Fuzzy control versus conventional Control

1. It was an optimized block for PLC (memory and execution time).
2. You need to do  parameters adjustments when needed.
3. Not a perfect improvement compared to conventional method.

I brief, I am keen to test the structure text generator in Simulink because it can be easily converted to SCL and I have the process model most of the time .I hope it is optimized for PLC(memory and execution time).

------------------------------------------------------------------------------------------
Split from Can a fuzzy driver be used on a Tia Portal v14 or another model?.

Good health to you Mr. Hosseini.

Throwing a stone in the bush by playing devil's advocate --

May I assume that a strong background in fuzzy logic theory is required to make such algorithms work correctly? Or could any one just jump in and tweak values until something happens?

 I agree that Fuzzy control is suitable for Complex applications. In process automation , you can see multi-variable and non-linear models and mostly their products are expensive. You can use extra software packages in parallel for optimization by any method you want to ensure higher quality and more profit. It is normal to see such advanced control features in PCS 7 or Emerson DeltaV because of the reasons mentioned.

For a simple application in a factory automation , most of the time conventional methods working perfectly.

@ hdhosseini, good health and spirits to you as well!

Let's say your company is responsible for integrating HVAC systems in a large number of high-rise office towers, and a large number of residential high-rise towers of around the same height and size, one of each being phased in such that a new one (of each) becomes available every year.

Let's say that the architect firm designing both office and residential towers are trying to use the same HVAC systems in both.

Let's say the tower owners for both are responsible for HVAC maintenance, and conditioning common spaces and providing inputs from the various tenants into the overall HVAC system.

Would a fuzzy logic control system be better than a PLC based control system?  I have a sense that for the first year, the PLC based system would indeed be better, but if a fuzzy logic system were installed, after the first year, it would save considerable expense throughout the useful life of the system, with the first year being the "learning" mode for the FLC, and mostly manually controlled by a team of people.  After year 1, the FLC will replace the fleet of people and handle it all automatically.

Is this remotely close to reality with FLC vs PLC controls for such a contrived system?  If so, wouldn't the FLC scale so that once it was integrated and trained, that data could be transferred to all FLCs, and the only tuning would be for the first one, and that the cost for that fleet of people training the FLC for a year could then be spread across all the future towers.  That would make the expense of FLC well worth it in the long run.

If they stopped after a couple years, like another pandemic with new rules forcing a complete HVAC reworking, all that cost would be staring them in the face, and the next round of HVACs would NOT be FLC no matter the potential long-term benefits.

hdhosseini

Good health to you Mr. Hosseini.

Throwing a stone in the bush by playing devil's advocate --

May I assume that a strong background in fuzzy logic theory is required to make such algorithms work correctly? Or could any one just jump in and tweak values until something happens?

 I agree that Fuzzy control is suitable for Complex applications. In process automation , you can see multi-variable and non-linear models and mostly their products are expensive. You can use extra software packages in parallel for optimization by any method you want to ensure higher quality and more profit. It is normal to see such advanced control features in PCS 7 or Emerson DeltaV because of the reasons mentioned.

For a simple application in a factory automation , most of the time conventional methods working perfectly.