I've been having a bit of a play with this as well
Rather than having to guess / copy-paste the troublesome UTF-8 characters, I wonder if it might work to replace essentially the whole non-ASCII unicode range? it seems to work in 'sed'
I can't make it work in 'rename' though - maybe one of the perl folks on here can figure it out?
$ > a$'\u2215'funny$'\u2215'file
$ > another$'\u2215'funny$'\u2215'file
$ for file in *; do echo mv "$file" "$(sed s/[^U+0000-U+007f]/-/g <<< "$file")"; done
mv a∕funny∕file a-funny-file
mv another∕funny∕file another-funny-file
$ rename -nv 's/[^U+0000-U+007f]/-/g' *
a∕funny∕file renamed as ----f-------f---
another∕funny∕file renamed as ----------f-------f---