Download PDF Spring Bridge on Postpandemic Engineering March 14, 2021 Volume 51 Issue 1 This issue is dedicated to the future of manufacturing. A stellar slate of experts present diverse experiences and perspectives from industry, a national laboratory, and academia. Together the articles provide informative coverage and holistic views on the future of advanced manufacturing, leveraging new and emerging technologies, desired infrastructure, innovative approaches, and a resilient supply chain to fortify US manufacturing competitiveness in the coming years. The Local Factory of the Future for Producing Individualized Products Thursday, April 1, 2021 Author: Yoram Koren Mass individualization factories will create local manufacturing jobs and novel research areas in manufacturing operations. My vision for 21st century innovation in the manufacturing industry is the establishment of a new type of factory for producing individualized products for a single customer at an affordable price. Such factories will be located close to customers and designed for cost-effective, large-scale production of a variety of individual products. Two key technologies that enable the production of mass-individualized products are 3D printing machines (Murr 2016) and the manufacturing system architecture (the way machines are both placed in the system and connected to each other; Koren and Shpitalni 2010). One type of individual product is the design of car interiors to fit a customer’s specific needs and preferences. Realizing this vision requires new software that will enable an ordinary buyer to design the unique interior of her own car. Individual products will be manufactured in local factories because interaction with the buyer is an essential element in the success of this new industry. This contrasts with the globalization trend that started around 2000 in the manufacturing industry and facilitates the production of products in distant countries (Koren 2010). I believe that mass individualization factories will dramatically change the landscape of the manufacturing industry, establish many local manufacturing jobs, and create novel research areas in manufacturing operations. Historic Trends in Modern Manufacturing The history of modern manufacturing began around 1850, when the automobile industry started and cars were produced locally, one at a time, each for an individual buyer (figure 1). In 1913 Henry Ford’s moving assembly line was invented, which led to mass production, peaking in 1955, when only seven types of cars were produced by the Big Three in the United States. Mass customization was introduced around 1980, with more models and options offered by the manufacturer for selection by the individual customer. FIGURE 1 Product volume versus variety in four manufacturing paradigms over time: craft production, the prevailing approach until 1913 when Henry Ford introduced the assembly line; mass production (1913–80), which lowered both cost and product variety (mass production peaked in 1955, when the Big Three US auto makers offered just seven models); mass customization (since roughly 1980); and, with the ramp-up in globalization at the turn of this century, individual production, which will be initiated in the 2020s, thus returning to outfitting cars to the individual consumer’s needs but at low cost compared with craft production. Now the trend is toward individualization (Clifton 2018; Kanecko et al. 2017). The internet and social networks are changing the market, making it easy for customers to come up with new ideas and unique products (Jiang et al. 2016). As shown in figure 1, the societal shift from craft production to mass production to mass customization and now individual production is coming full circle—but from market-of-one for the wealthy to market-of-one for the ordinary buyer. Table 1 summarizes the differences among three consecutive paradigms: mass production, mass customization, and mass individualization (Koren et al. 2015). With mass production the factory scheduling is fixed—1000 products a day on the same production system with the same part program. Mass individuali-za-tion—each product is different and requires another set of manufacturing operations (that each takes a different amount of time)—entails completely new mathematical challenges in scheduling and in system operations. Another challenge is that new interfaces are needed to allow ordinary consumers to connect with the mass individualization factory of the future. Mass Individualization Factories I will elaborate on two types of mass individualization factories: A factory that can produce numerous individual products at affordable cost, thanks to manufacturing system architectures that enable the simultaneous production of a variety of market-of-one products. An assembly factory for building individualized vehicle interiors and other open architecture products (which comprise a fixed platform and modules that can be added and swapped; Koren et al. 2013) that require manual assembly of components chosen by customers. Mass Individualization Factory for Producing a Variety of Individual Products The mass individualization factory of the future that I envision enables the manufacture of many individualized products (e.g., a decorative garden fountain, a small metal roof for an outside door, or an individual prosthetic ordered by a hospital). FIGURE 2 Two manufacturing system architectures of a factory for manufacturing mass individualized products. (a) Simultaneous production of four individual products through various manufacturing phases. A gantry brings parts from the conveyor to the machines in the cell and when a machine finishes the task, the gantry takes the part back to the conveyor, which moves the part to the next cell (e.g., from CNC milling to 3D printing), and so on to the finished product. (b) Diagram of the whole factory (100) organization and product progression. Materials enter on one side (130) and finished products emerge on the other side (132), transported through the system on a loop conveyor (120). Hexagons (102) represent manufacturing cells, each with 4–6 machines (e.g., milling machines, 3D printers, welding machines) or assembly stations (104). Cell controllers (134) are connected to a central controller that coordinates the traffic of all products. CNC = computer numerical control. For this factory I propose two types of manufacturing system architectures, depicted in figure 2 (Gu and Koren 2018; Koren and Hill 2005). The system architecture is the most critical factor in making the factory responsive to market changes (e.g., type of products, quantities) and thus economically successful. The system architecture determines the layout of machines in the system and the way they are connected to each other. (Upon finishing a machine’s sequence of operations, the connectivity between machines determines to which machines the item is transferred to continue its production.) For a mass individualization factory the manufacturing system should enable these three features: 1. transfer of the product from any machine to any other machine in the system, 2. simultaneous production of several different individualized products, and 3. space reserved for the rapid addition of machines in the future. For example, the factory system in figure 2a has spots for four machines in every production stage, with reserved spaces for adding two milling machines and one drilling machine in the future. The diagram is based on modified system architecture of the reconfigurable manufacturing system (RMS) architecture (Koren 2010), which has been in use by Chrysler, Ford, and GM since the turn of the century. It is an ideal framework to deal with the “unpredictability of market requirements and the frequent changes induced by technological innovations. For this reason, firms are more and more addressing the need to be responsive at affordable cost” (Napoleone et al. 2018, p. 3815). In addition to a typical RMS architecture, figure 2a depicts a return conveyor (or a gantry), which allows the transfer of a product from any stage to any other stage in the system, and integrates 3D printing machines (i.e., additive manufacturing) (Rogge et al. 2017), computerized numerical control (CNC) machines, and assembly stations in a single, coherent system whose operations are synchronized for the simultaneous manufacture of four different products (shown with purple, green, red, blue lines). Another system architecture for the mass individualization factory is featured in figure 2b (Koren and Hill 2005), which depicts eight manufacturing cells (102), each with four to six manufacturing machines (e.g., milling machines, 3D printers, welding machine; 104) or assembly stations with workers. The system comprises 45 machines and processes (e.g., assembly). A gantry or loop conveyor (120) connects all machines and assembly stations. The fact that every assembly station and machine can be connected to any other maximizes the ability to produce simultaneously a variety of individualized products. In addition, it is easy to replace machines or controllers, which are integrated rapidly like plug-and-play modules. This system architecture is an enormous advantage in a rapidly changing world in which new processes are invented quickly. Cell controllers (134) for the manufacturing operations are connected to a central controller that coordinates the traffic of all products. Materials enter on one side of the factory (130) and finished products emerge on the other side (132). Figure 2 shows two options to connect a variety of machines in a manufacturing system. Another emerging option is to move parts between machines in the system with autonomous mobile robots that use obstacle avoidance algorithms (Borenstein and Koren 1991) to avoid collisions with machines and workers. In complex systems this method will require a large floor space. Impediments to implementing cost-effective individual production may include the need for coatings or postprocessing stages (e.g., for heat treatments). Customer Design and Assembly of Vehicle Interior Imagine that a car interior is an open space and every individual buyer can design it according to his or her individual needs and preferences, subject to safety constraints (Koren 2005). Buyers will use software to select from internet hardware modules designed and produced by third-party companies (the cloud in figure 3) that may design furniture or appliances to fit into the car, or dog seats, or any other items that a customer may want. All options will have the defined mechanical and electrical interfaces needed to attach to the chassis. FIGURE 3 Schematic of a software system facilitates the customer’s (10) interior design of a vehicle and automatically identifies design conflicts (30). Adapted from Koren et al. (2006). The equivalent software example is in smartphones. Many individuals and companies invent applications and phone buyers decide which apps to upload. Potential interest in (and the market for) individual customer design of car interiors is enormous. There has been substantial progress toward autonomous driving, but the inside of the vehicle has not changed in 100 years. The 1908 Ford Model T had a driver and passenger in the front seats and three passengers in the back seat; the typical car interior is the same today. The challenge is to accommodate many individual customers in the design of their car interior. The software system illustrated in figure 3 could facilitate this. Customers would select from a database of available interior components to design their dream car, possibly assisted by cloud-based design algorithms (Wu et al. 2015); the software would identify design conflicts (30 in figure 3). The source of the drawing in figure 3 is a 2006 patent application (Koren et al. 2006). Although it was abandoned, a search on this application shows that 34 new US patents are based on it, including those of Ford Motor Co., Mazda, Boeing, Johnson Controls, IBM, and Lear Corp. But implementation of the concept of the individual buyer’s design of a car interior raises a nontechnical challenge: Buyers are confused by too many choices (Iyengar and Kamenica 2010). When the car interior design concept is realized, a new profession will probably emerge: designers of car interiors who assist buyers with their choices. Assembly of Open Architecture Products The auto interior design method may be generalized to a broader category of open architecture products (OAPs; Koren et al. 2013) (also called open platform products), comprising a fixed platform with defined mechanical and electrical interfaces. The OAP platform enables the addition of a variety of modules that may be designed and produced by different manufacturing companies, as long as the modules are designed with the mechanical and electrical interfaces defined by the platform manufacturer to allow their easy and safe integration. Customers choose their desired modules. The integration of a variety of hardware modules in a platform, using the method depicted in figure 3, adapts the product functionality to the user’s specific needs and preference. A major challenge is the development of OAP design software for use by nonprofessional customer-designers. Following are examples of open architecture products: a hospital bed to which mechanical and electrical components and devices can be added to accommodate a patient’s medical or physical condition an office chair designed to allow the addition of accessories that fit the buyer’s desire and needs a refrigerator interior with interfaces for components such as special shelves or a rotary platform that facilitates finding things a machine tool designed to add a variety of machining and milling heads for different applications. The assembly of automobile interiors and OAPs must be done in local factories because close interaction with buyers on product design and specifications is essential. Challenges As manufacturing practice moves toward individualized products, customers will place their orders and factories will have to convert them for implementation. Software needs to be developed to help customers turn their ideas into digital models for factory production. Algorithms will schedule the production of each individual product, while the system is simultaneously producing a variety of individual products. Optimal scheduling should maintain the maximum number of machines busy nearly continuously even as each product needs a different processing time on several machines. It will be a challenge to build software that schedules different individual orders while optimizing system operations. Artificial intelligence will have a major role in such scheduling in the production system. AI software can group similar products for production so that the system is synchronized to optimize efficiency. This synchronization will require digital twin models (Rosen et al. 2015).[1] Economic Impacts of Local Factories for Individualized Products Factories for individual production and shops for auto interior assembly cannot be located offshore because in most cases direct communication between the customer and the factory will be needed to avoid misunderstandings. Local factories will also reduce the time from order to delivery—buyers will want their customized product as soon as possible, not in 3 months from an overseas producer. Establishing many such factories will create numerous local manufacturing jobs and greatly benefit local economies. More broadly, with traditional manufacturing, when it “bids farewell [through offshoring], engineering and production know-how depart as well, and innovation activities eventually follow. We can trace how this happened in the US” (Kota et al. 2018). In contrast, with individualized production, technologies and practices will remain in the United States. The experience of strong economies around the world shows that a nation needs both R&D and manufacturing activities to maintain a healthy 21st century industrial ecosystem. I hope the US government will find ways to support US industry in developing mass individualization factories, and that agencies will encourage the US research community to develop the science base for the emerging mass-individualization paradigm. Conclusions Mass individualization shifts the product design focus from a large manufacturer to the individual buyer and multiple companies that invent and offer personalized modules to suit customer needs and tastes. I predict that 10 years from now the norm will be that the customer sends requirements to a local factory that produces individualized products for delivery in a timely manner. Manufacturing enterprises will introduce consumer products with open platform hardware that will facilitate the integration of mechanical and electromechanical modules. The number of options in certain individualized products (such as auto interiors) will depend on the creativity and ingenuity of both the customer and the companies that produce modules for integration in the product. The discussion and figures in this paper effectively illustrate the concept of individualized products and suggest practical implementations of manufacturing systems to produce them. The emergence of many small innovative companies will sustain continuous steady growth in the manufacturing sector and boost the economy. References Borenstein J, Koren Y. 1991. The vector field histogram: Fast obstacle avoidance of mobile robots. IEEE Transactions on Robotics and Automation 7(3):278–88. Clifton L. 2018. The factory of the future. Siemens Magazine. Gu X, Koren Y. 2018. Manufacturing system architecture for cost-effective mass-individualization. Manufacturing Letters 16:44–48. Iyengar SS, Kamenica E. 2010. Choice proliferation, simplicity seeking, and asset allocation. Journal of Public Economics 94(7-8):530–39. Jiang P, Lang J, Ding K, Gu P, Koren Y. 2016. Social manufacturing as a sustainable paradigm for mass-individualization. Proceedings, Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture 230(10):1961–68. Kanecko K, Kishita Y, Umeda Y. 2017. In pursuit of personalization design. Procedia CIRP 61:93–97. Koren Y. 2005. Reconfigurable manufacturing and beyond. 3rd Internatl Conf on Reconfigurable Manufacturing, May 10, Ann Arbor MI. Koren Y. 2010. The Global Manufacturing Revolution: Product-Process-Business Integration and Reconfigurable Systems. Hoboken NJ: John Wiley & Sons. Koren Y, Hill R. 2005. Integrated reconfigurable manufacturing system. US patent #6,920,973. Koren Y, Shpitalni M. 2010. Design of reconfigurable manufacturing systems. Journal of Manufacturing Systems 29(4):130–41. Koren Y, Barhak J, Pasek Z. 2006. Method and apparatus for re-configurable vehicle interior design. US patent application 20070156540A1, filed Jan 5. Koren Y, Hu J, Gu P, Shpitalni M. 2013. Open-architecture products. CIRP Annals 62(2):719–29. Koren Y, Shpitalni M, Gu P, Hu SJ. 2015. Product design for mass-individualization. Procedia CIRP 35:64–71. Kota S, Talbot-Zom J, Mahoney T. 2018. How the US can rebuild its capacity to innovate. Harvard Business Review, Oct 23. Murr LE. 2016. Frontiers of 3D printing/additive manufacturing: From human organs to aircraft fabrication. Journal of Materials Science & Technology 32(10):987–95. Napoleone A, Pozzetti A, Macchi M. 2018. Framework to manage reconfigurability in manufacturing. International Journal of Production Research 56(11):3815–37. Rogge MP, Menke MA, Hoyle W. 2017. 3D printing for low-resource settings. The Bridge 47(3):37–45. Rosen R, von Wichert G, Lo G, Bettenhausen KD. 2015. About the importance of autonomy and digital twins for the future of manufacturing. IFAC PapersOnLine 48(3):567–72. Wu D, Rosen DW, Wang L, Schaefer D. 2015. Cloud-based design and manufacturing: A new paradigm in digital manufacturing and design innovations. Computer-Aided Design 59:1–14. Dedication This paper is dedicated to the memory of my daughter, Esther (Asi) Koren, who passed away at age 49 in our house on October 28, 2020, a couple of days after I finished the draft of this paper. She had cancer. Asi established and managed the Ana House to help many girls who experienced difficulties in their life. [1] In 2001 my colleagues and I tested the first prototype of the digital twins on our manufacturing system at the University of Michigan’s NSF-sponsored Engineering Research Center for Reconfigurable Manufacturing Systems. About the Author:Yoram Koren (NAE) is the James J. Duderstadt Distinguished University Professor Emeritus of Manufacturing and Paul G. Goebel Professor Emeritus of Engineering at the University of Michigan, Ann Arbor.