I've made occasional use of pdfSAM (Split And Merge) - the Lite version is free.
You cannot split HTML randomly and expect to do anything sensible with it - it needs to split on a boundary of some sort. Try using calibre's recently added calibre-debug --explode-book and --implode-book options, see ==>>
calibre-debug.
That should allow you to use standard html and other text editing tools.
BR