We have a SISO level process plant for which we controlled the water level using Fuzzy Logic Controller successfully. As a validation to the result we were also asked to simulate and compare it with real time controller. So what we did was kept the process in open loop (without controller)and Applied Step and Ramp singals to the pneumatic control valve (actuator element in our plant) and logged the level of the tank.
Input are of range 0-100% whose range is later convereted into 0-10V for applying it to the control valve via cFP (We Used 0 to 100 because in fuzzy logic we set the output range as 0-100)
Outputs - Level of water in tank in cm (0-25)
Logged in both inputs and outputs for step and ramp with fixed sampling interval and brought the data into Matlab for system identification. The estimated Process Model resulted in a transfer function that is same for both the signals.
Then, We used the control design and simulation toolkit in labview to check the process response for the same fuzzy logic to the transfer function we obtained using sys ident. If we are correct so far then the time taken to reach setpoint in both simulation and in real time should be almost same. The simulation for the above transfer function is resulting in the response in which the output reaches the set point, but I have confusion reg. the time taken to reach the setpoint
It took 40 sec to reach a level of 15cm when applied on the harware. I have two questions :
1. Is the procedure I followed for identifying the transfer function correct ?
2. In order to validate the results, the time taken to reach the set point in simulation should be almost equal to the time taken when applied on hardware. So how can I have the simulation in labview show the time properly in seconds
Waiting for some valuable suggestions
I have read your post but I'm still missing the possibility to see the real implementaion.
From my expirence and from reading your post it sounds like your don't saturate your control signals in the simulation. The control signals should offcourse be saturated so they match the maximum water speed achieveable.
National Instruments Denmark