I get what you mean. The database makes an automatic recover after a shutdown abort (implicit, using the redologs to apply the changes not written to disk already, once this is an assync process due to performance concerns). This process can take long, indeed, if there is too much data to be applied.
My recommendation is first of all, after your next start, consider increasing the number of DBWriters.
Some trick that may help: Mount the database, go to RMAN, open several parallel channels, issue “recover database”. It will apply the relogs in parallel. This may help on parallelizing the recover, eventually making it faster. But the recover can’t (or shouldn’t) be avoided.
Ideally, take your actions to ensure a faster recover on the next time. And try to avoid Aborts.