Working capital management : implementation of integrated inventory management solutions for raw materials

Abstract

In the Working Capital Management context, the industrial group Knauf Insulation launched the Cash for Growth project which consists in releasing sleeping funds in inventory to improve the profitability of the company. The studied plant, located in Visé, produces glass mineral wool and white blowing wool. This paper focuses on the inventory management of the raw materials. A distinction is made between the raw materials, which form the glass mineral wool mattress, and the direct materials which are applied on the mattress for product differentiations. Basic inventory management concepts are exposed: inventory holding cost, inventory control system and ordering policy, models for the reorder point and safety stock calculation and the well-known lot-sizing model: the Economic Order Quantity model. In the second part of the paper, how the theoretical elements can be applied to the case of a continuous manufacturing like Knauf Insulation is discussed. A modified version of the reorder point and safety stock model is proposed for the raw materials for which Excel tools are developed to control the inventory level and the planning process. Concerning the direct materials which are managed on SAP ERP, improvements in the parameters of the MRP are proposed: the automatic safety stock calculation is activated. For slow-moving items, a reorder point strategy is used with a special key to obtain a fixed reorder point. Implementation of the proposed solutions results in a decrease of 10% in the stock value, which represents a value of €250,000. Finally, a relationship between inventory management and a scheduling problem is established: changeovers between products consume materials and produce waste. This problem refers to a sequence-dependent setup times problem which need to be minimized. Besides the description of the current situation of the schedule construction, a literature review about the studied problem in a single-machine environment is presented. The problem needs further investigations before being developed and implemented as a tool to help the schedulers.