Data collection and statistical data analysis in preparation for simulation of a furniture manufacturing company

Systems that are simulated are generally made up of one or more elements that have some uncertainty associated with them. These systems may eventually evolve in a manner that is not completely predictable and thus become stochastic in nature. Simulation of stochastic systems requires that the variability of the elements in the system be characterized by probability distributions or concepts. An 'As-Is-Analysis' of the plant layout and product process flows was carried out at a furniture manufacturing company. Process data for four of their main products namely, pallets, baby tenders bunk beds and standard coffins was collected using a specially designed data sheet. An analysis of the product flow times was carried out by grouping the data into four time variables namely, material movement, processing and waiting (idle) times before and after processing at the active workstation. The distributions of these variables were obtained using graphical methods in which smooth distribution curves were generated. The gamma distribution with shape parameters of α = 3 and α = 2 characterized the baby tenders while the product flow times for pallets were characterized by both the gamma distribution, α = 3 and the exponential distribution with the mean varying between 14.95 and 271.78 seconds. All the data and analysis carried out produced useful information for input to the design of experiments for simulation.