Flow Analysis and Optimization in Machining Lines - Analytical Models and Applications to a Real Case

The present master thesis work originates from the analysis of a real engineblock transfer line sited in a multinational automotive company. Beside the manufacturing system performance evaluation and improvement, a new analytical model for performance evaluation of two-machine lines, characterized by deterministic processing times, finite buffer capacity and Time Dependent Failures (TDFs) is developed. In particular the analysis is presented both for the single and the multiple failure modes case. In the latter case the closed solution is derived only for two-machine lines and we allow the machines to be affected by two different failure modes, the first one operation-dependent and the second one time-dependent. In literature no
analytical solution is presented to account for such a situation, but only for lines characterized either by different operation-dependent failures or by one single time-dependent failure. Nowadays, especially in long transfer lines working 24 hours per day, TDFs are assuming such an importance that it cannot be neglected during the system modeling, in order to derive more and more accurate estimations. In fact, being the machines made of components
which can fail even if the stations are not processing a part, it is useful to take into account also these kinds of failures because of, depending on their occurrence frequency, they can sensibly affect the system performance. Model consistency and quantitative comparisons with other remarkable models existing in literature are also shown.
With regard to the real industrial case, the line is analyzed following three main steps. At first machine failure parameters are derived through an opportunely built Matlab algorithm operating on shop floor data stored by the line supervisor in the system database. Secondly, the system performances are evaluated through a well-known approximate analytical method,
also discussed in the literature overview. The machine disruptions mainly affecting the performance of the line are detected through a detailed analysis of the evaluation results. The last step is represented by the line performance improvement: three different improvement options are considered and the increase in throughput is quantify. In particular a buffer capacity optimization, a time to repair reduction study and a repair crew optimization are proposed.
All the methods and the suggested solutions have been shown to the Scania’s research environment with successful feedbacks from all the staff: in such a way their production engineers have the opportunity to eventually implement one or more improvement action, given their internal trade-off structure.