Humidity control using Fuzzy logic?

[Aidy]

Hello,

I am using a Humidifier to change the humidity of sample air flow in my instrument.

For controllig the humidifier.........I have two probes.......a Temperature probe (T) and an RH probe (RH).

So I have four values:

RH actual (the current RH value).......................... T actual (The temperature of the humidiifer)

RH set point   (The RH I want)............................ T set point (The temperature I want)

Now based on the inputs of RH actual and RH set point I want to control T actual and generate T set point values.

How to proceed with this using Fuzzy logic toolkit?? 

The number of input variables? output variables? rules? membership functions? shape? etc etc.......many more things!!!!

Any help will be highly appreciated 🙂

[arvin_labrat]

Hi Aidy

You can't control both.  They are tied together.  My recommendation would be to close the loop around temperature and see what humidity you get.  You can always change the temperature of the gas down stream from you humidifier to increase or decrease you RH.

Hope that helps.

One more thought,  If you are using a low efficient humidifier like a pass over you can not control both.  If you can independently control the temp of the gas and the amount of water vaporized into the air stream then you can.

Message Edited by arvin labrat on 06-09-2010 01:17 PM

[Aidy]

Hello Arvin,

I think I did'nt describe my probelm fully!!

What you are saying is correct. But here the case is different. I want to controll only the humidity of the gas. Since I know what humidity I want (Set point RH), I can ask the fuzzy controller to generate a set point for the heater (Set point T). The heater will heat up a water tank there by adding water vapor to the gas flow thus humidifying it.

The problem is I am not getting how to define these parameters and other required variables in the Fuzzy controller toolbox!!

See the attached VI which is just for an idea.

[Rico P]

Hi Aidy,

Sorry in advance that this will be my only post as I'm going away.

There are examples in the Example Finder in LabVIEW. One I looked at was called "FuzzyEx Tanks with PI Control.vi". To get the Example Finder open LabVIEW and go to Help >> Find Examples. In the Example Finder in the Browse Tab go to Toolkits and Modules >> PID and Fuzzy Logic Control >> Fuzzy Logic

Regards,

[Aidy]

Thanks Rico.

The examples are great. They gave a lot more information than I thought.

But I am still stuckup with choosing the right membership functions!!!!! How to be sure what membership is to be used and what are the points corresponding to that??

Thanks.

[Barp]

Aidy,

The type of membership and values depends exclusively from the problem you are trying to solve. In general, you can just start with triangular and trapezoidal and adjust the limits based on experience and/or knowledge. More you have know about the system, better you can model the information.

Our manual has some information and examples that shows how to define and enter this information:

http://www.ni.com/pdf/manuals/372192d.pdf

Notice that the version 2009 has improved greatly our offering and you can even define you own membership function programmatically.

Hope this helps...