If you have carried out similar scientific experiments in the past, i.e. if you have used the same methods in previous studies, you can involuntarily plagiarize yourself in the manuscript where you are using the same scientific method. It does not have to be copy-paste from another article, but if you have described the same method numerous times, then the wording can be linguistically so incorporated in your mind, that you involuntarily happen to write the same as you did in one of your previous papers.
It is important to realize, that you are not allowed to write the same text even though you have written it yourself in one of your previous papers. The text is actually owned by the publisher, which owns the journal and you are therefore not the owner of your own sentences so to speak. An exception of course is if your previous paper was published in an open access journal, then you are allowed to use the same sentences if you just give reference to your original article where it first was used. If you, however, publish your study in a non-open access journal, then as mentioned, the publisher will own your language.
Of course there is a pain threshold at which it becomes too much for the publisher and this limit is not a fixed number. When they run your article through plagiarism software, then some wordings will of course have been used before in the entire scientific literature. There is no official limits to how much of your paper can be plagiarism, but in editor-circles we tend to talk about something in the vicinity of 20-30% of the paper can be language that has been used before. This is only if it is sentences here and there in the manuscript, whereas if full paragraphs have been copied, then it is a totally different scenario.
Instead of using the same formulations as you did in one of your previous papers it is advantageous to just refer to earlier works where the method is described in detail.