Abstract
This paper is aimed at formalizing a model for hybrid production systems where the interactions between the continuous process parts and the manufacturing sub-systems are given by minor stoppages. The proposal is to represent the effects of the continuous process dynamics on discrete manufacturing sub-systems using Autoregressive Conditional Duration (ACD) models originally conceived to treat high-frequency and irregularly spaced financial transaction data. The proposed methodology has been tested on a real-life fibre-glass production plant exploiting simulation techniques. The physical model of the furnace and spooling-bushing de-partment has been run in two different conditions: (i) the fibre-glass breakings, i.e. minor stoppages, on the spooling-bushing machines are generated according to the proposed approach; (ii) historical fibre-glass breakings data are used. The compari-son between the simulation results shows that for three spooling-bushing machines only out of 24, the daily throughput of completed spools and the empirical distribu-tions for the uncompleted spools weight are dissimilar under the two different con-ditions. Therefore, the ACD-based model proved useful for representing the occur-rence of fibre-glass breakings on the spooling-bushing machines and, in more gen-eral terms, for the logical modelling of the hybrid production systems, where the re-lationships between continuous and discrete parts are given by minor stoppages.
