with the tidy ( brush icon) active, this will work
find <p Chapter
replace with <h2 Chapter
( so long as the word chapter does not also start a sentence anywhere - tidy will fix the trailing tag
this is safer - find replace set to regex + match case, all files - in code view
find <p Chapter ([\d+])</p>
replace <h2 Chapter \1 </h2>
the \d+ finds any number with 1 or more digits and the \1 in the replace syntax re-uses the value found in the find process.
if find replace complains that there are no matches, check very carefully for superfluous
whoops I see you wanted h3, not h2 - just amend the above examples
it is a good idea to click replace ( once) a few times to check that it is working, before running replace all.
variations of that syntax will fix up other types of chapter headings, but if you need to use a period ( a full stop) in regex you must place a backslash in front of it \.