Product Line Engineering: Intelligent manufacturing for intelligent products

By Brandon Lewis


Embedded Computing Design

February 10, 2017

Product Line Engineering: Intelligent manufacturing for intelligent products

  The Internet of Things (IoT) is ramping up the creation diverse, intelligent, and connected products, in many cases much more quickly than the manufacturing people, processes, and technologies...


    The Internet of Things (IoT) is ramping up the creation diverse, intelligent, and connected products, in many cases much more quickly than the manufacturing people, processes, and technologies used to create them can support. This requires a new approach to engineering and product lifecycle management, according to Dr. Charles Krueger, CEO of BigLever Software, and in this interview he explains how large-scale organizations can abstract complex processes and portfolios to gain a competitive edge.

    The Internet of Things is beginning to take hold in almost every industry and application, not the least of which is product lifecycle management in factory automation. What pain points are you seeing in large scale factory automation and how can digitized assets and agile practices benefit here?

    KRUEGER: One of the major challenges is connecting the digital assets that are used throughout the engineering process, and keeping those digital assets connected when the products are sent out into the real world. This requires a complete understanding not only of the physical products, but also of all the critical data needed to manage their digital assets. A product’s digital assets – referred to as the digital twin – are the digital images or representations of what that product contains, its current status, and any metadata about that product. Maintaining this connection from the product in the field back to the digital twin in the company’s “back office” creates another layer of product complexity.

    From the product line perspective, we see this problem multiplied as many different “flavors” of a product are produced. For example, automotive manufacturers must deliver an ever-growing array of options and variants in their vehicle product line offerings. If an automotive company sold a fleet of 30 million vehicles last year, this equates to hundreds of thousands of different vehicle configurations. With the Internet of Things (IoT), there must now be a digital twin for all of these configurations, representing each of the vehicles in the field. The result is a tremendous level of complexity, which is extremely difficult to manage using traditional product-centric engineering approaches.

    Feature-based Product Line Engineering (PLE) provides a way to dramatically simplify this complexity. With this approach, the auto manufacturer can create and manage the digital twin for each vehicle in its product line based on the features contained in that vehicle. Feature-based PLE allows the manufacturer to establish a “feature catalog” for the entire product line and use a “bill of features” to clearly characterize, within the digital twin, which features uniquely differentiate one vehicle from another. As a result, the manufacturer can much more easily track, understand, and communicate with all the products in the field – based on a complete, yet simplified, feature-based view of the digital twin for each product.

    We find that manufacturers who adopt this feature-based PLE approach achieve a discontinuous jump in competitive advantage – which comes from order-of-magnitude improvements in efficiency, cost reduction, portfolio scalability, time-to-market, and product quality.

    Regarding agile, an important characteristic of feature-based PLE is that it is agnostic to development methodologies, so organizations can choose the methodologies that best fit their needs. However, many organizations implementing PLE choose to use agile/iterative development approaches. An organization that has a large number of products in its product line must inherently “be agile” in order to effectively accommodate the many changing requirements, needs, and inputs received from the individual products that need to be supported in the product family under the PLE approach.

    What would PLE look like if implemented at a large-scale manufacturer, and, given that it is an “organizational” approach, how does it affect the day-to-day jobs of stakeholders, from the C suite down to product engineering?

    KRUEGER: We’ve all seen how manufacturers have expanded their product lines to offer more flavors of their products – whether in consumer electronics such as televisions or cell phones, or transportation products like cars, trains, ships, or airplanes. What’s changed more recently is the introduction of much higher levels of intelligence built into these products, which some people are calling Industry 4.0. Self-driving automobiles are a good example of that. All of this diversity and sophistication introduces order-of-magnitude greater complexity into the engineering, deployment, and maintenance of these products.

    Now that complexity is reaching levels beyond which humans can easily manage, new engineering methods and tools are needed to tame this complexity. This is where PLE comes in. The intent of PLE is to help organizations dramatically reduce that complexity by enabling a fundamentally different way – a much more efficient way – to manage this product line diversity.

    Feature-based PLE provides the ability to abstract, which allows organizations to dramatically simplify the way they describe and manage the variations across the entire product line. When managing these product variations at the detailed asset level, there can be thousands, even up to a million, points of variation across those engineering assets. Using the automotive example, rather than looking at all these individual pieces and parts, and how they interact in order to make features emerge on a vehicle, PLE starts from the other end. The organization looks at the features that differentiate the products – which is a much simpler view.

    This is where automation plays a key role. With PLE, the organization creates a “superset” of digital assets that are shared across the entire product line. These digital assets are then equipped with all the feature options offered in the product line. PLE provides an automated production system that assembles and configures the shared digital assets automatically, based on the features that are selected for each product variation. This approach allows organizations to replace very complex and error-prone manual activities with automated ones. Ultimately, PLE removes much of the time-consuming, non-value added tasks involved in managing product complexity, giving engineers more time to focus on creating and bringing product innovations to market faster.

    Once an organization has embraced PLE as a much more efficient way to engineer, produce, and deliver its product family, then the question becomes, “How do we change our organization to embrace a PLE approach, as opposed to the traditional product-centric approach?”

    The metaphor we like to use is the “PLE factory” – much like a typical manufacturing factory, except that this factory is used to configure and produce digital assets, or digital twins, for the products. Organizations have to bring engineers and technical managers into the new PLE factory and train them to operate it effectively. Today, under a product-centric approach, the engineer’s job could be thought of as working in a “single-person garage,” where the work is focused on an individual product. However, with PLE, everyone has to be trained and equipped to come work inside this new PLE digital factory where assets are shared across the entire product line. This requires not only training and mentoring, but also a well-defined strategy and plan for managing the necessary organizational and cultural change.

    [Figure 1 | onePLE is a Product Line Engineering (PLE) solution that provides the business strategy, organizational change, and technology infrastructure for companies to establish a feature-based “PLE factory” that can be used to configure and produce digital assets for each product in a product line based on its features.]

    onePLE was created to provide an end-to-end solution – everything an enterprise needs to predictably adopt a product line approach – and make this transition successfully. onePLE allows a company’s leadership to define one organization-wide product line vision, and provides one unified approach for the organization to adopt and execute on that vision. It helps all of the parts of the organization to navigate each stage of this journey – from how they are producing products today to producing products as contributors to the PLE digital factory.

    What is preventing large-scale manufacturers from moving towards PLE approaches and agile practices?

    KRUEGER: As the complexity within products grows through greater sophistication, such as digital intelligence – and as more and more product diversity is required for competitiveness – manufacturers are recognizing that they need a new approach to deal with these challenges.

    The manufacturing industry is changing very rapidly. Organizations understand they need to move and evolve quickly to accommodate that change. Yet, change is very hard, especially when it involves organizational change. Moving too fast can break pre-existing assumptions about how an organization performs and gets the job done.

    So, it’s not that manufacturers haven’t implemented PLE, or haven’t done it effectively. It’s more a question of how enterprises can make the transition to PLE in a way that avoids organizational disruption. PLE, and certainly the emergence of IoT, introduces a significant amount of change. Many organizations are moving very deliberately, yet somewhat cautiously, to protect themselves against the risk of destabilizing too many things at once. As a result, it takes time for the success stories from the PLE arena to be seen and awareness of these successes to spread out into the broader industry.

    We’ve found that, with the right approach, organizations can speed up this transition and begin achieving the benefits of PLE immediately while maintaining good organizational stability and meeting important goals, such as production schedules. This requires a solution that’s both holistic and incremental. It must be holistic, in that it must encompass all the business, organizational, and technology needs of the company making the transition. Yet, each of these areas must be carefully planned and implemented incrementally so that each stage in the transition brings benefits and enables the next stage in the transition. This has been the major driving force behind our creation of the onePLE solution.

    What’s the next step in factory plant automation? What types of organizations do you see leveraging in the future, and what practices can you prescribe for the digital shift in manufacturing?

    KRUEGER: In the global marketplace, manufacturers must deal with significant variations in manufacturing requirements from one country to another. There are major legislative differences across regional boundaries that require manufacturers to use, track, and maintain different parts and materials, as well as deal with differing supply chains to support their plants. Or, there may be certain types of materials that are more readily available in some regions, so the choice of materials could be based on what is physically convenient or more cost-effective due to geographic location. Manufacturers, and the providers of manufacturing plant systems, need a better way to deal with the complexity of this.

    We’re seeing some innovative work coming into play here, which is poised to become an important part of where the future will take us. In these scenarios, each manufacturing plant around the world can be considered a member of a product family. PLE can provide a significant leap forward as a new way for plants and manufacturing systems to be designed, built, and maintained. With PLE, organizations can get real engineering data across the full engineering V, use digital twins, and create feature-based configurations for very fast design, implementation, and deployment of these manufacturing systems and plants based on a rigorous and well-defined product line approach.

    This is a very interesting future, and organizations that are now making the move in this direction are gaining a critical leg up on their competitors.

    BigLever Software


Brandon is responsible for guiding content strategy, editorial direction, and community engagement across the Embedded Computing Design ecosystem. A 10-year veteran of the electronics media industry, he enjoys covering topics ranging from development kits to cybersecurity and tech business models. Brandon received a BA in English Literature from Arizona State University, where he graduated cum laude. He can be reached at [email protected]

More from Brandon