How do we model and design a controller for a machine has no mathematical model? whats the best way ?

my task is to control a machine (process control) has no first principle model(mathematical model), but we manage to collect real data from this machine, i.e flow rate and concentration of liquid, i would like to design a feedback controller for this machine (PID) and i need your help to advice me with the right step that i can start with. i tried system identification and i will appreciate if you could advice me with the right way (guide).

for the hard ware i think PLC is the best option to use so is there away to convert the Matlab simulink if i get success designing the controller to PLC, and VFD will be used to control the feed rate of the pump.

thank you very much and i am looking forward to hearing from you.

Answer 1

[Arkadiy Turevskiy]

You can follow the workflow described here .

[Magdi Nabi]

is it possible from read data i get from machine (2 input and 1 output) to simulate this machine ( nonlinear model),

i did use system identification tools to get transfer function but i am looking for another method to represent my nonlinear machine.

on the other hand, from linear model used by system identification and connected transfer function to MPC block to control, is it possible to convert this simulink to PLC and if so is it going to fit to the real machine? i know i have to do manual tuning of course but is it applicable?

Thank you very much