|
...to meet required throughputs for different products.
We applied the method at Woodward Aircraft Engine Systems, Rockford,
Illinois, which produces high-precision components with long setup
times. Using the spreadsheet, plant supervisors can quickly analyze
several alternate scenarios. They have reduced WIP levels in some
cells to one third of their original values.
(Queues: networks. Industries: machinery.)
Companies are increasingly using lean manufacturing concepts to conduct their manufacturing operations. In general, these concepts succeed when applied to manufacturing organizations with stable demand and a small product mix. For such organizations, there is a well-developed methodology, based primarily on the Toyota Production System (Ohno 1988, Monden 1993). A key feature of lean manufacturing systems is their use of pull signals to trigger production. The pull mechanism is typically implemented using kanbans or the constant-work-in-process (CONWIP) protocol. These implementations place a cap on the work in process (WIP).
Placing a cap on WIP has a number of benefits. It reduces flow times, reduces variation, and improves quality (Hopp and Spearman 2000, Suri 1998). However, setting WIP levels is not straightforward. It is well known that pull systems set the WIP levels and observe throughput. Determining WIP levels that satisfy a set of throughput requirements is not a trivial task, especially for plants that manufacture high-variety, low-volume products. For such plants, a change in the WIP level for one product could dramatically affect the throughputs of other products that share common resources, particularly if these resources are potential bottlenecks. Long setup times may exacerbate the problem because they often lead to production in large batches, resulting in high levels of WIP.
Woodward Aircraft Engine Systems was aware of the benefits of placing a cap on WIP and had identified the CONWIP protocol as an effective means of promoting flow in the plant. Woodward is a leading producer of fuel-control systems and components for aircraft and industrial engines and turbines. Its headquarters are in Rockford, Illinois, and it serves a global market from locations worldwide. Woodward has about 3,600 employees. The company's products and services are used in power generation, oil and gas processing, rail, aviation, and marine markets in addition to many light and heavy industrial applications. Woodward Aircraft Engine Systems provides approximately 45 percent of the total company revenues.
The Rockford plant manufactures a large variety of products at low volumes, some as low as 100 per year. The management at Woodward approached the University of Tennessee for assistance in determining the correct WIP levels for the different products that would provide the desired throughputs. We addressed the problem Woodward faced. We developed an algorithm for setting the batch sizes (number of parts placed on a pallet) and the number of pallets for each product. The WIP level set for a product is then simply obtained as the batch (pallet) size multiplied by the number of pallets. Woodward has implemented the method, adopting a hybrid of the CONWIP protocol and a drum-buffer-rope protocol (Goldratt and Fox 1986) to release pallets into the cell using a simple visual method.
The plant at which we conducted our study has about 900 employees. In 2000, Woodward had converted a number of cells in this plant from a traditional push system to a pull system, using the CONWIP protocol to release parts into the cell. When it moved to the pull system, the supervisors typically continued to maintain WIP inventory at or slightly below the levels observed under the push system. Occasionally they found that the WIP level for a part type was too low and had to run overtime to compensate for the low production volumes. For other part types, they often found they were carrying too much inventory.
The supervisors were trying to use a formula often discussed in the context of pull systems to set WIP levels (Monden 1993). This formula assumes that the kanban cycle time, which is the time that it takes to replenish the raw material for the manufactured product from the moment it is requested, is known in advance. The WIP levels are calculated based on the assumed kanban cycle time, the demand for the product, and a safety factor that buffers the variation in demand and cycle time. The shop-floor supervisors, however, were reluctant to use the formula. They found it difficult to estimate cycle times because they were unable to determine congestion and queueing delays at the various machines in the cell. Furthermore, they did not know how to set the safety factors. Consequently, they did...
NOTE: All illustrations and photos
have been removed from this article.
More articles from Interfaces
The analytic hierarchy process at Dar Al-Hekma, Saudi Arabia., July 01, 2003
Looking for additional articles?
Search our database of over 3 million articles.
Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication
name or publication date.
About Goliath
Whether you're looking for sales prospects, competitive information, company
analysis or best practices in managing your organization,
Goliath can help you meet your business needs.
Our extensive business information databases empower business
professionals with both the breadth and depth of credible,
authoritative information they need to support their business
goals. Whether it be strategic planning, sales prospecting,
company research or defining management best practices -
Goliath is your leading source for accurate information.
|
