What is the effect of each of the three terms on output response of a pid controller?
i mean how will P term shape/effect the output response similarly what is the effect of I and D terms on output response respectively?
0
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
You have a few questions about PID control. Let’s break it down to its basic concepts.
The P term (for Proportional) is what you might think of as the strength of the controller. If you want a system to reach a particular state, and it is instead at some other state, the P gain multiplies the error to drive the system toward the desired state. In a perfect world, the system gets there. However, if there is a lot of friction, or the P gain is too low, the system may only get partway to the goal. This is overdamped behavior. Conversely, if there is little friction, or the P gain is too high, the system may travel past the goal state and overshoot the goal. In this case, the error will change sign, and the P gain will cause it to go back toward the goal state. This is underdamped behavior. In general, you want the P gain to be relatively high so the system responds strongly (quickly) to error states.
The D gain (for Derivative) helps compensate for underdamped behavior. When the derivative of the error is large, so that the system is approaching the desired state quickly, the D gain responds to counter the rapidly-changing state. In effect, this adds damping to the system to reduce or eliminate overshoots. When the system is close to the goal state, and the error is not changing much, the D term does very little. When the error is reducing quickly, the influence of the D gain can stabilize the performance. Note: if the D gain is very large, it can lead to oscillatory behavior due to the rapidly-changing derivatives.
The I gain (for Integral) serves to drive a system to its goal state in those situations for which the P (&D) gains are not strong enough to overcome the system’s characteristics and cannot precisely reach the goal state. This is really only useful when trying to achieve a steady-state position. The I gain is a multiplier of the integrated error, so as a small fixed error remains over time, the I term grows until it can drive the system to the goal. Note that real physical systems always lag the controller (due to friction, inertia, power supply slew rates, etc), so the integral term should not be enabled until the system is within a threshold of the desired state. Otherwise, the accumulated errors from the lagging response will cause the I term to be large, causing overshoot, due to the accumulated errors during motion, instead of just considering the errors when at steady-state.
$\begingroup$
$\endgroup$
Check this video: https://www.youtube.com/watch?v=4Y7zG48uHRo
It explains how PID algorithm work for controlling self driving cars.
