You are on page 1of 2

PID Control

Proportional-integral-derivative
explained
By Javier Gutirrez All these systems share one Open-loop systems lack the How a PID Controller Works
Product Manager common characteristic: they have information needed to change PID controllers have three com-
National Instruments Corp. one input that can be controlled the value on the controller to ponents: the first component,
to produce a desired change in correct error. To do this, en- proportional control, makes the
Proportional-integral-derivative the output signal. gineers must add feedback actuation by multiplying the
(PID) control is the most com- A mathematical model can ap- information, resulting in a current error. The second compo-
mon control algorithm used in proximate how the plant behaves. closed-loop system. nent, integral control, takes into
industry today. The popularity of A “black box” with one input and What is the purpose of the account the error history. Finally,
PID controllers can be attributed one output can represent this controller? It takes into ac- derivative control acts based on
to their effectiveness in a wide plant model. In the case of a DC count the desired output of the the rate of error, as it is an indica-
range of operating conditions, motor connected to a conveyor plant (setpoint) and its actual tor of how the error is going to
their functional simplicity and belt, the input would be the volt- state and thus finds the error. evolve in the near future.
how easily engineers can imple- age applied, and the output would Based on this error, it works Proportional control acts
ment them using current com- be the speed at which it turns. to apply the input necessary much like a human operator
puter technology. This article dis- On a hydraulic system, the in- to obtain the desired output. would. The bigger the error, the
cusses PID control and practical put would be how open the valve Consider a plant that has an bigger the actuation applied.
implementations and provides is, and the output would be the input set to zero and that is This can lead to oscillations or,
a brief overview on how to tune height on the tank. Finally, on a stable. An engineer could set because of the dynamics, per-
PID controllers. pH reactor plant, the input is the the input up a step to observe manent error if the error is not
Why do engineers need con- rate at which a chemical reactor how it behaves. This is called significant enough to create a
trol, and which types of control is fed, and the output is the pH the “step response” in control corrective action. This is called a
can they use? Engineers use measured. theory. “steady-state error.”
control to either modify plant
behaviour to keep system out-
put stable and improve response
time (the time the plant takes to
go from its current state to the
state defined with the new in-
put) or minimize the energy the
plant uses to transition between
states. To accomplish this, en-
gineers can operate controllers
themselves or use mechanical
devices or PC or PC-related Figure 1: All plant systems have one input that can be controlled to produce a desired change in the output signal.
technology. Thanks to the de-
velopment of computer-based
systems, engineers can use PCs
and appropriate hardware to in-
teract with plants and read their
output and input signals.

Definitions
A plant is a physical process that Figure 2. Open-loop systems such as this lack the information needed to change the value on the controller to correct error.
must be controlled, meaning a
PID controller needs to increase
the stability or performance of
the process. A tank with one in-
put valve and a fixed output flow,
where height level is regulated,
is one example. Other examples
include a deposit where a chemi-
cal reaction takes place, and pH is
maintained precisely, or a motor
connected to a conveyor belt
whose speed must be sustained. Figure 3. A closed-loop system includes a feedback mechanism to correct error.

EE Times-India | eetindia.com 
components “ including aca- How to Build a PID Controller
demic, parallel and serial. In an As mentioned before, there are
academic implementation, each different ways to implement PID
term has a constant: Kp for pro- controllers: pneumatic, hydraulic
portional, Ki for integral and Kd and so on; however, PC-based
for derivative. control tools, programmable logic
The proportional actuation is controllers and programmable
common, so “integral time” (Ti, automation controllers (PACs) pro-
equals to Kp/Ki) is used for the vide the easiest and fastest ways to
integral component, and “de- design, prototype and deploy PID
rivative time” (Td equals Kp*Kd) controllers.
is used for the derivative compo- Modern PC-based systems are
nent. This decouples the propor- especially prepared for PID imple-
tional part of the controller from mentation because they are highly
the dynamics of the controller cost-effective and have high pro-
related to Kp and Ki. cessing power and data-logging
In a parallel implementation, capabilities, as well as hardware
each component is calculated capable of interacting with physi-
independently and then added cal sensors and actuators.
together, while in a serial imple- The new generation of graphi-
mentation the output of each cal programming languages
component feeds the next, so provides flexible and powerful
they are interconnected. The programming capabilities “ users
latter is the most restrictive of with little or no control or pro-
the three and was used in older gramming experience can imple-
pneumatic controllers. ment advanced PID algorithms.
Figure 4. PID controllers have three components: the first component, It is important to consider that An example of this architecture
proportional control, makes the actuation by multiplying the current error. some controllers do not have all is a PC equipped with a National
The second component, integral control, takes into account the error history. three components; it is possible Instruments Compact FieldPoint
Finally, derivative control acts based on the rate of error, as it is an indicator of to have P, PD or PI controllers, or CompactRIO PAC and NI
how the error is going to evolve in the near future.
where one or two components LabVIEW graphical development
are missing. software.
There are several ways to set
up the different constants, de- Limitations of PID
pending on whether the plant Although PID has been widely
can be offline or online and on adopted, it has some limitations.
how long it takes for a change in For example, it can only handle
the input to be reflected in the systems with one input and one
Figure 5. This example uses Ziegler-Nichols rules, which consist of having an output. The following example output. It expects the system to
offline plant and implementing a PID but with Ki and Kd set to zero.
uses Ziegler-Nichols rules. behave in a linear manner through-
Integral control considers gains may also cause the system These rules consist of having out the entire working range, and
the history of the error. It keeps to become unstable. an offline plant and implement- it does not take into account the
adding the value of the error, Derivative control evolves ing a PID but with Ki and Kd set actuators’ physical or mechanical
consequently removing the with the rate of change. This al- to zero. If engineers increase un- limitations.
steady-state error once the sys- lows the controller to estimate til the system oscillates, they can To overcome these problems,
tem is near its final state. how the error will change in the consider the following rules to engineers can use a wide variety
In systems with long transi- near future. It makes the system find the different gains using the of techniques, including gain
tions, however, integral control react faster to changes, but it also proportional gain implemented scheduling, cascading controllers
integrates over time, causing can make it unstable. Another (Kc) and the period of the oscil- or the use of filters with the PID
“integral windup,” where the drawback is that derivatives are lation (Pc). algorithm. Also, with computer-
system evolves beyond the very sensitive to noise, making These are only rules of thumb based technology, engineers can
desired setpoint. The output the actuation itself noisy. and can vary for different types build custom, PC-based PID con-
overshoots “ that is, it goes over There are several PID imple- of plants. Engineers may need to trollers capable of overcoming
the desired value before going mentations “ or methods for tune the plant manually to fur- these limitations.
back and settling. High integral calculating the different PID ther improve the overall system.

 eetindia.com | EE Times-India

You might also like