I used JENA to write into Model. Have faced typical UTF-8 wrongly encoded from Latin 1/ISO-8859-1, e.g., "Ã©" instead of "é".
Are there any parameters I need to tweak in Jena? Or should I consider another library?
if you run into BAD IRI warnings, you can ignore those.
If you want to try a different library, try RDF4j. I prefer the later to pass triples around or hold them in a simple container model. It is less complicated an offers nice Interfaces to hook into proceedings. JENA is more sophisticated when it comes to ontologies and querying.