Skogestad’s Half Rule

Consider the following two first-order Taylor

approximations of a time delay transfer function:

s s 1 1
e  1  s and e  s  (6.57)
e 1  s
IMC

An ‘‘inverse response time constant” T0inv (negative

numerator time constant) may be approximated as a
time delay
T0inv s
(T s  1)  e
0
inv

This is reasonable since an inverse response has a

deteriorating effect on control similar to that of a time
delay
EKC 462: Advanced Control System for Industrial Process
1
 Similarly, a (small) lag time constant 0 may be
approximated as a time delay:
1  0 s
e
 0s  1
Since

( T s  1)  0 s
inv
IMC

 0 s T0inv s  0 s
0
e  e .e .e
 0s  1
( 0 T0inv  0 ) s s
e e
The effective delay  can be taken as the sum of the
original delay 0, and the contribution from the
various approximated terms
EKC 462: Advanced Control System for Industrial Process
2
 Skogestad’s Half Rule: the largest neglected
(denominator) time constant (lag) is distributed evenly
to the effective delay and the smallest retained time
constant (i.e. one half of its value is added to the
existing time delay (if any) and the other half is added
to the smallest retained time constant).
Time constants that are smaller than the “largest
neglected time constant” are approximated as time
IMC

delays using Equation (6.57)

In summary, let the original model be in the form

 (T inv
j0 s  1)
j  0 s
e
 i
i0 s 1
EKC 462: Advanced Control System for Industrial Process
3
 where the lags i0 are ordered according to their
j0  0
inv
magnitude and T denote the inverse
response (negative numerator) time constants.
Then, according to the half rule, to obtain a order
model e s /( 1 s:  1)
 20  20 h
 1   10  ;   0    i 0   T inv
j0 
2 2 i 3 2
IMC

and, to obtain a second-order model we use

 30
 1   10 ;  2   20  ;
2
 30 h
  0    i 0   T inv
j0 
2 i4 j 2
where h is the sampling period (for cases with digital
implementation).
EKC 462: Advanced Control System for Industrial Process
4
 Example
Consider a transfer function:
K ( 0.1s  1)
G0 ( s ) 
(5s  1)(3s  1)(0.5s  1)

Derive an approximate FOPTD model using

Skogestad’s half rule.
IMC

ANSWER: The largest neglected time constant = 3

 The new time constant,  = 5 + ½(3) = 6.5
Total time delay,  = 1.5 + 0.1 + 0.5 = 2.1
Then G(s) can be approximated as:
2.1s
Ke
g (s) 
(6.5s  1)
EKC 462: Advanced Control System for Industrial Process
5
 Approximation of positive numerator time
constants
If we have positive numerator time constants T0 in the
original model G0(s). It is proposed to cancel the
numerator term (T0s+1) against a ‘‘neighboring’’
denominator term (0s+1) (where both T0 and 0 are
positive and real) using the following approximations
IMC

T0 /  0 for T0   0   ( Rule T 1)
T0 /  for T0     0 ( Rule T 1a )
T0 s  1
 1 for   T0   0 ( Rule T 1b)
( 0 s  1)
T0 /  0 for  0  T0  5 ( Rule T 2)
(~0 /  0 ) def
for ~0  min( 0 ,5 )  T0 ( Rule T 3)
(~ T ) s  1
0 00

EKC 462: Advanced Control System for Industrial Process

6
 Here  is the (final) effective delay, the exact value
of which depends on the subsequent approximation
of the time constants (half rule).

If there is more than one positive numerator time

constant, then one should approximate one T0 at a
time, starting with the largest T0
IMC

Normally select 0 as the closest larger denominator

time constant (0 >T0) and use Rules T2 or T3. The
exception is if there exists no larger 0, or if there is
smaller denominator time constant ‘‘close to’’ T0, in
which case we select 0 as the closest smaller
denominator time constant (0 <T0) and use rules T1.

EKC 462: Advanced Control System for Industrial Process

7
 Example: Consider the process with the
original transfer function
2(15 s  1)
g 0 (s) 
( 20 s  1)( s  1)(0.1s  1) 2

first introduce from Rule T2 the approximation

15s  1 15s
  0.75
20 s  1 20 s
IMC

(Rule T2 applies since T0=15 is larger than 5, where 

is computed below). Using the half rule, the process
may then be approximated as a FOPTD with
k = 2 × 0.75 = 1.5;  = 0.1 2 +0.1 = 0.15;
1 = 1 + 0.1 2 = 1.05
or as a second-order time delay model with
k = 1.5;  = 0.1 2 = 0.05; 1 = 1;
2 = 0.1+ 0.1 2 = 0.15
EKC 462: Advanced Control System for Industrial Process
8