Hello all,
I have the following challenge (I apologise for the amount of text, but this had to be done without accompanying figures/schemes/pictures).

I am looking for a control configuration for the following system: A static mixer which has two inflows A and B and one outflow C.

The set-up is as follows:

Flow measurement and Temperature measurement in outflow C.
One control valve in inflow A (valve A).
One control valve in inflow B (valve B).

Flow A is main process water system of the plant with temperature depending on weather conditions between 15°C in winter and 30°C in summer. Flow B is hot water with a stable temperature of 55°C. Outflow C has the following requirements/setpoints as a consequence of different demands (sequence operated):

1) 30°C and flow about 9000 kg/hr
2) 40°C and flow about 40000 kg/hr
3) 55°C and flow about 30000 kg/hr

Of course, temperature and flow are coupled in this system. I would like to implement this in the DCS so that the system is decoupled.
I cannot figure out how this should be done.

I am familiar with the following steady state equations:
flow C = flow A + flow B
temp C = (flow A * temp A + flow B * temp B) / flow C

I also know the classic example of the heat exchanger and bypass, flow and temperature measurement/control; one control valve in stream through heat exchanger and one control valve in bypass.

if the output of the flow controller is "a" and the output of the temperature controller (note control action: reverse) is "b" then, this system is decoupled when the outputs to:

-the bypass valve = b-a
-the heat exchanger valve = b+a
another (classic) solution is:
-the bypass valve = b*(1-a)
-the heat exchanger valve = a*b

I would like to have something like this in my static mixer problem.
The objective is that if flow setpoint is increased, both control valves A and B are opened further proportionally as to not change the temperature of the mixture. This means that output of the flow controller "a" should be a factor to both valves. When setpoint 55°C is required, valve A completely closed. I don't know how to implement the "b" in the outputs to the control valves to have the best stable control.

I would appreciate hints and tips

thanks in advance
Sven

 

If you want to use first principles you might have to use the heat capacity of the products in your temperature calculation. In the old days one used PID controllers of which one (the least important parameter) is detuned.

You can also use a MIMO controller (multiple input multiple output - 2x2 in this case) for which several APC algorithms are available.
Another option would be to have two PID controllers driving the two flows. One flow controller's setpoint is in ratio to the other. Driving the setpoint of the main flow controller results in both flows being changed (use a total flow controller). Changing the ratio results in the temperature being changed (use a temperature controller). Also here you might need to use the heat capacity of the products.

 

Thanks for your answer.

The idea is to use MIMO control, like in the heat exchanger example:

one flow controller and one temperature controller in DCS together with a calculation block, which calculates the outputs to both control valves using the formula's like in the example. The inputs to the calculation block are the outputs of the controllers, a and b.
The various Advanced Process Control algorithms you mentioned, maybe that is what I'm looking for.

output to valve A=f(a,b)
output to valve B=f(a,b)

Do you have more info about the APC algorithms?
Thanks