Feedback algorithms Flashcards
Types of feedback systems
on-off
dead zone
proportional
hysteresis
on-off system main features
binary: on or off
epsilon = controller error = how much the output value (y) deviates from the reference value (r) = r - y
what determines the command of the controller in a controller system with an on-off feedback algorithm?
if epsilon, the controller error > 0, the output value is too low, so the command is umax, which intends to make the next output value the maximum it is possible
if epsilon, the controller error < 0, the output value is too high, so the command is umin, where the system aims at decreasing the output value to the minimum
disadvantages of an on-off feedback algorithm
too much oscillation
leading to wear of the controller system
dead zone algorithm
an algorithm that implements the on-off algorithm but includes a deadzone with previously defined values of controller error (epsilon), epsilon min and epsilon max, in which the output command of the controller is 0, so it tells the system to maintain its current settings
advantages of the dead zone algorithm
prevent oscillation of the output
reduce wear on the system
main features of the hysteresis algorithm
it also has a defined range for the controller error where it will maintain the same command setting, u, as it did previously
value of the command depends on previous history of the system
until the controller error exceeds the defined range, the command will stay in either umin or umax. Afterwards, it will swap. Process will repeat
advantages of the hysteresis feedback algorithm
preventing oscillation
reduce wear on control system
why is a proportional controller algorithm used instead of a hystere
what is a proportional controller algorithm?
it changes the value of the command, u, for small values of the controller error, epsilon
the small values of the controller error are described within the defined range of epsilon - < epsilon< epsilon +
the u command becomes the proportional controller gain, k_p multiplied by the maximum or minimum values of the epsilon, depending on the previous history of the controller system
PID controller overall description
A type of feedback algorithm which considers the current, past and future error of a system
the System output is a combination of the errors in the integral, proportional and derivative branches
PID controller: integral controller
As the controller error reaches the steady state error, the error becomes close to 0. When integrating a non-zero error, the result will lead to a linear graph. When this result is added to the system output, it will cause the error to increase, therefore continuously increasing the system output, and decreasing the controller error.
PID integral controller disadvantages
There will have to be a point where the integral of the error will have to be negative, so there is a negative error added to the sum which lowers the controller error, and allows the system output to remain at the reference level. This overshooting effect is not desirable.
How to solve the issues introduced by the integral branch in the PID controller?
add a derivative branch.
This branch calculates the rate of change in the error.
As the system output converges towards the reference value, the controller error decreases and the controller command decreases. The rate of change of the error will therefore become negative. When this result is added to the system output, it causes it to decrease, preventing the overshoot.
what is a proportional controller?
a controller which changes the value of the controller command based on the controller error.
