Someone else will need to comment on why that's causing Calibre to cause the epub to fail the flightcrew check, but you can prevent Calibre from splitting the file by enabling 'do not split on page breaks' under epub output.
Also delete the xpath for 'insert page breaks before' under structure detection, and if you already have a TOC you're happy with then you can also delete the chapter detection xpath - If you want Calibre to re-build a TOC but not split the file set the Chapter detection to not insert a page break.
All that advice is predicated on you having a backup of the original epub, if you don't you'll need to use Sigil to re-merge the file. Sigil will rebuild the TOC based on the file's <h1> tags, so you can safely remove the <a> tag Calibre inserted, just delete this part:
Code:
<a id=”anchor-14-anchor” class=”calibre3”></a>