Digital Twins (DTs) often require maintenance throughout their life cycles, as their Physical Twin (PT) counterparts undergo maintenance and evolution. This necessitates software updates to the DT, but when should these updates be done? Updating the DT at the wrong time can lead to inconsistencies between the DT and PT, as well as failures and increased downtime.

This study investigates the practicality and usability of applying Semi-Markov Processes (SMPs) to represent the connections and state transitions of the DT-enabled system. SMPs can be used to calculate the probability that all components in the system are collectively in a safe state, that is, in a state where updating the DT would result in minimal disruption to the DT-enabled system’s operation.

We also discuss the limitations of the approach and the future work required to make it robust. Lastly, we present how SMPs can be used for an industrial concrete mixer as our case study, to remove the need for fixed maintenance intervals, which are costly.