PIDtuningapp

Here you can run a javascript app, or download an app for Android for tuning PID controllers based on the open loop step response of the plant.

It requires the user to introduce some measurements obtained from an open loop step experiment in the plant. The measurements are:

The final output value, to calculate the static gain, K.
The times that takes the output to reach, the 5%, the 35.3%, and the 85.3% of the final value.

With those measurements, the application computes an approximate model with three real poles and a time delay defined by parameters : K, tau, L and alpha.

First, the user selects the type of controller: PI or PID. The desired robustness is then selected by the user introducing the value of Ms (through a slider or a numeric input).

If the controller is PI, the user can select the desired detuning factor (gamma<1). For gamma=1, the optimum PI controller is obtained (the one that minimizes the disturbance IAE). For lower values of gamma one obtains slower PI controllers with the same robustness.

If the controller is PID, the user can select the derivative filter parameter, N, to find the required compromise between performance and measurement noise amplification.

The application shows the performance indicator IE=1/Ki, and the high frequency noise amplification, Kp for PI, or Kp(1+N) for PID, to guide in the selection of the controller. The expected response to a step change in the reference and to a step input disturbance is also shown.