This piece was originally published in the September/October 2019 issue of electroindustry.
Dirk Schulz ABB Corporate Research, Ladenburg, Germany dirk.schulz@de.abb.com and Lars Simora Busch-Jaeger Elektro, Lüdenscheid, Germany lars.simora@de.abb.com
From automotive to commodity goods, mass customization has become state of the art in discrete manufacturing. In Industry 4.0 (the current trend of automation and data exchange in manufacturing technologies), one of the drivers of mass customization is to manufacture highly customized consumer goods in small batch sizes while also cutting costs. We’re literally down to “lot size one.”
There are two main challenges to making a perfectly customized product: to contain costs, the production system must self-configure for any individual product design without the need for human involvement; and, as virtually no spares are produced, any single product must be of perfect quality.
However, when designing the corresponding production line, it may turn out that established integration concepts will not provide the flexibility or quality control needed for this endeavor. Instead, manufacturers can achieve that goal using concepts from Industry 4.0 such as the digital twin and machine- to-machine (M2M) protocols such as OPC Unified Architecture, a leading Standard across systems in automation and security technology.
Quality and Perfection in Every Step
Numerous innovations are possible in the context of Industry 4.0, such as autonomous machines that negotiate production schedules with each other, or smart products that steer themselves through the production process. The first challenge, then, is to decide which concepts are useful for a particular production environment.
Considering the design freedom the consumer has, every control element is potentially a unique product. Quality becomes both a top priority and a top challenge because a quality issue requires the complete item to be manufactured again and thus might delay the entire customer order.
Today, lack of digitalization is a main contributory factor to quality issues: material is procured and transported manually, guided by order data on paper slips. Mix-ups happen but are hard to detect. While design data are available in digital form, they are manually transmitted to the machines on removable media. Should quality inspection of the finished product reveal any flaws, auditors can’t review quality- relevant data such as calibration settings because the machine didn’t capture that data.
In individualized production, it is imperative to detect flaws in the product as early as possible. Even better, avoid flaws altogether. Individualized production must be highly adaptive and resilient. Human operators and machines need to collaborate seamlessly to close the quality loop.
Complete Digitalization Using Digital Twins
Individualized production has two key enablers: adaptive material transformation within machines and digital interoperability between machines. While material technologies like NIR lasers or digital printers already exist, machines still need to be taught how to work together naturally to achieve efficiency and quality.
Today, machines are hardwired or coupled through Programmable Logic Controller software. While this arrangement achieves a tight integration, it comes at a significant effort and does not yield much flexibility for new product variants except for the originally designed product. For truly adaptive production, achieve machine integration with a much looser coupling so the entire line can be flexibly reconfigured with minimal effort.
This is achieved by providing digital twins that represent the specific skills of each machine in a common format on the network, independent of any wired or programmed connection. To avoid “media breaks,” represent product design and order and quality data in the same manner. By orchestrating the machine digital twins based on the ordered product design, any reconfiguration in software seamlessly translates to production steps in the physical world, and quality information is mirrored back into the software domain.
Digital Twin: Production Steps for the ABB-tacteo KNX Sensor
Part of the production line for the ABB-tacteo KNX control element for smart buildings. Established production line concepts struggle with small lots sizes. However, Industry 4.0 concepts such as the digital twin and M2M protocols allow complete flexibility right down to a lot size of one. All photo courtesy of ABB.
We’ll use a capacitive control element for intelligent building automation in high-end luxury hotels, offices, and public and residential buildings as an example of how this works.
The number of functions provided by the control element is variable and determined by the customer’s specific needs and wishes. Individually configured according to the customer’s desire with an easy-to-use online configurator, each sensor is unique in design and function.
This means that the consumer is given direct control over the design of the end product, while the wholesalers take care of distribution and the expert installers integrate building automation products according to the consumers’ needs, as is always done in a traditional network.
Back at the factory, using the digital twin concept, the control element passes each production station as follows: Upon arrival, the glass is scanned and its identity is used to get the approval, design data, and process parameters for the next production step. To this end, the machine contacts an orchestration service that automatically replicates the product recipe from the enterprise resource planning (ERP) software and holds the production history of the particular glass.
Once the glass has been processed, this is reported back to the orchestration service along with machine data used for quality tracking. The glass now departs for the next station as indicated by operations management and as shown on the human–machine interface to direct the worker. For normal operating steps, station operators interact with the machine using a Standardized automation panel. Manual steps, such as material transport, are electronically guided and supervised.
For example, if a glass is inserted into the printer before it has been lasered, it is rejected, and the human operator is instructed to take it to the laser station instead. If glasses are transported from the laser to the printer out of order, the printer still is guaranteed to receive the correct data matching the lasered icons on the glass. Before leaving the printer, the current calibration data are captured for the particular glass. The same holds true when calibrating the icon brightness on the product itself during automated testing. In this manner, a quality issue with the product can be traced to the potential root causes such as problems with the machines or suboptimal process parameters. Above all, quality issues are raised right when they occur. Any flawed product is automatically removed, and production of the corresponding design is restarted.
Machines and their digital twins are derived from a common design template, which helps support how human operators, machines and services work on them. Machines following this template are also easier to add or replace because they naturally plug into the existing production system.
In this production concept, the control element qualifies as a smart product: The identity of each glass actively drives its own production, and the outlined process steps allow the quality loop to be closed based on the global product memory.
Designing the Digital Twins
The production process described depends on a mix of very different data: some relate to the product design, some to the manufactured product, some to a specific machine—and some describe information common to all products or machines. For example, for each product design, there are machine-independent characteristics such as the layout of the control icons; there are generic machine properties such as the current operating state; and there are properties like the printer calibration data, which are machine- specific but do not depend on the product. Product design is translated into machine-specific artifacts like rasterized icons, and there are test results for each manufactured product.
To create these digital twins, the best practice was to first create a top-down Open Platform Communications Unified Architecture (OPC UA) information model from the perspective of the overall production steps. Only in the next step were the model split and parts allocated to the individual machines and software services. To upgrade existing machines with this model, the best method was to wrap them using an embedded OPC UA gateway. To this end, close collaboration between information designers, the integrator, and the various machine builders were required to ensure the real machines could simply be plugged together in the factory to form the desired production system.
Industry 4.0 for All
Designing a production line provided firsthand insight into the needs of machine builders and production designers on the road to Industry 4.0. Particularly for individualized production, perfect quality control and automated reconfiguration are key challenges to be addressed by an automation solution. These challenges can be met based on Industry 4.0 concepts such as digital twins, smart products, and M2M technologies like OPC UA.
While today, individualized production requires a significant upfront investment, future solutions must be covered by established annual investment budgets to be broadly competitive.
To this end, specialist machines need to be taught how to work together out of the box. This requires automation vendors to supply processes, tools, and Standards that make it easy for production designers, machine builders, and integrators to build their digital twins independently and then run virtual integration testing before physical machines are built and commissioned in the factory. Currently, premium products are welcome catalysts to drive this type of innovation. The vision, however, is to offer Industry 4.0 not as a premium product but as an off-the-shelf solution that can be applied to any type of production, from large factories to small and medium-sized enterprises. ei