If that pattern of text is always to be ignored in searches, then it can be ignored during indexing, using the regex token filter. Given that you’re looking for substrings within words, your custom analyzer using the regex token filter would need to start with the keyword tokenizer.