Transportation Lag

HIGHER-ORDER SYSTEMS:
SECOND-ORDER AND TRANSPORTATION LAG
CHAPTER 7
SECOND-ORDER SYSTEM
Transfer Function
A second-order system can arise from two first-order systems in series.
Some systems are inherently second-order, and they do not result from a
series combination of two first-order systems.
Inherently second-order systems are not extremely common in chemical engineering
applications.
Most second-order systems that we encounter will result from the addition of a
controller to a first-order process.
Second-order system or a quadratic lag

Consider a simple manometer the pressure on both legs of the manometer is initially
the same.
The length of the fluid column in the manometer is L.

At time t= 0, a pressure difference is imposed across the legs of the manometer.
Assuming the resulting flow in the manometer to be laminar and the steady-state
friction law for drag force in laminar flow to apply at each instant, we will determine
the transfer function between the applied pressure difference ∆P and the manometer
reading h.
Perform a momentum balance on the fluid in the manometer:
The term for the skin friction at the wall is obtained from the Hagen-Poiseuille
relationship for laminar flow .
Note that V is the average velocity of the fluid in the tube, which is also the
velocity of the interface, which is equal to
The momentum correction factor b accounts for the fact that the fluid has a
parabolic velocity profile in the tube, and the momentum must be expressed as
b mV for laminar flow.
The value of b for laminar flow is 4/3.
Substituting the appropriate terms produces the desired force balance equation
for the manometer.
• Dividing both sides by rg ( p D2/4), we arrive at the standard form for a second-
order system.
If the fluid column is motionless ( dY/dt =0) and located at its rest position
(Y= 0) before the forcing function is applied, the Laplace transform
All such systems are defined as second-order.
Note that it requires two parameters, t and z, to characterize the dynamics of a

second-order system in contrast to only one parameter for a first-order system.
Response of a second-order system to some of the common forcing functions,

namely, step, impulse, and sinusoidal.
Step Response
If the forcing function is a unit-step function,
In terms of the manometer shown in, this is equivalent to suddenly applying a

pressure difference [such that X (t) =∆ P / rg= 1] across the legs of the manometer
at time t = 0.
Superposition will enable us to determine easily the response to a step function of
any other magnitude.
The quadratic term in this equation may be factored into two linear terms that
contain the roots
-
For this case, the inversion of Eq. yields the result
Since z < 1, indicate a pair of complex conjugate roots in the left

half-plane and a root at the origin.
The constants C1, C2, and C3 are found by partial fractions.

• Y (t) is plotted against the dimensionless variable t /t for several values of z.
• Note that for z < 1 all the response curves are oscillatory in nature and become
less oscillatory as z is increased.
• The slope at the origin is zero for all values of z.
• The response of a second-order system for z < 1 is said to be underdamped.
• If we step-change the pressure difference across an underdamped manometer,
• The liquid levels in the two legs will oscillate before stabilizing.
• The oscillations are characteristic of an underdamped response.
Show that the roots s1 and s2 are real and equal.
The response, which is plotted in Fig. is nonoscillatory.
This condition, z=1, is called critical damping and allows the most rapid approach
of the response to Y= 1 without oscillation.
CASE III STEP RESPONSE FOR z > 1
For this case, the inversion of Eq. gives the result
where the hyperbolic functions are defined as

• The response has been plotted for several values of z.
• Notice that the response is nonoscillatory and
becomes more “sluggish” as z increases.
• This is known as an overdamped response.
• As in previous cases, all curves eventually approach
the line Y = 1.
Actually, the response for z > 1
This is true for z > 1 because the roots s1 and s2 are real,
may be factored into two real linear factors
Note that if t1 t2, then t = t1 = t2 and z= 1.

Terms Used to Describe an Underdamped System
Overshoot
Overshoot is a measure of how much the response exceeds the ultimate value
Following a step change and is expressed as the ratio A/B in Figure
The overshoot for a unit step is related to z by the expression

Why are we concerned about overshoot?
Temperature in our chemical reactor cannot be allowed to exceed a specified temperature
to protect the catalyst from deactivation, or if it’s a level control system,
These physical limitations, we can determine allowable values of z and choose

our control system parameters to be sure to stay within those limits.
Decay ratio
The decay ratio is defined as the ratio of the sizes of successive peaks and is given
by C/A in Figure.
The decay ratio is related to z by the expression
Notice that larger z means greater damping, hence greater decay.
Rise time This is the time required for the response to first reach its ultimate value
and is labeled tr in Figure.
The reader can verify from Figure that tr increases with increasing z.
Response time This is the time required for the response to come within 5
percent of its ultimate value and remain there.
The response time is indicated in Figure. The limits 5 percent are arbitrary, and other
limits can be used for defining a response time.
Period of oscillation The radian frequency (radians/time) is the coefficient of t in the

sine term; thus,
Since the radian frequency w is related to the cyclical frequency f by w = 2pf,
it follows that
where T is the period of oscillation (time/cycle), T is the time elapsed between peaks.
It is also the time elapsed between alternate crossings of the line Y = 1.
Natural period of oscillation
If the damping is eliminated [ B= 0, z = 0], the system oscillates continuously without

attenuation in amplitude.
Under these “natural” or undamped conditions, the radian frequency is 1/ t, as shown
by when z= 0.
This frequency is referred to as the natural frequency w n
The corresponding natural cyclical frequency fn and period Tn are related by the
expression
Thus, t has the significance of the undamped period.

Natural frequency is related to the actual frequency by the expression
Notice that for z < 0.5 the natural frequency is nearly the same as the actual
frequency.
In summary, it is evident that z is a measure of the degree of damping, or the

oscillatory character, and t is a measure of the period, or speed, of the response of
a second-order system
Impulse Response
If a unit impulse d (t) is applied to the second-order system, then from Eqs. and the
transform of the response is
As in the case of the step input, the nature of the response to a unit impulse will
depend on whether the roots of the denominator of Equation are real or complex.
The problem is again divided into the three cases shown in Table , and each is
discussed below.
CASE I IMPULSE RESPONSE FOR z < 1
CASE II IMPULSE RESPONSE FOR FOR
For the critically damped case, the response is given by
CASE III IMPULSE RESPONSE FOR.

However, the impulse response always returns to zero.
Terms such as decay ratio, period of oscillation, etc., may also be used to describe the
impulse response.
Sinusoidal Response
If the forcing function applied to the second-order system is sinusoidal
The inversion of Equation may be accomplished by first factoring the two

quadratic terms to give
Here sa and sb are the roots of the denominator of the transfer function and are given
by Eqs.
For the case of an underdamped system ( z < 1), the roots of the denominator are a
pair of pure imaginary roots (+jw, -jw) contributed by the forcing function and a pair
of complex roots
We may write the form of the response Y( t)

If the constants C1 and C2 are evaluated,
Depending upon the values of z and wt. This is in direct contrast to the sinusoidal
response of the first-order system, where the ratio of the output amplitude to the input
amplitude is always less than 1.
The output lags the input by phase angle |f|.
TRANSPORTATION LAG
A phenomenon that is often present in flow systems is the transportation lag.
Synonyms for this term are dead time and distance velocity lag.
System with transportation lag.
If a step change were made in x (t) at t = 0, the

change would not be detected at the end of the
tube until t later, where t is the time required
for the entering fluid to pass through the tube.
response y ( t) at the end of the pipe would be identical with x ( t) but again delayed
by t
The transportation lag parameter t is simply the time needed for a particle of fluid to
flow from the entrance of the tube to the exit, and it can be calculated from the
expression
Relationship between y (t) and x (t) is

Deviation variables X= x – xs , and Y= y - ys give
If the Laplace transform of X( t) is X( s), then the Laplace transform of
X( t - t) is e-st X ( s ).
We shall see in a later chapter that the presence of a transportation lag in a control
system can make it much more difficult to control.
In general, such lags should be avoided if possible by placing equipment close
together. They can seldom be entirely eliminated.
APPROXIMATION OF TRANSPORT LAG.
The transport lag is quite different from the other transfer functions (first-order, second-
order, etc.)
It is not a rational function (i.e., a ratio of polynomials.)
• The transport lag can also be difficult to simulate by computer.
• For these reasons, several approximations of transport lag that are
useful in control calculations are presented here.
• One approach to approximating the transport lag is to write e-t s as
1/ et s and to express the denominator as a Taylor series;
the result is
Keeping only the first two terms in the denominator gives
This approximation, which is simply a first-order lag, is a crude approximation of

a transport lag.
An improvement can be made by expressing the transport lag as
Expanding numerator and denominator in a Taylor series and keeping only terms
of first-order give
This expression is also known as a first-order Padé approximation.

Another well-known approximation for a transport lag is the second-order
Padé approximation:
Equation (7.48) is not merely the ratio of two Taylor series; it has been optimized to
give a better approximation.
• The step responses of the three approximations of transport lag presented in Figure.
The step response of ets is also shown for comparison.
• Notice that the response for the first-order Padé approximation drops to 1 before rising
exponentially toward 1.
• The response for the second order Padé approximation jumps to 1 and then descends
to below 0 before returning gradually back to 1.
• Although none of the approximations for e-ts is very accurate, the approximation
for e-t s is more useful when it is multiplied by several first-order or second-order
transfer functions.
• In this case, the other transfer functions filter out the high-frequency
content of the signals passing through the transport lag, with the result that the
transport lag approximation,
• when combined with other transfer functions, provides a satisfactory
result in many cases. The accuracy of a transport lag can be evaluated most clearly
in terms of frequency response, a topic covered later in this book.

Transportation Lag

Uploaded by

Copyright:

Available Formats

Transportation Lag

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

Transportation Lag

Uploaded by

Copyright:

Available Formats

HIGHER-ORDER SYSTEMS:

SECOND-ORDER AND TRANSPORTATION LAG

Second-order system or a quadratic lag

The length of the fluid column in the manometer is L.

All such systems are defined as second-order.

Note that it requires two parameters, t and z, to characterize the dynamics of a

Response of a second-order system to some of the common forcing functions,

If the forcing function is a unit-step function,

In terms of the manometer shown in, this is equivalent to suddenly applying a

Since z < 1, indicate a pair of complex conjugate roots in the left

The constants C1, C2, and C3 are found by partial fractions.

where the hyperbolic functions are defined as

may be factored into two real linear factors

Note that if t1 t2, then t = t1 = t2 and z= 1.

Following a step change and is expressed as the ratio A/B in Figure

The overshoot for a unit step is related to z by the expression

These physical limitations, we can determine allowable values of z and choose

Period of oscillation The radian frequency (radians/time) is the coefficient of t in the

Natural period of oscillation

If the damping is eliminated [ B= 0, z = 0], the system oscillates continuously without

Thus, t has the significance of the undamped period.

In summary, it is evident that z is a measure of the degree of damping, or the

For the critically damped case, the response is given by

CASE III IMPULSE RESPONSE FOR.

If the forcing function applied to the second-order system is sinusoidal

The inversion of Equation may be accomplished by first factoring the two

We may write the form of the response Y( t)

System with transportation lag.

If a step change were made in x (t) at t = 0, the

Relationship between y (t) and x (t) is

Keeping only the first two terms in the denominator gives

This approximation, which is simply a first-order lag, is a crude approximation of

This expression is also known as a first-order Padé approximation.

You might also like