How to tidy string data that doesn't meet a variable pattern

I'm looking to solve a problem I encounter on a daily basis at work. I have to convert a bill=of-materials into usable data. These can sometimes extend to hundreds of lines of data per BOM, so it's a tiresome task and open to mistakes. Under the instructions header (see attached sample) are varied strings, of which (due to the good people on this forum) have been sorted out, stripping out the alpha character, cleaning the numerical data, and then adding the alpha character back in behind each numerical character again.

What I'm trying to achieve is highlighting any sub-strings that interfere with the above process, so that I can spot them easily and remove them manually. The problem lies in that each row will contain various differing prefix characters, and the string sections that do not fit the typical pattern may consist of instruction mistakes (an additional character) or a line of text after the initial string patterns duplicates between different lines, or even in the same string occasionally happens too. Any help or guidance would be hugely appreciated