March, 2021 by John Gøtze
A few weeks ago, I got a letter from The Danish Motor Vehicle Agency. They informed me that the manufacturer of my 3-year-old car had issued a recall due to possible problems with the engine timing belt. I, therefore, had to bring the car to an authorized repair shop for inspection and potential repairs.
The manufacturing industry (including the automotive industry) is of great importance to modern society and is constantly evolving. The manufacturing industry deeply shapes people’s everyday lives by providing them with products such as furniture, clothing, and technological devices such as our cars.
Sustainability, the internet of things (IoT), and predictive maintenance are some of the leading trends in the manufacturing industry and profoundly impact it. Manufacturers are becoming increasingly connected with their customers through the IoT and smart products, while supply chains are becoming more integrated and companies more connected.
Manufacturing is moving towards connected products and the integration of engineering across the entire value chain. Factories and production sites are becoming increasingly connected and processes more integrated. An empowered workforce is crucial to the Industry 4.0 transformation, and employees are becoming increasingly connected through smart equipment.
Smart Products and Digital Enterprises
To better understand the changes occurring in the manufacturing industry, it is important to understand the transformation of products and enterprises.
Consider the example of my car. Being a conventional car, it communicates information concerning performance that it may be experiencing to me the driver but cannot share them directly with the manufacturer. The manufacturer and its authorized repair shops can assess the condition of a car when it is serviced without knowing how it has been used, its performance, and where and in which environmental conditions it has been driven. When it comes to servicing cars, most manufacturers require them to be serviced at predetermined times or after certain distances. Still, a manufacturer cannot know whether a vehicle actually needs to be serviced. In the event of a manufacturing flaw, conventional cars have to be recalled to fix the issue. In 2009 and 2010, Toyota had to recall 9 million vehicles due to a pedal issue, and during the same years Ford recalled 14.9 million vehicles to fix a cruise control issue. In 2015, Toyota recalled 6.5 million vehicles due to faulty power window switches. In the same year, Volkswagen recalled more than 4.5 million cars when it was revealed that the manufacturer had used software to conceal its vehicles' emissions. Such issues endanger both people’s lives and the survival of car manufacturers.
The Industry 4.0 transformation is leading to improvements in the capabilities and value of products. Smart products consist of three types of components:
- Physical components are the mechanical and electrical parts of a product.
- Smart components include sensors, processors, and the operating system embedded in a product. These components collect data concerning a product’s condition, use, and environment.
- Connectivity components include ports, antennae, and communication protocols. These components enable smart products to send and receive data (e.g. product updates).
“Just” a Car?
As an example of a smart product, consider not my dear Citroen C3 but a Tesla car which essentially is also “just” a car, though it costs 5 times more, but is also smarter, than my C3. A Tesla car has sensors that continuously monitor the vehicle’s environment and performance, including the state of its physical components and the way the vehicle is used, as well as traffic, road, and weather conditions. Smart and connectivity components enable the car to continually share data about its condition and use with the manufacturer. These components make it possible, for example, to service the car when doing so is actually required, not at predetermined intervals. For example, the vehicle might indicate that an oil change is necessary based on the performance, viscosity, and working temperatures of the oil, thus preventing unnecessary changes of oil that could still be used. When a car service is required, the vehicle schedules an appointment with a mechanic and notifies the owner for confirmation. There is even a Mobile Service option that allows a car to be serviced by a mechanic at the vehicle’s location, thus avoiding the need to drive it to the mechanic. Another example of improved capabilities and value is the manufacturer’s ability to improve the cars’ performance and fix issues over the Internet, without recalling the cars. For instance, a car magazine tested a Tesla car and reported poor braking performance. A few days after the test, Tesla updated the car’s software over the Internet, recalibrating its braking algorithm and thus instantly improving its braking performance. These examples illustrate some of the differences between conventional and smart products.
The Industry 4.0 transformation is not restricted to smart products but also affects enterprises. An enterprise comprises different functional units. In the manufacturing industry, the functional units within an enterprise work to deliver products and services. The research and development (R&D) functional unit researches new technologies and designs new products. It develops product specifications, bill of materials, and computer-aided design models. The manufacturing functional unit uses these specifications to manufacture the product. The sales functional unit is responsible for market analysis and selling the product. The information technology (IT) functional unit manages the enterprise-wide computing infrastructure that supports other functional units’ work. The service functional unit provides after-sale services to customers.
For enterprises, embracing the Industry 4.0 transformation means evolving into digital enterprises. A digital enterprise is an enterprise that applies data and information to enhance the enterprise’s products, processes, and services. A digital enterprise leverages the enterprise’s digital models to support the use of data and information and facilitate integration among functional units and other digital enterprises.
The transformation into a digital enterprise revolves around applying data and information from functional units and smart products to enhance an enterprise’s performance. In this transformation, data from smart products is fundamental because it can generate insights that can improve the performance of the enterprise and its partners and increase the value offered to its customers. For example, smart products can be better controlled and optimized through software updates.
A Digital Twin of the Digital Enterprise
To apply data and information in a digital enterprise, digital models, feedback loops, and integration among functional units are required. Digital models are models that represent an enterprise and its products, processes, and services. Practitioners refer to these models as the “digital twin of an enterprise.” Gartner defines a digital twin as “a dynamic software model of a thing that relies on sensors and/or other data to understand its state, respond to changes, improve operations and add value”. Feedback loops are the data and information that are the output of a product, service, or process used as input for a product, service, or process. These feedback loops often involve data and information from different functional units and enterprise information systems in a digital enterprise.
Figure 1: Communication among functional units, enterprise information systems, smart products, and enterprises.
In a digital enterprise, functional units have new goals, and digital models, feedback loops, and integration play an essential role in achieving these goals. As shown in Figure 1, in an enterprise that offers normal products, there is no communication from the product to the enterprise, and the limited communication that occurs across functional units is one-directional. One functional unit produces output that is used by another functional unit without timely feedback loops. The same figure indicates that, in the case of a digital enterprise that manufactures smart products, there is communication from the smart product to the enterprise, as well as communication between smart products and/or with other enterprises. In addition, the frequent communication that occurs across functional units is two-directional. Feedback loops are in place, and functional units are integrated. In a digital enterprise, functional units rely on each other to a significant degree. For instance, the R&D functional unit is responsible for integrating connectivity and smart components into new products, as well as for developing features that leverage these components. The IT functional unit needs to provide the infrastructure required by these components. The service functional unit creates new after-sale services that leverage the new components and product data.
The Tesla example illustrates the Industry 4.0 transformation of an enterprise. Its R&D functional unit develops cars with connectivity components that share data about a car’s conditions, use, and environment with Tesla. The R&D functional unit developed various modes for these cars, and it receives feedback from customers on these modes. For instance, a Tesla car has a remote heating mode that allows its owner to activate the vehicle’s heating system through a mobile app. This is possible because Tesla has a digital model of every car it produces. The digital model manages and organizes the data about the car. Tesla uses the data communicated from the car in a feedback loop that provides insights with which the vehicle can be improved. The remote heating mode would not have been possible were the functional units within the enterprise not integrated. For the IT department to be able to provide the connectivity infrastructure necessary to establish connections among the company, its vehicles, and its customers the Research and design and IT functional units need to be integrated.
Transforming the Enterprise
The Industry 4.0 transformation refers to when an enterprise that manufactures conventional products transforms into a digital enterprise that manufactures smart products. To summarize, the Industry 4.0 transformation involves both products and enterprises. Products are significantly transformed due to the incorporation of smart and connectivity components and thus differ considerably from conventional products. An enterprise is also deeply transformed by smart products and the use of data and information to enhance its performance. Digital enterprises improve their performance through digital models, feedback loops, and integration. In a digital enterprise, functional units are integrated using digital models and feedback loops. The Industry 4.0 transformation requires each functional unit to transform with other functional units. It is an enterprise-wide transformation.
Enterprise architecture (EA) is a discipline that aims to support the implementation of enterprise-wide transformations. Enterprise architecture involves the creation of digital models of the enterprise. Enterprise architecture models structure and represent functional units, their processes, and enterprise information systems. Enterprise architecture can also support the development and implementation of feedback loops, as well as the integration of functional units using EA models. Therefore, it is fair to say that EA serves as the foundation for the Industry 4.0 transformation and plays a key role in implementing this transformation within digital enterprises.
The manufacturing functional unit performs the process of converting raw materials, components, or parts into finished goods that meet a customer's expectations or specifications. This functional unit leverages a man-machine set up for large-scale production. In the example of car manufacturers, their manufacturing processes often aim at minimizing costs, and their production lines often remain unchanged for several years.
Digital manufacturing is the application of data and information to enhance manufacturing products, processes, supply chains, and services. It leverages digital manufacturing models to support the use of data and information and facilitate integration within a digital enterprise.
The transformation to digital manufacturing is based on the application of data and information from enterprise information systems and smart products and equipment to enhance manufacturing. Data from smart products and equipment can provide insights into improving manufacturing and other functional units.
To apply data and information in digital manufacturing, you require digital models, feedback loops, and integration. Digital models in manufacturing represent products, manufacturing processes, components, and resources. Digital models also support data access in enterprise information systems and the use of data and information for enhancing products and processes. Both the manufacturing unit and other functional units are involved in feedback loops. The data and information about a product or process provided by the manufacturing functional unit are used as inputs in other functional units. Finally, integration in manufacturing requires a shift from working in silos to a way of working that is integrated with other functional units.
Considering the example of Tesla, the company has a digital model of every car it manufactures. Data from the equipment assembling a car is communicated to this digital model to collect accurate data about the assembly process. For instance, the torque applied during assembly for every part fitted is collected. In the event of quality issues, the digital model’s data can provide feedback to the production line and adjust the torque level applied when assembling certain parts. This adjustment requires the functional unit managing the cars’ digital models and data to be integrated with the manufacturing functional unit to allow the company to continuously learn about and improve its cars and assembly processes.
Researching the Digital Manufacturing Challenges
To understand enterprises’ efforts to achieve digital manufacturing, QualiWare and Aalborg University collaborated on a three-year research project where we examined 21 enterprises and collaborated with eight of them. These enterprises mostly operate in the manufacturing industry; they range from small and medium enterprises to large international enterprises. Based on the interviews conducted and data gathered, the challenges related to digital manufacturing can be classified into three types:
The information availability challenge relates to difficulties in accessing manufacturing information stored in enterprise information systems, a lack of understanding of manufacturing’s resources and processes, and a lack of standardization of manufacturing information.
The process and data heterogeneity challenge has been observed in large manufacturing companies’ manufacturing functional units. In these cases, production sites operate in different countries with different regulations, cultures, and historical conditions. Sharing and comparing processes and data within a functional unit is problematic, and it is even more challenging to share and compare processes and data across functional units. In manufacturing, process heterogeneity refers to differences in terms of processes across production sites. In manufacturing, data heterogeneity refers to the inconsistent storage of data on different enterprise information systems. Process and data heterogeneity are mostly due to various environmental and historical conditions.
The silo mentality challenge relates to a lack of interest in pursuing collaboration across functional units and diffidence regarding openly sharing data and information.
To better describe the three challenges identified above, the following paragraphs present examples of each. First, manufacturing companies have reported difficulties accessing information because few people know how to use an enterprise information system well enough to extract the required information.
Another company reported that they lack an understanding of their production processes and equipment because they only develop production process models when they develop a new product and do not keep their models aligned with the production site's process.
An example of a lack of standardization is the lack of unique identifiers for resources, processes, customers, and parts. One enterprise identified clients through addresses, internal identification numbers, or national identification numbers. As a result, the enterprise does not know whether a customer has bought the correct product for the software license they are selling because the customer is identified in different ways in the enterprise resource planning (ERP) system and the customer relationship management system.
Second, several companies encountered the challenge that, while they identify production equipment and parts numbers according to a rigid convention, the naming of and the level of detail concerning processes in these companies’ respective enterprise information systems and documentation are not standardized. Therefore, the same production process used at different sites is often documented at varying levels of detail and using other names for the activities in the process. As a result, production processes are complicated to understand and compare. Furthermore, production processes are often specified differently for each production site. As a result, the data collected during the production process and documentation for the production process were also very different, even though the same product was manufactured. This challenge hinders the comparison of production processes. In addition, it made it problematic for production managers to share data from heterogeneous processes when attempting to achieve process efficiency.
Third, the service functional unit focused almost exclusively on its goals, refusing to share data and information with the R&D functional unit. Therefore, the R&D functional unit had limited insight into how their products were used.
While digital manufacturing requires digital models, feedback loops, and integration, the manufacturing functional units in enterprises experience a lack of information availability, process and data heterogeneity, and silo mentality challenges. These challenges affect the transformation to digital manufacturing.
The development of digital models is affected by the information availability challenge. The difficulty in accessing data and information in the first place hinders the development of digital models representing data and information and support access to data and information. A lack of understanding of the resources and processes involved in manufacturing limits the development of digital models of products, processes, components, and resources. Digital models are based on data and information, and a lack of standardization hinders their development.
The development of digital models is also affected by the process and data heterogeneity challenge. The fact that processes vary across production sites complicates the development of digital models representing manufacturing processes. The fact that data is stored inconsistently on multiple enterprise information systems also complicates the data representation in digital models and the use of data for enhancing products and processes.
The digital models need to provide operational support for enhancing manufacturing products and processes. Operational support refers to monitoring products, resources, and processes to keep them running and manage errors and problems. In addition, the extensive application of digital models requires efficient modelling approaches.
Feedback loops and integration are affected by the silo mentality challenge. Enterprises encountering these problems rely on solutions intended to address information availability and process and data heterogeneity challenges. Once solutions for these challenges are available, incentives for promoting collaboration across functional units and openly sharing data and information need to be developed.
Enterprise Architecture to the Rescue
Enterprise Architecture supports the implementation of enterprise-wide transformations such as the Industry 4.0 transformation. Enterprise architecture can support digital manufacturing by addressing the three challenges encountered in manufacturing presented above. When using enterprise architecture, digital models can represent manufacturing’s products, processes, components, and resources. The models could be applied to address information availability and process and data heterogeneity challenges. Furthermore, the architecture models could be further developed to provide the operational support necessary to enhance products and processes.
Using the solutions developed to address these challenges, Enterprise Architecture can be applied to address silo mentality. It could also support developing incentives for promoting collaboration across functional units and openly sharing data and information.
Figure 2: Role of Enterprise Architecture in supporting functional unit integration.
As shown in Figure 2, an Enterprise Architecture team can collaborate with functional units to develop models representing their key elements, processes, and resources. Enterprise architecture models can represent different aspects of an enterprise. A common Enterprise Architecture model is a business process model that identifies the structure of processes and subprocess (in QualiWare called business process networks). These types of models can support the management of process heterogeneity.
Other Enterprise Architecture models relate to enterprise information systems and data. These models specify the structure and relationship of data and enterprise information systems. These types of models could support the management of data heterogeneity.
A common challenge in “hand-held” Enterprise Architecture is that models sometimes are not related to the data stored in an enterprise’s information systems, even though these systems support the respective processes of the functional units. Therefore, QualiWare’s platform is a connected repository with several integrations to the enterprise’s core information systems and digital backbone.
Researching Digital Manufacturing
QualiWare sponsored Marco Nardello’s PhD from 2016-2020.
The project title was “Enterprise Architecture for Digital Manufacturing, EA Models and an Automated Modelling Method to Support the Industry 4.0 Transformation.” Marco collaborated in several industry and research projects with Danish and Swedish manufacturing companies and Aalborg University on Industry 4.0 and smart manufacturing. The PhD thesis is available here.