Did you model the drives to allow motor contribution through the drives? If they aren't regenerative drives and don't have bypasses or have bypasses that won't be used or can't be used without damaging the compressors or the process they are part of you could check or uncheck the boxes that allow motor contribution and see what that does. Probably will help some but not get you there. I looked up the noise filter and it says it is for applications using regenerative drives so you may be stuck if that is what you have. If you are allowing fault contribution through the drives, does the noise filter impedance have an effect on motor fault contribution? You'd have to model those to, to figure that out.
I don't see the sccr/withstand rating on the filter data sheet:
https://coselusa.com/wp-content/uploads ... _TSD-S.pdf but I do see that rating on this page:
https://www.cosel.com/en/product/noisefilter/TSD and here:
https://www.cosel.com/en/lp/lineup/appl ... index.html I spent enough time looking into it but it seems as though Cosel has had some of their filters tested in combination with some breakers and if they haven't the "assumed" value for that filter size is 10kA. Here is another link on this if you want to research further:
https://www.cosel.com/en/coselplus/technical/111Is this an existing installation or new design?
Reactors are another option.
Artificially increasing feeder length could be another option but I expect it would border on ridiculous.
Can you reduce the 2 sets of 750's to 600's or 500's or is voltage drop an issue? Your feeders are pretty long. Reducing those 750's to something smaller would help.
If this is a new project, could you locate B-2HA closer to the compressors and make it's feeder longer and the compressor feeders shorter? That may help or may make it worse. I don't know without modeling it.