embello

Embedded software from JeeLabs

Embello Home Page
Hardware Documentation
Forth Library Documentation

PID Control Library

A proportional–integral–derivative controller (PID controller) is a control loop feedback mechanism (controller) commonly used in industrial control systems. A PID controller continuously calculates an error value as the difference between a desired setpoint and a measured process variable and applies a correction based on proportional, integral, and derivative terms (sometimes denoted P, I, and D respectively) which give their name to the controller type. (Source: Wikipedia)

This library implements a PID controller. It does not do anything by itself but provides an API to do the calculations. You feed the currently measured variable (process variable) in the PID controller and get a correction value (control value) back. Where you get the process variable from and what you do with the control value is your responsibility. You also have to make sure you call pid periodically.

You do have to initialize the PID (pid-init) before using it for the first time. The PID controller will start in manual mode and always return 0 until you set a desired setpoint and switch it to auto mode.

API

: pid-init ( f_kp f_ki f_kd s_sampletime s_outlimit -- )
: tuning  ( f_kp f_ki f_kd -- ) \ Change tuning parameters on a running PID

: pid ( s_is -- s_corr )        \ Calculate new PID value

: set ( s -- )      \ Change setpoint of a running PID
: manual ( s -- )   \ Manual output override
: auto ( -- )       \ Switch back to auto mode
: autohold ( -- )   \ Switch back to auto mode and hold the current process variable

The first words are used to initialize the PID and change the tuning parameters of a running pid controller. All the values for Kp, Ki and Kd are always fixed-point values. Sampletime is in milliseconds and outlimit is the maximum value the control value (output) will reach. The output will always be a value from 0 to outlimit (including).

pid is where the magic happens. You are responsible for calling this word periodically with the current process variable on the stack, even if the PID controller is currently in manual mode. The word pid will return the override value as long as it is in manual mode or the calculated control value if it’s in auto mode.

The last group of commands is for the normal operation. Changing the setpoint and switching between auto and manual mode. Switching to manual mode requires a fixed control value (output value) which will be returned by pid as long as it is running in manual mode. Switching back to auto mode using auto will keep the last supplied setpoint and tries to reach it again. If you use autohold to switch back to auto mode the PID will use the current process variable as new setpoint and tries to hold this value.

Examples

Initialize and enable PID with Kp=120, Ki=1.5 and Kd=0.0075. 100ms period, maximum output value set to 10’000

120,0 1,5 0,0075 100 10000 pid-init

You have to implement something which calls pid periodically (every 100ms in the example above). It does not matter how you implement this, so this part is omitted here.

Now set initial setpoint and start the PID:

3100 set \ Set desired setpoint to 3100
auto     \ enable PID

Now the controller is running and tries to reach the given setpoint.

If you do want to stop the output (set it to 0) you should use this command:

0 manual

You do not have to change anything in the PID itself and you still have to call it every 100ms as before. The return value of pid will be fixed to the value 0.

Set a new setpoint and enable PID again:

2000 set \ Set desired setpoint to 2000
auto     \ enable PID