Proportional plus Integral plus Derivative (PID)

Control of First-Order Systems

Babar Ahmad

April 15, 2022

1 PD Controller
We shall start with the Proportional plus Derivative control of a First-Order plant.
With reference to Figure 1 below, the transfer function of the controller is now given
by
D(s) = (K p + sK d )
and that of the plant by
1
G(s) = .
(T s + 1)

Figure 1: Basic structure of a unity-feedback control system with a Controller D(s).

The closed-loop transfer function is given by

C (s) K p + sK d
= (1)
R(s) (T + K d )s + K P + 1

The input to the system is, once again, a unit-step (i.e., R(s) = 1/s). Thus

1 K p + sK d
C (s) = (2)
s (T + K d )s + (K P + 1)

After some algebra, the inverse Laplace Transform of the above equation yields

Kp Kp
¸ 1+K p
Kd
·
− t
c(t ) = − − e T +K d (3)
1 + Kp 1 + Kp T + Kd

1
Babar Ahmad MET 353

which, upon further simplification, reduces to

Kp Kp T − Kd
· ¸ 1+K p
− t
c(t ) = − e T +K d (4)
1 + Kp (1 + K p )(T + K d )

From the above equation, it is clear that the time-constant of the closed-loop system
is given by
T + Kd
Tcc =
1 + Kp
and the steady state error by
1
e ss = lim e(t ) =
t →∞ 1 + Kp

In contrast to a system with a Proportional Controller only, the output of the system
with a Proportional plus Derivative is not zero at t = 0, but is given by
Kd
c(t )|t =0 =
1 + Kd
The main implications of incorporating a Proportional plus Derivative Controller
are:

• The steady-state error, which was unity in case of Derivative only control, can
now be reduced by choosing an appropriate value of K p . Note that the value of
K d plays no role in determining the steady–state error, which depends entirely
on the magnitude of K p .

• The magnitude of c(t ) at t = 0 depends on the value of K d . Therefore, by choos-

ing a suitable value of K p , the Derivative Controller speeds up the system re-
sponse during the initial phase of operation. As the time evolves, the Propor-
tional Controller takes over and is mainly responsible for the control action.
The Derivative Controller is most effective at the time interval close to zero.

Figure 2 shows the response of the system subjected to a unit-step input. The
time–constant T of the open-loop first-order system is set to unity. The response
of the closed–loop system can now be shaped/tuned by appropriately changing the
values of the controller gains, i.e., K p and K d .

2 PI Controller
For a Proportional plus Integral Controller, the controller transfer function is given
by
Ki
D(s) = K p +
s
where K i is the gain of the Integral Controller.
The closed-loop transfer function is given by

C (s) sK p + K i
= 2 (5)
R(s) T s + (K p + 1)s + K i

Page 2
Babar Ahmad MET 353

Figure 2: The response of a first-order system with a Proportional plus Derivative Controller. The
Kd
value of K p = 20 and K d = 1. Note that the value of c(t ) at t = 0 is given by 1+K d and is equal to 0.5 as
expected.

which may be written in the standard form as

C (s) (sK p + K i )/T

= (6)
R(s) s 2 + K p +1 s + K i
T T

From the above equation, it is clear that the order of the closed-loop system is two,
and hence exhibits second-order response to a unit-step input. On comparison with
the transfer function of a prototype second-order system,

C (s) ω2n
=
R(s) s 2 + 2ζωn s + ω2n

it is clear that
Kp + 1
2ζωn =
T
and
Ki
ω2n =
T
Since the nature of the response of a second-order system depends on the values
of ζ and ωn , the response can be shaped by choosing the appropriate values of K p
and K i to meet the desired performance specifications. The effect of incorporating
either a PD controller or PI controller is the addition of a zero to the closed–loop
transfer function, and the exact location of the zero can be selected by choosing the
appropriate values of K p and K i .

Page 3
Babar Ahmad MET 353

The figure below shows the unit-step response of the system, for various values
of K p and K i .

Figure 3: The unit-step response of a system with Proportional plus Integral Controller. The time-
constant of the original open-loop system is set to one. The curves are intended to depict the qual-
itative nature of the response only. As is clear from the figure, if we try to reduce the rise time, the
overshoot of the response increases. By choosing suitable values of K p and K i , we can achieve a suit-
able trade–off between the rise time and an acceptable system overshoot.

3 PID Controller
For a Proportional plus Derivative plus Integral Controller, the controller transfer
function is given by
Ki
D(s) = K p + sK d +
s
where K p , K d and K i are the gains of the proportional, the derivative and the integral
controller, respectively.
The closed-loop transfer function is given by

C (s) s 2 K d + sK p + K i
= (7)
R(s) (T + K d )s 2 + (K p + 1)s + K i

which may be written in the standard form as

C (s) (s 2 K d + sK p + K i )/(T + K d )
= K p +1
(8)
R(s) s2 + s+ i
K
T +K d T +K d

Page 4
Babar Ahmad MET 353

Once again the order of the closed–loop system is two, and the closed–loop system
will exhibit typical second-order response to a unit–step input. On comparison with
the transfer function of a prototype second-order system

C (s) ω2n
= 2
R(s) s + 2ζωn s + ω2n

it is clear that
Kp + 1
2ζωn =
T + Kd
and
Ki
ω2n =
T + Kd
In contrast to the PI or PD controller case, we now have an additional parameter, that
is, K d , that can be tuned to achieve the desired value of the damping ratio, ζ, and the
undamped natural frequency, ωn .
From equation 7, it is clear that the degree of the numerator polynomial is the
same as that of the denominator polynomial, which means that two zeros are now
added to the closed–loop transfer function, unlike any previous case. This has a sig-
nificant effect on the nature of the transient response, and the controller coefficients
(K p , K d , and K i ) need to be chosen very carefully to achieve the desires transient
response.

Page 5