There aren't any duplicates here, and nothing that needs fixing.

There is one row for each partition. It contains an object_id which is teh object_id of the table of which this partion is part. So if a table is partitioned into 3 partitions, sys.partitions contains 3 rows which have that table's object id in them - one for each of that table's three partitions.