Manufacturability Starts with Communication, and DFMA Can Help

https://www.flickr.com/photos/31274959

Manufacturing Engineering Magazine August 2018
Bruce Morey

Controlling cost and complexity starts in design. Capturing and applying experience in a disciplined manner is vital, for both design engineers and their manufacturing counterparts. In the mid-1980s, the industry was understanding that many designs were not only difficult to make, but debilitatingly complex. Boothroyd Dewhurst Inc., a pioneer in the field, was founded by two university professors who had been developing solutions for a related problem, Design for Assembly, or DFA. DFA is all about producing elegant designs with as few parts as possible that still satisfy customers.

What ultimately controls the cost of a product is part count. Fewer parts mean both a less complex end product and fewer parts flowing through the plant and supply chain. Thinking about cost control in this sense can be a ‘lightbulb’ moment. Engineers typically think of how to make individual parts as inexpensively as possible without considering the total cost savings from reducing part count, even if individual parts cost more.

From DFA grew Design of Manufacturing, or DFM, which turns back to thinking about cost and manufacturability of the part. We needed to understand the impact on cost of these more complex, fewer parts resulting from DFA. DFM methodology contains a series of about 30 manufacturing cost models to look at those design decisions. A design team now using integrated DFMA (Design for Manufacturing and Assembly) can potentially look at three, four, or five different but appropriate methods of producing parts.

We assign accurate and credible costs to the parts, using software that calculates manufacturing cycle time through embedded equations developed by the company over the last 30 years. They include factors such as machine rates, operator rates, material cost, and scrap value. They can be provided as default values in a general analysis or tuned to and edited by an individual manufacturer.

The trend went from using DFA in simplifying the design and shifted to DFM to challenge suppliers on the cost of piece parts. But practice seems to have run its course. Now people are moving back to the whole of DFMA after understanding that DFA alone and pressuring suppliers has its limits.

Profiting from Lean Design

Another visionary that embarked on a similar professional journey in the mid-1980s was Sandy Munro, founder of Munro and Associates. From his position at a major automotive OEM, he developed a different emphasis on the same basic idea of reducing complexity by coining (and trademarking) Lean Design, a companion to Lean Manufacturing. Through his research and experience, he developed the notion that 70% of a product’s profit is established in the design phase.

The best design is the simplest on that works. However, over the years what we measure, along with simplicity, has become a bit more encompassing. There is now Design for Quality; Design for Producibility, or Design for Service, as well as Design for Manufacturing and Assembly. The area where Munro is focusing on is Design to Cost, developing its software package Clike2Cost, which is part of the Lean Suite “Design Profit”. Costing is a black art and fewer experts are available as many have retired or are at retirement age.

Munro developed extensive, region-specific costing libraries with hundreds of materials, machines, and processes that allow engineers and the costing community to create accurate cost quickly. These knowledge-based models are adaptable within each organization. No one company is the same.

After 20-plus years of DFM, DFA (and Design for Cost), Foreman observed that acceptance of the technique ranges from indifference to enthusiasm. Some, have fully adapted Lean Design into their engineering culture with big success. Others may use it when they find themselves in a tight squeeze. The demand for Lean Design is so large in China that we now have an in-depth Lean Design Certification program that many companies are requiring their engineers to achieve.

Integrated CAD Platforms

Teamwork and collaboration can happen not only within companies but in the services provided by software vendors as well. While organizations have been practicing DFM/DFA for a long time, the advent of computerized engineering processes, including CAD, CAE, and other simulations, has enabled greater leverage.

Earlier approaches relied on handbooks, training sessions, and manual reviews. However, these processes were slow, error-prone and hard to sustain and scale. Now we have CAD integrated design assistants like DFMPro which make the implementation of DFM and DFA easier.

DFMPro is “seamlessly” integrated inside popular CAD programs including PTC CREO Parametric, Siemens NX, and Dassault Systems SOLIDWORKS.

Vision of DFM/DFA goes beyond the acronyms of DFM or DFA. To get measurable benefits, one must take a holistic approach and adopt DFX or Design for eXcellence. DFX captures  industry best practices available via DFM rules for core manufacturing processes like machining, injection molding, sheet metal fabrication, casting, and assembly, which DFMPro software does.

Beyond that, DFX means that the software is a platform that can incorporate more than DFMA by allowing third-party technologies to operate in the CAD-integrated environment DFMPro enables. It also, according to the company, integrates easily with PLM, ERP, and MES systems.

The positive impact from Industrial Internet of Things (IIoT), and other smart manufacturing initiatives on DFX is to have a free flow of information in the right direction at the right point in time. The IIoT and other smart manufacturing initiatives enable this. Design best practices for DFM/DFA which were earlier derived based on rules of thumb can be derived from real-time manufacturing data and prescriptive rules can be derived from the huge volume of data using machine intelligence.

New Tools for the Problem

Since the time that DFMA was first developed and introduced, CAE simulation tools have progressed by leaps and bounds. From their origins in applying basic, first principles of stress analysis or fluid flow to complex shapes, simulation tools have expanded into optimization and generative design techniques as computers have become more powerful. No longer is CAE simulation constrained to “testing” if a design can meet performance requirements, it can help create them.

Optimization techniques can improve existing or initial design ideas. Beyond optimization, the emerging generative design field uses computer algorithms to propose many design concepts based solely on constraints provided by the engineer. Generative design mimics nature’s evolutionary approach to design, testing and learning from each iteration what works and what does not, according to Autodesk Inc.. In either case, manufacturing realities can be an imposed constraint.

Autodesk is focusing especially on generative design. With this technique, you can look at different shapes and topologies and make tradeoffs between different manufacturing processes. Generative design minimize part count and assembly complexity among other factors. The AI algorithm can explore making, say a gear, from additive manufacturing, casting, or machining from a billet, including finishing operations, and present many—possibly hundreds—of different solutions.

In April 2018, Autodesk began offering a cloud-based generative design service for its Fusion 360 Ultimate customers. Using the cloud allows computer-intensive operations to be transparent and fast to the user, according to Autodesk. From the many choices the algorithm produces, users can select which ones to convert into a B-Rep CAD representation and download into their own systems for further downstream work.

Manufacturing engineers can influence generative design techniques by making sure that the language and constraints of manufacturing are in the generative design processes and knowledge base of a company. The manufacturing engineering and assembly process still is treated as an afterthought. Generative design can provide collaboration so that people can look at a multitude of ways to solve a problem and start to provide that feedback and insight.

Information Infrastructure

Advances in simulation tools and evolving standards in how information is being distributed have an impact on how designers incorporate DFM/DFA. One such communications foundation is the emerging use of product manufacturing information, or PMI, attached directly to a CAD model.

PMI is not a new concept. The idea is decades old and the enabling standards were developed at least 10 years ago. What we are seeing now in our customer base is a real transition—mass adoption as people are moving to [PMI] and using a model with PMI attached across the enterprise.

Why is this important for the implementation of DFMA? It provides not only a communication tool from the designer to the manufacturing process, but also allows employing other tools just as important for understanding manufacturability.

For example, we have Monte Carlo-based techniques for setting tolerances—they can simulate thousands of variations in assembly configurations and tell you which ones will fail, where tolerances are actually too tight and how to relax them to make a cost-effective process. With PMI attached to CAD, not only can manufacturing processes like machining and fabrication be programmed, so can inspection programs and a host of other manufacturing-related processes.

Other simulation tools that have come into their own are discrete event simulations, such as Tecnomatix Plant Simulation offered by Siemens.

Approaching the problem from this perspective might not always impact design, but it could change the manufacturing process or system. By exploring options through a manufacturing simulation, fixturing could be adjusted or operations performed in a different sequence to get the lowest cost system rather than simply the lowest cost design. Combining DFM tools with other simulation allows all those checks before you ever go into production.

So, almost 30 years after these concepts were introduced, what has been the general acceptance in industry? “I am still not seeing the level of acceptance I would expect for as long as it has been around,” said John MacKrell, chairman of CIMdata. There are a few places where it is well received, such as automotive and aerospace, where either the volume or complexity of the manufacturing process makes it paramount to consider DFM in the engineering process.

“In other companies, I get mixed feelings about it; [acceptance] often depends on the engineer rather than DFM being supported] by a company process,” he said. Obviously, there is room for education. One area where DFM will be vital is additive manufacturing, MacKrell asserted. “While the basic principles of DFMA have not changed that much, additive requires both designing the part for the manufacturing process, and additive manufacturing enables things like reducing part count,” he said.