Once the chamber model had been validated the new control algorithms could be developed.
One of the algorithms chosen was model predictive control (MPC). This advanced algorithm uses a dynamic model to predict future system outputs, and can perform better than PID in systems with larger time delays or high-order dynamics.
After the algorithms were developed using the simulated chamber, they were then fine- tuned by running the controller model on a Performance real-time target machine from Speedgoat, controlling the real chamber.
Simulink Coder was used to automatically generate the code to run on the real-time target machine. The control relays for the chamber (heating, cooling, fan and compressor) were driven by the real-time target machine’s digital outputs, and the chamber’s resistance thermometers were measured using an external conditioning module fed into the analog inputs of the target machine.
The results from the simulation and measurement of the new algorithms were then compared with the previous algorithms.