2. How to Improve your PID Controller
Javier Gutierrez
LabVIEW Product Marketing
2
3. Benefits of Advanced Control and Tuning
Model-based Manual
• A poorly tuned control control < 1% control
valve costs additional
$880/year*
• A bad pH loop incurred
chemical waste of
$50,000/month*
• A bad kiln temp loop cost
$30,000/month* PID needs
PID is fine
manual tuning
*Sources: Cybosoft and ExperTune
3
4. Agenda
• What is PID?
• How to improve performance
Hardware considerations
Upgrade PID Algorithm
Advanced Controllers
• Conclusion
4
5. Agenda
• What is PID?
• How to improve performance
Hardware considerations
Upgrade PID Algorithm
Advanced Controllers
• Conclusion
5
6. What is PID
• Set Point (SP) – Desired control point
• Output (OP) – Controller output
• Process Variable (PV) – Plant/process output
• Error = SP - PV
error OP
SP PV
6
7. PID Parameters
• Proportional
Drive to setpoint
Error → 0, OP → 0
“Steady-state error”
• Integral
Eliminate steady state error
OP proportional to ∫ error
• Derivative
Increase response rate
OP proportional to rate of change of error
7
10. PID Control – Pros and Cons
• Advantages
Proven
Easy to implement
• Disadvantages
Not easy to tune
Not suitable for all systems
• Backlash, friction, and so on
10
11. Agenda
• What is PID?
• How to improve performance
Hardware considerations
Upgrade PID Algorithm
Advanced Controllers
• Conclusion
11
12. How to program PID
Function Blocks
Windows/Real Time
FPGA Control and Simulation
12
18. Feed-Forward
• Commonly used to compensate for a
measurable external disturbance before it affects
a controlled variable.
• e.g. product feed rate changes
18
19. Gain Scheduling
• Used to change gain on real-time depending on
OV.
• Bumpless transfers
19
20. Adaptive PID
• Mixed of On-Line system identification and
common PID control.
• Can handle time-variant systems
20
21. Agenda
• What is PID?
• How to improve performance
Hardware considerations
Upgrade PID Algorithm
Advanced Controllers
• Conclusion
21
22. Advanced Controllers
National Instruments
• Optimal Controllers (LQR, LQG)
• Model Predictive Control (MPC)
• Kalman Filters
• Fuzzy Logic
Third Party Partners
• Neural Networks
• Genetic Algorithms
• Model Free Adaptive
22
23. How to create an advanced Controller
Control Design
• Datalogging • Deployment
• System • Design • Test
Identification • Simulation
• Model Validation
Plant Modeling Implementation
23