A Product Variation Flight Checklist

The Airbus A380 has approximately 4 million parts and 100 million lines of code, plus avionic and interior variations, all of which add to its design complexity. Image courtesy of Airbus.

Complex products such as robots, tractors, airplanes and battleships are now much more than mechanical products; they are sophisticated digital systems with a mechanical purpose. The Airbus A380 has approximately 4 million parts and 100 million lines of code embedded in processors scattered throughout the aircraft. Such a difference between part count and code line count is now normal in automotive, aerospace & defense, medical, high tech, shipbuilding and many other industries.

An individual part or assembly may stay the same on every A380 delivered, but there are wide variations from customer to customer for seats, interior design, avionics and the software systems driving virtually every aircraft operation. Managing the type of complexity that variation inevitably brings is emerging as a distinct engineering discipline in many industries.

Successful management of product variation must take into account hardware, software and systems. Mass manufacturing techniques must be changed to support customization; software and hardware engineers must collaborate; and systems engineers must be involved in all stages of engineering.

There are various approaches to managing product variation. Some companies create unique bills of material (BOMs) for each variation. But what is the role of the BOM when a company needs to build one item one thousand different ways? Such product variation requires an integrated approach to product lifecycle management (PLM) platforms, where multiple existing systems collaborate to add customer value, reduce transaction costs, trim lead times and control production.

Some manufacturers are rethinking the use of their existing PLM tools, such as configuration management, program management, engineering change management and engineering to order (ETO) or design to order (DTO). Change management tools in particular are being seen as more than a method to track how errors were found and eliminated.

The goal is to have an engineering-based plan for variability. Managing product variation means being proactive about reuse, whether it is parts, code or systems. Such variability management must account for both commonalities and differences, and model them both in a common and consistent way.

The internal cost in time and labor from poor management of change and variation can be enormous. The independent research team of Terwiesch & Loch conducted a survey on engineering change orders (ECOs) in 1999 and found engineering changes consume between one-third and one-half of total engineering capacity and represent between 20% and 50% of total tool costs in manufacturing. Given the ever-present need for speed to market and rapid innovation turns, such high statistics for handling ECOs should be unacceptable no matter the size of the company.

Reconfiguring Existing PLM Tools

Revising existing PLM products and methods is one approach. IT equipment manufacturer Inphi used a variety of software products to work with internal engineering data, but realized its processes for change management and variation control were not affected. The lack of communication among departments and systems forced employees throughout the company to spend time manually searching for needed data. With help from U.S.-based PLM implementation specialist Razorleaf, Inphi conducted an internal discovery process to determine the data points and business processes that would need to be integrated to achieve a proactive change management system that connected all of the affected processes and teams.

“Our customer response time, overall customer satisfaction and sales team efficiency have all improved,” says Robb Johnson, director of Technology at Inphi. Inphi connected its change management to Autodesk Fusion Lifecycle and Autodesk Vault for engineering, and for operations and sales to the company’s Oracle-based business systems and Salesforce.

Specific Tools for Product Line Engineering

The management of software inside manufactured products has become an important aspect of managing product variation. In some products, it is not the manufactured parts but the software “parts” that change the most from one product variation to another. The engineering discipline known as product line engineering (PLE) is gaining traction as a way to manage a portfolio of products and a shared set of software assets within an efficient means of production.

PLE was originally used only for managing product software, but in recent years has become a new way to manage the engineering process behind products with significant elements of variability and modularity. Using PLE requires a shift from building products to meet requirements—with occasional discreet modifications for variations—to planning around the notion of “systems of systems” where processes, code and parts are designed for reuse across families of products.

PTC is a proponent of the product line engineering approach. Derek Piette is the product management director for PLE at PTC. He says PTC customers using PLE take one of two approaches, the physical or the abstract. “An example of a customer implementing PLE in the physical domain is a truck manufacturer that uses our CAD software to design every possible configuration of specific vehicle features—like the height of the door or wheel, or size of the cab,” says Piette. “No two trucks are alike.”

By comparison, the more abstract use of PLE is to think about modularity and variability at the backend to identify standard components that can be shared across an entire product line. Piette says these standard components are then designed “in ways that make them easy to configure.” He says Volkswagen designs one engine for all car models and then fine tunes it based on horsepower. PTC considers this to be “front-end PLE.”

Another software vendor working specifically in the PLE space is BigLever Software. The company says managing product variation with PLE means moving from a product-centric point of view to a production system point of view. The production system is designed to automatically produce all of the products in a specific line or portfolio. The focus is on a singular means of production rather than on all the specific variant products.

BigLever CEO Dr. Charles Krueger says the product line engineering approach consolidates many activities that have been separate in the past, “to reduce the duplication of effort and activities. As a result we get very high levels of productivity, efficiency, scalability and time-to-market improvement.” Krueger says the increases do not come as incremental improvements: “These are business transformation levels of improvement that can be factors of three to factors of 15 improvements.”

Lockheed Martin is a BigLever customer. They adopted PLE when government contracts shifted toward encouraging commonality and use of shared assets in defense products. According to BigLever, when Lockheed Martin adopted its Gears PLE solution, it reported cost avoidance of over $139 million in a three-year period. Lockheed Martin also reported a 40-60% reduction in test cases. The company achieved these gains while producing variations of a weapons line for deployment by the U.S. Navy, the U.S. Coast Guard and the U.S. Department of Defense Missile Defense Agency, as well as for the navies of unnamed “key allies.” BigLever says Lockheed also uses the feature-based configuration capability in Gears to provide full, automated traceability from features to delivered products, to demonstrate compliance with a variety of tough export control requirements.

A detailed case study titled “The Challenges of Applying Service Orientation to the U.S. Army’s Live Training Software Product Line,” from the U.S. Army and General Dynamics and published by the Association for Computing Machinery explains the evolution of their product management to a product line engineering approach. The product line in the study is a live weapons training system comprised of many elements and many variations. The transformation to product line engineering is called Live Training Transformation (LT2) in the study.

“LT2 has realized significant improvements in cost savings and cost avoidance totaling hundreds of millions of dollars in development and sustainment of live training systems,” according to the case study. The study also cautions that “empowering product teams to drive to a more effective state of product line engineering can only be accomplished when product teams are motivated to change.”

Paul Clements, VP of Customer Success at BigLever calls the cost benefits of product line engineering a “superlinear” phenomenon. “We call this effect superlinear cost avoidance because it exceeds the cost avoidance predicted by the linear cost models that, until now, have been posited in product line economics work,” he says.

The Challenge of Increased Complexity

When a new process increases efficiency, the natural tendency is to test the boundaries of the process. A review of academic research shows there is a tendency among companies that are managing product variation with PLE to go big. One such researcher, Kyo Kang of Pohang University in Korea, sums it up: “When a family of products is derived from a common asset base and when these products and assets evolve over time, the complexity of managing products and assets increases an order of magnitude compared with managing versions of a single product.”

The solution to complexity brought about by product variation is a high-level, top-down recognition of the issue. When senior management considers complexity management a discipline or job function in the organization—equal to any primary engineering task in the product development lifecycle—engineering teams will have the necessary resources to reap the potential benefits.

More Info

• Autodesk
• BigLever
• Inphi
• Lockheed Martin
• Oracle
• PTC
• Razorleaf