Is “corrosion” an instance of onomatopoeia?

I mean, obviously "corrosion" isn't actually onomatopoeic, because corrosion doesn't make a sound (or at least not one that humans can hear). Yet it seems to me that the word corrosion sounds like its meaning, in some hard-to-define way.

Is there a term for words that suggest their meanings, without necessarily recreating the sound, but that aren't quite onomatopoeic? There are a lot of similar ones, like grind, shuffle.

Words whose sounds refer to, suggest, or otherwise are associated with a particular meaning are cases of sound symbolism. Although onomatopoeia - direct imitation of a real-world sound - is one type of sound symbolism, it is not the only one.

A common sound symbolism is sound iconism. With the related clustering, this is the re-use of sounds across a set of words with related meanings. Note that this is the re-use of sounds, not morphemes. One example is found in the set of words stamp, stomp, tamp, tromp, and tramp which have the common /-mp/. The final /-mp/ sound is strongly suggestive of stomping or stamping, though it's actually not imitative of it.

Another set of words is glisten, gleam, glint, glare, glam, glimmer, glaze, glass, glitz, gloss, glory, glow, and glitter. The /gl-/ is associated with shining, though it's not imitative in any way since shining is visual, not auditory. Nevertheless English speakers hear these words as related.

This is not unique to English of course. For example, in Japanese linguistics one finds the terms phenomime and psychomime for similar phenomena.

I don't believe corrosion has any (historical or current) onomatopoeic qualities whatsover. But the question as now rephrased addresses the more general case covered by John Lawler's answer here...

Phonosemantics, or sound symbolism.

Per that Wikipedia link, onomatopoeia is the least significant type of symbolism. But it's very difficult for the layman to have a meaningful opinion on whether or not any given word has phonosemantic qualities. Once you know a word, you'll always associate the meaning with the sound - but it's not easy to say which came first.

There's also the well-known Bouba/kiki effect whereby the majority of speakers of any language tend to associate bouba with a "rounded" shape, and kiki with a "spiky" one. Personally, I doubt there are many such examples. But I do think neologisms such as cruft have more chance of gaining/retaining currency if they sound like words with related meanings (crud, crust, crude, for example, in that case).