Digital Twin (DT) implementations can reduce latency drastically in future communication systems for steering and control of cyber-physical systems (CPS). When the CPS and its DT periodically exchange information to repair possible desynchronisations between the two systems, this can lead to large data traffic flooding the networks. This work proposes a goal-oriented communication approach based on message identification (ID), which only sends repair data when necessary, but adds control traffic and a probability of not repairing some desynchronisations between the two systems. We compare optimal ID codes and a hash function for detection whether desynchronisation correction is necessary. Our method can reduce the traffic to within 0.3% of its optimal value while repairing 99.97% of all desynchronisations for 4 kbit of data and grows even better for larger data.