Simulation of a Plant-Wide Inventory Pull System

Abstract

Successful implementation of inventory pull (kanban) systems has become a widely discussed topic in recent years. This paper examines one such effort which uses simulation modeling as a means to quantitatively design such a system. The Normalize area is responsible for beat treating raw material for Delco Chassis wheel spindle bearing products. Within the area, product travels in containers that eventually are delivered to various machining lines. Since its inception, the area has been scheduled with a traditional "push system." The result has been high inventory, long lead times, floorspace organization problems, and a lack of employee involvement. A Normalize pull system was proposed to provide for better control of product flow through the area. Furthermore, the pull system maintains a maximum inventory level of approximately two days for all part sizes at the machining lines. The simulation uses historical production data to replicate demand for Normalized product. By using actual production data, the Normalize environment is simulated as if the pull system had been in place for ten months in 1992 and 1993. The simulation's results confirmed the feasibility of the pull system. The pull system would have caused an acceptably low amount of starvation in the machining lines. With the pull system, plant-wide Normalized inventory can be reduce by approximately 48%. Furthermore, plant Industrial Engineers will be able to dedicate floorspace to properly store Normalized inventory.