Using discrete event simulation for scheduling and long range capacity planning of a high volume press shop

Abstract

The thesis expresses the essential requirement for and the use of Discrete Event Simulation (DES) in a high volume press shop. The press shop produces blanks and panels for the body shop, which manufactures three car models. DES is used to combat the battle between shop efficiency and low inventory. The process used to choose the most appropriate software package is described and then current situation in the press shop is discussed. The procedures involved in model creation follow set model construction guidelines. There are several assumptions made, which together with the constraints of the system, provide the limitations of the inputs facing the system. There is a trade off between model complexity and accuracy, so the setting of the constraints and assumptions often provided difficult decisions. Validation of the model is very important, so this was a lengthy process, involving using a series of dummy buffers to check inputs such as cycle times and batch quantities. The validated model is used to monitor the methods used to reduce inventory on the shop floor over a period of eight weeks and then used for 'What If? Scenarios, to ascertain the systems capacity and inventory levels underdifferent conditions. The scenarios include using volumes that are 100% higher on some models than the current situation and 20% less than currently. The findings are examined and proposals made for the introduction of the proposed volumes where possible. Findings of the scenarios highlight bottlenecks in the shop and areas for improvement. Using the model, the schedules can be changed quickly and easily to try and eliminate the bottlenecks and improve capacity. Conclusions discuss the problems encountered during the modelling process as well as the benefits. The integration of DES into the current scheduling processes in the shop poses no problems and the model will be used as an aid for capacity planning in the future.