All of the errors shown relate to the USR/Nilan path.
Yes. Error handling is time consuming and does just that.
Alright, so what’s gone wrong?
Modbus-TCP is the usual Modbus master-slave, query-response stuff.
Each query generated by openHAB is given a unique ID.
Each response (from the same target IP) is checked to see if has a matching ID.
No match, it’s garbage to be rejected.
What can go wrong - maybe the message is truly all garbage, maybe the target is just horribly out of sync in a sequence.
I would first check in PaperUI to see you have your expected number of TCP Bridge things,in case you’ve managed to generate a duplicate somehow and their messages are getting cross-wired.
If that all looks good, give the USR a kick in the pants and reboot it.