asimov
Member level 3
Hi All,
This is not going to be another post asking for help with code or error finding! I was trying to implement a heater controller using popular PID algorithms. I followed lectures from Bruce Lander from Cornell university as his lecture seemed most useful and simple one in terms of describing the PID algorithm from algorithm perspective and not just writing second order diffrential equation. URL;
Most books are lectures seems to mathermatically intense and lack the details of implementation as an algorithm, this lecture seems to bit different listen to the lecture from 31 onwards upto 43 minutes. I wrote the following code for the PID implementat. I need inputs from someone if the integral term i am calculating is correct (including anti-integral windup)
I need feed back from experienced users about the integral term calculation is correct and if not what it should be.
The PWM limit of the timer is 0-1000 and this code generated a value between 0-1000 that is used fby PWM routine.
Any comments and reviews will be appreciated.
This is not going to be another post asking for help with code or error finding! I was trying to implement a heater controller using popular PID algorithms. I followed lectures from Bruce Lander from Cornell university as his lecture seemed most useful and simple one in terms of describing the PID algorithm from algorithm perspective and not just writing second order diffrential equation. URL;
Most books are lectures seems to mathermatically intense and lack the details of implementation as an algorithm, this lecture seems to bit different listen to the lecture from 31 onwards upto 43 minutes. I wrote the following code for the PID implementat. I need inputs from someone if the integral term i am calculating is correct (including anti-integral windup)
I need feed back from experienced users about the integral term calculation is correct and if not what it should be.
The PWM limit of the timer is 0-1000 and this code generated a value between 0-1000 that is used fby PWM routine.
C:
unsigned int PID_Calculate(float SetPoint, float InputValue, char flag)
{
float Err_Curr, Derivate_Term, Proportional_Term, Integral_Term, Result;
static float Err_Prev, Acum_Integral_Err;
Err_Curr = SetPoint - InputValue;// Error is calculated by calculating difference of Set_point and measured value.
if (PID_First_Time ==1) // Since the controller is running for the first time there will be no accumilative errors
{
PID_First_Time = 0;
Err_Prev = 0.0;
Acum_Integral_Err = 0.0;
}
// Proportional term is calculated by multiplying current error by Kp easy peasy!!
Proportional_Term = Err_Curr * PID_Kp;
// This term is about calculating rate of change of error over time(PID loop cycle times) multiplied by Kd.
Derivate_Term = ((Err_Curr - Err_Prev)/ PID_CYCLE_TIME)*PID_Kd;
// Integral Term is calculated below
Acum_Integral_Err = Acum_Integral_Err + (((Err_Curr + Err_Prev)/2) * PID_CYCLE_TIME);
if (Err_Curr <= 0)
{
Acum_Integral_Err *= 0.90;//scale down the integral term to reduce over shoot see this in Bruce Land video around 40 minute
}
Integral_Term = PID_Ki * Acum_Integral_Err ;
// This is integral windup
if (Integral_Term < -200.00) // Integral term limited to 20% of full sale value
Integral_Term = -200.00;
if (Integral_Term > 200.00)
Integral_Term = 200.00; // Integral term limited to 20% of full sale value
Result = Proportional_Term + Integral_Term + Derivate_Term; //Adding P + I + D terms
if (Result > 1000.00) // Result bound to max upper limit of the PWM
{
Result = 1000.00;
}
Any comments and reviews will be appreciated.