Kalman Filtering (as presented by Terry Blevins in 2013)

Emerson Community,

I have customer that is interested in trying out the Kalman filtering composite developed by Terry back in 2013. The files were shared on Applications Exchange, however, I get an error when trying to download. Would anyone be able to provide me the files from the entry below?

  • Hi, Brad. It looks like this link is active now: http://deltav.com/Appsnew/

    Ray
  • In reply to Ray Emerson:

    Thanks Ray. Yes, the site is up but, the download links result in errors. Also, it looks like my screenshot didn't load in my original question, if there is any confusion about the files I'm looking for, please let me know and I can walk through additional details.
  • In reply to Brad Bouchard:

    Ah, I see. There's a permissions conflict on the server (Error 500).

    Rachelle McWright - Do you know who to talk to about this on the Emerson side of the fence? Thanks very much!

  • In reply to Ray Emerson:

    Hi all~ I was waiting to see if a technical resource would chime in. My understanding is that they are sunsetting the applications site anad looking for a new home for the info. That MAY be a new permissions-only private group in EE365, but that is tbd. Will keep everyone apprised as I hear more....

    Best Regards,

    Rachelle McWright: Business Development Manager, Dynamic Simulation: U.S. Gulf Coast

  • In reply to Rachelle McWright:

    I just posted the DeltaV V12.3 configuration files for the Kalman Filter implementation as well as related presentations on the EE365 Media Gallery.  They can be accessed at the following link. 

    Let me know if you have any questions!

    James

  • In reply to James Beall:

    Awesome! Thanks James!
  • In reply to James Beall:

    Hi James - the calc block receives the PID block gain and reset when PROC_SEL=1 and then it transforms those inputs into process gain and process time constant to derive the first order discrete model of the process to be used by the discrete Kalman Filter. However, the PID gain and reset are the loop tuning parameters and not the process gain and time constant identified using for instance using InSight. The calc block calculation for process gain is 1/PID tuning gain and for process time constant = 1*RESET. Is this correct or perhaps the calculations in the calc block are just for a specific example?