In general, cutting tools account for about 3 to 5 percent of total manufacturing cost. By itself, this level of expense may not be enough to attract the attention of top management at major manufacturing companies. However, activity-based costing has tended to reveal that the cost involved in the sourcing and replenishment process can exceed the purchase value of tools—an insight that has led many manufacturers to switch to full-service supply. Also, there is the painful irony that even though $50,000 worth of cutting tools may be tied up per manufacturing line or flexible cell, stock-outs are still a recurring problem. Many facilities have looked to shopfloor tool management to help address this problem.
But there is only so much that can be done at the shopfloor level. To look to the shop floor for the solution to these problems is to take for granted the tool as it was released from engineering, and treat tool management as merely a logistics and delivery challenge. The flaw in this thinking is revealed when one considers the sources of the inefficiencies that the shop floor confronts. These sources include . . .
- The large number of distinct components the shop floor must manage.
- The variety of obviously redundant tooling from multiple sources.
- The differing tool layouts.
- The error-prone nature of entering tool data into shopfloor systems.
What this list suggests is that many of the decisions affecting the difficulty of tool management have already been made before any physical tool comes into play. In fact, close to 70 percent of the difficulty in shopfloor tool management is created in engineering. Here the tool layout is fixed, and so are the supplier and the operation. Thus 70 percent of the life cycle cost of the tooling is fixed as well.
A concept called integrated tool management allows for the fact that a cutting tool, over the course of its lifetime, changes from an engineering item to a logistics item. During engineering, tooling is an area for process innovation. During the manufacturing cycle, tooling becomes a productivity driver for the machine tools.
Both of these separate realms can optimize their own flow of information. However, these areas typically don't view one another as customers. Engineering has no means to provide value-added service to manufacturing, and manufacturing has no means to channel its experience and requirements back to engineering. Providing these means—and therefore the integration—is the mission of integrated tool management.
Usually the tool layout is the handover document transferred from engineering to the shop floor. A tool layout captures the tool information in the language of engineering, consisting of drawings, bills of material and parameter lists. A single tool layout refers to a single tool assembly for a certain operation performed with a specific spindle on a specific machine tool. The layout documents the components of the tool assembly, including spare parts. On average, 30 to 50 tool assemblies are assigned to a machine tool in engine manufacture, with each assembly including some 230 to 350 components. Tooling engineers undertake great effort to document all of these components, and yet the effort is often of little value for downstream activities.
From the tool layout, machine operators and purchasing personnel (or a full service supplier) pick relevant information for their own downstream systems. Most of the information needed for these systems is not available in digital form, so information has to be obtained and keyed in at various stages. Here are some examples:
- Procurement generates tool packages—that is, bills of material used to generate purchase orders.
- Shopfloor tool management assigns storage IDs to populate the inventory management system. Also, tool management adds the distinctions of perishable versus durable tool components and returnable versus non-returnable tooling.
- The tool crib physically assembles the tool according to the tool layout and performs presetting. Correction values are sent to the NC control.
- The tool crib inspects returned tool assemblies and generates failure reports.
- On-site cost reduction teams improve cycle times—changing speed and feed rate or calling for alternate tooling—and thereby change the tool specification, creating the need for a new release from engineering.
Now consider what effort is involved in these activities. Lost time can be attributed to these factors:
- Tool search and legacy case studies. Including communication with tool suppliers and machine tool companies, this work alone can consume up to 50 percent of the tooling group's time.
- Generating tooling packages. This takes approximately 3 days per package, with each package addressing one particular spindle.
- Populating downstream systems. At a rate of about 2 minutes per item, this translates to about 500 hours for the average of 15,000 tool items per transmission plant. For an engine plant, the figure is approximately 200 hours.
Thus the potential savings that integrated tool management can realize are significant. And these savings do not even include the savings from higher quality engineering processes. They also don't include the savings from "deproliferation" programs that reduce the number of tool items in use.
So why do so few companies practice integrated tool management? These circumstances, which are typical in large-scale manufacturing operations, serve as the most frequently cited reasons:
- No commonality in tool descriptions.
- Documentation divided between different systems.
- Graphics and tool data not in digital format.
- No digital information about tool performance.
- No information about which tools are actually in use.
Compounding these obstacles, collabora-tion with outside suppliers is difficult because, in general, there is no electronic supplier integration beyond electronic data interchange links with major suppliers. Also, no communication standards are available for collaboration with engineering partners such as cutting tool and machine tool companies.
These difficulties apply even in the rare situation where the engineering cycle is supported through effective electronic data management and the supply cycle is supported through effective shopfloor tool management. That's because engineering pulls together all relevant information for a machining process and consolidates this into a single tool assembly, whereas purchasing must disintegrate the tool assembly to create procurement packages and populate the downstream systems with data. In addition, shopfloor management requirements may differ in different locations, even within the same company.
The situation at the powertrain operations of a major automobile manufacturer reveals the scope of the problem. Here are the key details:
- Five engine and two transmission plants consume an annual average of $200 million worth of cutting tools, sourced from approximately 150 suppliers (excluding occasional sources).
- Shopfloor tool management is each plant's responsibility, dealing with 5,000 to 18,000 tool items in each plant.
- Different full-service suppliers are in charge of insourcing, replenishment and cutter grind at each location, their incomes depending largely on proven cost savings.
- Process engineering is centralized, but production processes are supported by local tool engineers.
- Advanced manufacturing is part of the company's tech center.
A preliminary study quantified the tool-management problem. While work related to documentation accounted for 60 percent of tooling engineers' time, shopfloor tool managers and full-service suppliers alike testified that no useful information on tooling was available. Also, 25 percent of the time on the shop floor was spent on tool searches, mainly to establish cross references from supplier IDs to the company's IDs. And a random search at one of the transmission plants showed 28 percent of tool items are either obsolete (not used for the past 24 months) or redundant (one of several tool items for the same application). All of other plants confirmed that this figure was representative for their sites as well.
Management decided to take action at the root of the problem. The company decided that a system should be put in place that would allow personnel to reduce the number of tool items in use. The company also decided that a knowledge base should be established as a communication platform for future best-practice studies and to promote common engineering standards.
One other requirement was that the system had to reflect the corporate structure of centralized engineering and purchasing with decentralized plant operations. No group should be forced to adopt another group's perspective on the overall process, a requirement that relates most significantly to the underlying classification structures. If an engine plant uses an item called "mini-drill," for example, then another plant might refer to the same product as "micro-drill," and the system needed to allow for both designations to be correct. From the start, the company's project team concluded that the push for a classification scheme that offered only one perspective would lead to a fatal level of resistance.
The system the company put in place uses Cimsource's corporate software "CS-Enterprise." At the core of this software is a relational database interfaced to shopfloor tool management and shopfloor requisitioning. For process engineering, a direct link to the drawing management system was established. A browser-based interactive user interface was adapted to each user group's priorities. Import profiles were designed so they could be tuned to different suppliers' content, allowing the database to be populated automatically—whether from the supplier directly, or from Cimsource's "ToolsUnited" master server, which stores data covering the product ranges of various tool suppliers.
Two important features led to the successful implementation:
1. The complete separation of content from applications.
2. The establishment of a master classification scheme that could serve as a central reference for the different views of the system seen by different users.
The system was implemented on an existing Oracle database, making it possible for different views to be cross-referenced, as in the case of supplier catalogs cross-referenced with internal product codes. In the application view, cross-referencing allowed engineers and plant personnel alike to search the database by application. This capability was pivotal in the attempt to deproliferate the tools in use.
Such deproliferation in the past was compared to a visit to the hairdresser—you had to do it once in a while. Now it's an ongoing process. A tooling engineer with a machining problem can specify the problem in a search through the internal database. If no match is found, the same request goes out to the ToolsUnited server in search of standard tools to solve the problem. If there still is no match, only then can a new tool be created. This is a dramatic departure from the seemingly unbounded creativity that led to the high level of tool proliferation in the first place.
Deproliferation is also driven by purchasing. Because tools can be identified by application, the buyer gains engineering intelligence that was previously unavailable. Redundant tooling becomes obvious. As a result, purchasing can blind these tools from the database and wait for someone to complain. When no one complains (the usual case), the tools can be removed from the database. Indeed, in a typical manufacturing facility, nearly 10 percent of all listed items can be deleted in this way within the first 4 months of the system's life.
Best-practice benchmarking also becomes more feasible. A plant manager can access the database via the plant layouts, identifying a certain transfer line or flexible cell. From there, the manager can drill down to the process plan and tool layout, comparing a tool's specification with any similar line item. Test results from advanced manufacturing are also published to the system, and in addition, actual speeds and feed rates can be fed back into the system from the shop floor.
Lessons Learned: The Parametric Master Classification
Integrated tool management focuses on these goals:
- It offers different views of the cutting tool that are suited to the different perspectives of the various stakeholders in a tool's life cycle.
- It eases the flow of information through all of the stakeholders' IT systems.
How these goals are accomplished merits some discussion.
The "different views" objective essentially means that machining information and tool specifications have to be presented in different "languages." While the drawing is the language of the engineer, a buyer would rather refer to a catalog. While the NC programmer is comfortable with a CAD model, the machine operator wants a printed tool layout. Everyone accesses the same body of information in different ways. Separating content from application is the key to meeting this "multi-language" challenge.
Parametrics provide the way to achieve that separation. A cutting tool is described using all of the parameters necessary to meet the various management applications along its life cycle. Each application then accesses only the relevant parameters to populate its predefined models, templates or tables. This technique is common practice in the management of standard parts within CAD systems, and any management information system (MIS) works the same way.
The "easy flow of information" objective is another matter involving translation. This objective requires that different applications' exporting and importing activities refer to a common language. Basic data standards such as ASCII or CSV provide part of the solution, but meaningful information using one of these standards also has to have a standard structure. This structure can be predefined between the systems to be connected, but it can also be part of the transferred data, as in Extensible Markup Language (XML). The e-procurement "standards" of ARIBA and CommerceOne take the latter approach, using XML with a predefined structure to populate their respective procurement platforms.
A classification scheme addresses both structure and parametrics. However, the problem with this approach is that each company has its own classification scheme, so mappings from suppliers' classifications to their customers' classification are inevitable. What's more, large or multinational companies may maintain different classification schemes in different locations. The only practical way to address the discrepancy is with a master classification scheme able to be mapped to different views.
For the metalcutting industry, a consortium of tool manufacturers and their customers has worked to define a master classification scheme that includes requisite parameters for standard tools. The result is an industry standard for tool descriptions. This standard, called "StandardOpenBase," is promoted through a joint venture involving Kennametal, Sandvik, Widia, Plansee Tizit and Cimsource.
StandardOpenBase is now used by a large number of cutting tool companies. That means there are a large number of suppliers now prepared to quickly populate a newly implemented system for integrated tool management, allowing the system to begin delivering its payback that much more quickly.
About the author: Goetz Marczinski is CEO of Cimsource Software Company (Ann Arbor, Michigan).