Transition of interarrival time patterns between automated and manual configurations of assembly systems

The paradigm of hybrid automatic/manual assembly systems has recently been extended by reconfigurable systems that allow exchanging manual and automated production modules with minimum interruption of the system control and operation. However, both types of assembly systems exhibit different interarrival time patterns for layouts that employ asynchronous workpiece transport on tracks with multiple routes and loops. In this work, it has been shown analytically that for Gaussian processing time distributions the queue length follows heavy-tailed unimodal and bimodal distributions. Based on the resulting outlier probability, an analytical model for predicting the limits of allowable processing time variance for deterministic interarrival time models is proposed. This model is positively validated with simulation analysis by comparing different levels of processing time variance. It is concluded that established design practices such as cycle-time balancing and synchronization should be avoided in reconfigurable hybrid assembly system design because they introduce interarrival-time turbulence in the automated system.