liquid level grows to  p1  p2
reach a new steady state value which is greater than
t
that of tank with smaller resistance (𝐾𝑝1) Fig.17.5 effect of steady state gain
 Response of first-order systems to other inputs
o The step response discussed here is not the only way in which the input can change.
o Other inputs as listed below can also occur. Their Laplace transforms have been derived.
o Thus, if input changes according to any one of the forms shown below, procedure to
obtain response is the same as for step input described.
Standard process inputs
Ramp input 0 t 0 Sinusoidal  0
f (t )  
t0
f (t )  
at t  0 sin t t  0
Rectangular pulse Random inputs
• Stochastic and exhibits
0 t 0 random fluctuations.

f (t )   A 0  t  tw • Characterised in statistical
0 t  tw
 terms.
 Response of first-order systems to other inputs
Example 19.1
A stirred tank system described by the following equation is used to preheat a reactant
containing suspended solid catalyst at a constant flow-rate of 1,000 kg/h. The volume in the
tank is 2 m3 , and the density and specific heat of the suspended mixture are, respectively, 900
kg/m3 and 1 cal/g oC. The process initially is operating with inlet and outlet temperatures of 100
and 130 oC. 1 wC 1
T ( s )  Q( s )  Ti( s )
m m
s 1 s 1
w w
a) What is the heater input at the initial steady state and the value of 𝐾 and 𝜏?
b) If the heater input is suddenly increased by 30%, how long will it take for the tank
temperature to achieve 99% of the final temperature change?
c) Assuming the tank is at its steady state. If the inlet temperature is increased by suddenly
from 100 to 120 oC, how long will it take before the outlet temperature changes from 130
to 135 oC?
 Response of first-order systems to other inputs
Question 19.2
A heater for a semi-conductor wafer has first-order dynamics, that is, the transfer function relating
changes in temperature 𝑇 to changes in the heater input power level 𝑃 is
T ( s ) K

P( s )  s  1
Where K has units [𝑜𝐶/𝑘𝑊] and 𝜏 [𝑚𝑖𝑛]. The process is at steady state when an engineer changes
the power input stepwise from 1 to 1.5 kW. She notes the following:
1) The process temperature initially is 80oC.
2) Four minutes after changing the power input, the temperature is 230oC.
3) Thirty minutes later the temperature is 280oC.
a) What are 𝐾 and 𝜏 in the process transfer function?
b) If at another time the engineer changes the power input linearly at a rate of 0.5 kW/min, what
can you say about the maximum rate of change of temperature? When does it occur and how
large?
 Response of first-order systems to other inputs

Question 12.3
A thermometer has first-order dynamics with a time constant of 1 sec and is placed in a
temperature bath at 120 oF. After the thermometer reaches steady state, it is suddenly placed
in a bath at 140 oF for 0 ≤ 𝑡 ≤ 10 𝑠𝑒𝑐. Then it is returned to the bath at 100 oF.
a) Sketch the variation of measured temperature Tm(t) with time.
b) Calculate Tm(t) at t = 0.5 sec and at t = 15.0 sec.
 First-order systems: Ramp response

o For a ramp input 𝑓 𝑡 = 𝑎𝑡 the Laplace transform is 𝐹 𝑠 = 𝑎Τ𝑠 2 and the output is
thus
Y (s) K K a Ka
 or Y ( s)   .........................(17.43)
F (s)  s  1  s 1 s 2
s  s  1
2

o Expanding into partial fraction expansion yields

Ka  A B C
Y (s)   Ka    2  .........................(17.44)
s  s  1
2
 s  1 s s 
o Values of the constants are 𝐴 = 𝜏 2 , 𝐵 = −𝜏 and 𝐶 = 1, therefore
Ka 2 Ka Ka
Y ( s)    2 ........................................................(17.45)
 s 1 s s
o Using the table of inverse transforms then
Y (t )  Ka e  t   Ka  Kat  Ka  e  t   1  Kat......................................(17.46)
 First-order systems: Ramp response

o For large values of time 𝑡 ≫ 𝜏 , the ultimate

response is
Y (t )  Ka  t    .....................(17.47)

o Above equation implies that after an initial transient

period, ramp input yields a ramp output with slope
equal to 𝐾𝑎.
o However, the ramp response is shifted in time by
Fig.17.6 Ramp response of a first-
the process time constant 𝜏. See Fig. 17.6 order system

o Sinusoidal response
o Consider a sinusoidal input 𝑢 𝑡 = 𝐴𝑠𝑖𝑛𝜔𝑡.
 First-order systems: Sinusoidal response

o The Laplace transform of this input is given by


u ( s)  ............................................................(17.48)
s 
2 2

KA
o Thus Y ( s)  ..............................................................(17.49)
 s  1  s   
2 2

o Which can be expanded into partial fractions

KA  C1 C2 C3 
Y (s)   KA  
 s  1 s  j s  j  ............(17.50)
 s  1  s   
2 2
 
o The values of the coefficients are as follows
2 j   j  
C1  2 2 , C2  and C   ............(17.51)
  1 2 (1   2 2 )
2
2 (1   2 2 )
o And
 1 2 1 j   j   1 
Y ( s )  KA  . 2 2  .  .  .......(17.52)
 s  1    1 s  j 2 (1    ) 2 (1    ) s  j 
2 2 2 2
 First-order systems: Sinusoidal response

o Simplifying KA   2 j   j   
Y ( s)  2 2     ...................................(17.53)
   1  s  1 2  s  j  2 ( s  j ) 

KA   2  s  1  KA   2 s 1 
Y ( s)  2 2    
   1  s  1 s 2   2   2 2  1  s  1 s 2   2 s 2   2 
.................(17.54)

o And after taking the inverse transforms

KA
Y (s)  
 e t 
  cos t  sin t  .........................(17.55)
  1
2 2

o And by using compound angle identities

KA e  t  KA KA
Y (s)   cos  t  sin t.........................(17.56)
  1   1
2 2 2 2
  1
2 2

KA e  t  KA
Y (s)   sin t    ......................................(17.56)
  1
2 2
  1
2 2
 First-order systems: Sinusoidal response

o Where ∅ = 𝑡𝑎𝑛−1 (𝜔𝜏).