In a recent talk with a colleague, he mentioned that in Chemical Processes it is not worthy to perform frequency analyses at values higher that 100 Hz, is this true? if so, why? is there any reference for that?
The answer may depend on the scale of the process in question: is it industry scale or test tube scale?
If it is industry scale you may have to consider a kind of "coherence length", i.e. the correlated interaction between chemical processes on one side of a reactor with those on the other side.
For the most part I would agree with this statement, however, I would add that there is no clear cut rule on what frequencies should be considered and where the cut-off is for stopping the analysis. The reason for this is that the relevant frequencies depend upon the time constants of the process and the accuracy of the model.
Let me briefly explain: if you are dealing with a distillation column and the input-output behavior under in investigation has a time constant of 30 minnutes, then you likely do not include some faster phenomena, such as opening/closing of valves, etc, in your model. As such, your model will not be accurate at higher frequencies due to information that is commonly (and for all practical purposes correctly) ignored. However, opening and closing a value may have a time constant of a few seconds to a few minutes. As such, your model will not be accurate at those frequencies or frequencies that are even higher. 100Hz is a lot faster than the dynamics that got ignored in what I outlined. Your analysis should likely stop at even lower frequencies.
Now all of that being said, if you are dealing with very fast processes (not very common in industrial scale chemical processes, but some exist) then you may want to consider carrying out your analysis to higher frequencies. In short: your analysis should be appropriate for the time scale of the process and the model that you use to describe the process.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory