Process Engineering Gets a Digital Boost
Linking software systems captures best practices and speeds process design
By Bruce Morey
Remmele Engineering (New Brighton, MN), a $93 million/year business with four plants in Minnesota, specializes in low-volume parts that require a high degree of manufacturing expertise. Designing the most cost-effective, high-quality manufacturing process in as short a time as possible is vital. "Lead times are becoming weeks where before they were months," says Red Heitkamp, director of advanced manufacturing engineering at Remmele.
Improving process design required standardizing the company's digital foundation of manufacturing software and data. The company has completed the first phase of this program. Just three years ago, Remmele was using about a dozen different CAD/CAM systems and several different DNC systems, each requiring dedicated personnel for that particular system. Moreover, these CAD/CAM systems were focused on a particular industry (for instance one for aerospace and one for defense). If an order came in from another industry, it was difficult for Remmele to make use of all of its facilities, because of the CAM systems' incompatibility. Even the simple matter of knowing what milling or cutting tools were available and where they were located was, in reality, not so simple. "While it's important to continue improving machine technology, we feel that there is much to be gained from applying automation to the front end of our process," says Heitkamp.
Besides streamlining close to a dozen different CAD/CAM systems, the improvements have simplified the work of manufacturing engineers. What Remmele has done goes beyond standardization—they have integrated the software and incorporated machine-specific information to reduce the process-design workload.
Tool management alone has reduced the number of manual steps an engineer must perform. Like many contract manufacturers, Remmele has hundreds of unique tools, including drills, reamers, end mills, taps, face mills, slitting saws, collets, holders, arbors, and extensions. Remmele created a database for these tools, and introduced procedures to keep track of them. Now, the company's manufacturing engineers know what tools are available and where they are.
Tool management means more than knowing what and where. It also means knowing what the tool can do. The manufacturing engineer and CAM programmer now have, for certain machines, accurate cutter definitions, consistent feeds and speeds, and accuracy-simulation graphics for each tool—on-line at their desks—during the design process. Because this information is autoloaded, 50% of the manual steps that were part of the CAM programmer's job are eliminated.
Improved change-control procedures are another benefit of Remmele's digital foundation. For example, the company received a large (approximately 2 x 8 m) complex part in Catia V5 from a customer who then—after Remmele began designing its manufacturing process—requested several revisions. Identifying the changes to its process caused by these engineering design changes was easy, even after programming hundreds of cutter paths. "For cases like this, the amount of time to process a revision change used to be eight hours and is now 15–20 min," says Mark Conley, CAM/EDI center manager.
To achieve this digital foundation, Remmele needed common software. Then the company required common data, processes, and procedures to achieve the savings cited above.
Five basic software toolsets were identified as standards for their corporate-wide system.
- Computer Aided Manufacturing (CAM),
- Performance prediction software,
- Tool data management (TDM),
- Engineering data management (EDM), and
- Distributed numerical control (DNC).
The CAM product they finally chose is NX from UGS (Plano, TX). Choosing a single CAM product was a challenge. Although Remmele wanted to standardize on one common product definition description, it also had to recognize the diversity of its customers. "We deal with so many industries; we don't have the luxury of being picky about what kind of models we can receive," says Heitkamp, "we have to take them all."
Customers provide models to Remmele in Catia V4 format, PTC, and SolidWorks, among others. Remmele needs to translate multiple definitions into a single master.
Fortunately, software from Translation Technologies (Spokane, WA) worked well for them. "This is really a remastering package that preserves the intelligence in the model provided by our customers," explains Conley. "It saves us time and effort in not having to re-enter that intelligence manually, and goes beyond a nominal STEP conversion."
Remastering creates a universal smart model in UGS NX format that preserves parametric data, features, or GD&T that may have been present in the original.
Future growth was another reason to choose UGS NX. "UGS NX has an open architecture," says Conley, "it lends itself to customization and automation." NX has an application programmer's interface (API) that allows users to include C++, Java, and Visual Basic (VB.net).
Remmele engineers chose two performance simulation packages: Vericut by CGTech Corp. (Irvine, CA) and Metal Max by Manufacturing Laboratories Inc. (Las Vegas, NV). "We use CGTech's Vericut for all of our machining simulation needs, including Vericut machine and control files," says Conley. "With data exported from NX, Vericut simulates the cutting process on the shop floor utilizing the machine G-code to detect any issues prior to NC Program release." Remmele uses Metal Max, an accompanying performance analysis software package, to perform modal analysis to predict and eliminate chatter and other unwanted vibration. It provides recommendations such as surface footages and depth of cut in the material specified.
TDM Information Systems (Schaumburg, IL) provided the tool management function. This package is able to integrate with UGS NX and NX's simulation tools. Automated tool management alone was good; automating data transfer to CAM automation made it even better. "We needed a tool management system that was robust enough to talk with CAM automation," says Tom Shuga, CIM Manager for Remmele Engineering. TDM provided that.
Based on an Oracle platform, the engineering database management software is Remmele's own. Although there are systems such as Teamcenter from UGS or SmartTeam from Dassault available, Remmele chose to keep its own proprietary databasing software. It configures the master of the part data received from the customer, and shares the data with the plant. The data are synchronized nightly at each plant by the software to ensure data validity. Finally, they chose Predator DNC (Portland, OR) to populate machines with CAM data anywhere in Remmele.
Remmele has integrated these software packages so that the tool assemblies captured in TDM are imported to NX for CAM programming. Product data and CAM configurations are configured, controlled, and distributed on its single DNC system.
"With these systems, we are able to have configuration control of the engineering data at the start of the process," explains Shuga.
Purchasing a tool inventory system like TDM was only one step—populating it and establishing procedures for its use was critical. Remmele stores its tools in its TDM as virtual tool assemblies, rather than as individual tools. "Virtual tool assemblies have data tied to them, like build feeds and speeds for a particular material type and different processes. When a CAM programmer selects a tool assembly from the system, he pulls in not only geometric data, but also tool feeds and speeds," explains Shuga.
Remmele's tool-inventory-control system gives manufacturing engineers access to tools not only in their own plants, but also corporate-wide. It provides visibility to all crib inventories, receives order recommendations from all cribs, makes order decisions, sends order requests, and even faxes blanket orders to suppliers.
Technical data for each tool assembly stored in TDM need to come from somewhere. Remmele decided that calibrating the software to the reality of its shop floor, rather than using generic data, was vital to success. Each machine will be measured in terms of accuracy, feed and speed, rigidity, and control-technology limitations. Remmele currently has 75 premium machines that it will characterize.
Currently, Remmele has 13,200 individual tool components and 10,390 tool assemblies loaded into the TDM database. So far, manufacturing engineers and NC programmers have populated about 10% of those assemblies with data derived from machine performance and process-specific information.
Remmele takes advantage of both software analysis and actual testing for machine characterization. In addition, The Metal Max software provides simulated cutting-tool performance analysis. Making the on-machine testing meaningful for future use means keeping strict control of cutting-tool assembly configuration and inventories. In addition, Remmele incorporates manufacturer testing recommendations and shop preferences into the eventual machine calibration. This calibration provides TDM with technology data, like the speeds and feeds preferred for specific materials.
"We do this to find out what the machine can do through actual testing," observes Conley, "and incorporate that information into the database."
The mixed blessing of a world always in flux represents both opportunity and risk for companies like Remmele and its manufacturing engineers. Continued involvement from its own manufacturing personnel is vital. "The manufacturing engineers and shop-floor manufacturing experts" involvement is vital to provide input into all automation projects," says Conley.
The process of developing the knowledge foundation is complex but necessary for knowledge-driven processes to be useful, says Heitkamp. He believes manufacturing engineers will benefit from the knowledge base, which enables consistent process definition for engineers and NC programmers that takes advantage of best standard practices within the manufacturing environment.
Keeping the foundation database up to date will represent a challenge as new machines are acquired or new materials specified. In addition, the software will continuously advance as software vendors seek to improve it. Software upgrades also represent opportunity, as Remmele will find new ways to provide process design aids for its manufacturing engineers, and the engineers find ways to use the system and improve it.
Automating Process Design
Remmele has just finished standardizing its software corporate-wide, and doing so has yielded significant benefits. The next step is to prepare Remmele for true automation. The ultimate vision is to implement a knowledge-based automated engineering system they call the Intelligent Model-Centric Manufacturing System.
Variations in the manufacturing design process cause problems. Variations in how the same parts are made in different plants or by different people are especially troubling. "It's typical today for it to take several iterations to get to a good part," remarks Red Heitkamp. "We need a method to reduce schedule risk and provide a more predictable process.
"The advantage to us of building a system to capture this knowledge is that it is something you can't buy off the street," he explains. "UGS NX or TDM you can buy, the proven best practices we put into the system cannot be purchased."
The company benchmarked its processes and built a roadmap to reduce its lead-time by 50%. They plan to do this by capturing best practices in a rules-based decision tree. When a manufacturing engineer is presented with a part model from a customer, a manufacturing procedure will be established and autopopulated. Manufacturing experts in each plant will provide information on how to autopopulate the procedure.
In March 2007, Remmele will demonstrate the first use of their knowledge-based engineering initiative on a complex part. This type of part currently requires about 150 hr of programming and three weeks elapsed time to program. After implementing the knowledge system, engineers at Remmele believe they can reduce this time by 40%, delivering a program in less than two weeks.
This article was first published in the March 2007 edition of Manufacturing Engineering magazine.