Performance analysis of manufacturing systems composed of modular machines using the universal generating function

Abstract

The modularity and reconfigurability of the building blocks of modern manufacturing systems have to be considered when evaluating their performance. This paper proposes a model for evaluating system availability and expected production rates for manufacturing systems that are composed of unreliable modular machines with multiple functionally parallel production units. These units are treated as independent modules, where the breakdown/stoppage of one unit does not necessitate the failure/stoppage of the whole machine and its production. The considered systems are multi-state manufacturing systems (MSMS) that can handle multiple parts simultaneously, and their structure is that of nonbuffered flow lines allowing paralleling of identical multi-state modular machines (MSMM) in each production stage. In spite of the inherent computational complexity of the proposed analysis, due to the large number of system states, it was made possible by the use of the universal generating function (UGF) technique, which proved efficient for large MSMS. The proposed model was applied to a number of case studies for demonstration and verification. The case studies were based on a family of engine front covers. The results show that machines with a larger number of modules, usually thought of as having lower availability, provided higher overall system availability in the case of machines with multiple spindles. Based on the new analysis and results, it is recommended that system designers favorably consider machines with multiple spindles rather than increasing the number of machines in parallel. These results provide an important support for the use of modular/reconfigurable equipment compared with traditional equipment, in spite of the higher cost.