By < Rune Stolan >
March 30, 2021
For more than five years Industry 4.0 and the Industrial Internet of Things are accelerating the digital transformation of industrial enterprises, while in several cases disrupting their operations. Industrial enterprises are increasingly introducing sensors and Cyber-Physical Production Systems (CPPS) in their shopfloors to boost their ability to manufacture customized products (e.g., made to order), to increase the flexibility and resilience of their supply chains, and to implement ROI (Return-on-Investment) generating use cases like predictive maintenance.
Industry 4.0 provides a wealth of opportunities, which cannot be realized based on a few projects. Realizing the complete value potential of Industry 4.0 may take many years and will span the implementation of multiple projects. As such Industry 4.0 is more like a marathon than a speed race. In this marathon, it is important to start on the right foot i.e., to commence the Industry 4.0 journey on the right track. This is a key for gradually evolving the number and sophistication of Industry 4.0 use cases at the least possible effort and based on optimal value for money. In practice, industrial enterprises need to set up some prerequisite infrastructures that will ease the implementation of Industry 4.0 over a longer period of time.
Industry 4.0 is largely about collecting and analysing digital data from the shopfloor, towards enabling data-driven processes. As such the transition of an enterprise to Industry 4.0 hinges on its transformation towards a data-driven enterprise.
In this context, enterprises need to establish reliable data collection infrastructures, which will enable them to gather, store and manage data about their machines, tools and physical processes. In this direction, they must deploy CPPS systems in their production shopfloor, including machines with sensors, robots, smart objects, and machinery with digital capabilities such as digital interfaces for data collection.
Apart from enabling digital data collection, CPPS systems enable data-driven actuation and control functionalities which are key to realizing automation functionalities in the Industry 4.0 era. The Industrial Internet of Things is not only about analysing digital data: It also entails the implementation of data-driven automation functions that interact with field systems and influence the status of the physical production processes.
To ensure the scalable and cost-effective management of large volumes of digital data there is also a need for BigData infrastructures that can store very large numbers of heterogeneous data streams.
BigData infrastructures handle very large volumes of both streaming data and data-at-rest. They consist of a collection of operational databases, analytical databases (e.g., data warehouses) and data lakes, which must be deployed in conjunction with the legacy data stores and historians that are typically available in a plant. Along with the establishment of BigData management infrastructures, there is also a need for realizing a migration to cloud computing.
Cloud computing infrastructures ensure on-demand access to the required storage and computer resources, which facilitates the scalable processing of large data volumes. Moreover, cloud resources are accessed based on a pay-as-you paradigm, which minimizes the Capital Expenses (CAPEX) of the industrial enterprise. Specifically, cloud computing enterprises can start small and scale their infrastructures as needed i.e., as they capture more data and deploy more Industry 4.0 use cases.
Most importantly, cloud computing is a key prerequisite for orchestrating different manufacturing services from the various systems and part of a plant, including example supply chain management services, production planning services, and quality control services. This orchestration facilitates the streamlining and the virtualization of production processes, which is one of the key value propositions of the Industry 4.0 era.
Leveraging cloud computing virtualization, manufacturing enterprises are provided with the opportunity of configuring and deploying complex industrial automation processes as IT pipelines, rather than having to integrate low-level OT (Operational Technology) systems.
Along with the deployment of cloud and data infrastructures, industrial enterprises must foster the creation of an IT-based data-driven culture. In this direction, there is a need for reskilling and upskilling programs that will enable the workers to use data-driven systems and the business management to make educated, data-driven decisions. Relevant training programs are indispensable assets for the proper migration to the Industry 4.0 era, as they will provide the knowledge required for designing, deploying, and operating Industry 4.0 systems and processes.
The above-listed prerequisite infrastructures can enable the development of basic automation and analytics use cases in manufacturing environments. For example, by collecting and analysing data about the operation of a machine, manufacturers can gain insights into the performance of certain production processes, while improving service and repair processes. Likewise, thanks to cloud computing it is possible to implementing data-driven automation pipelines such as automation of the tools and machines configurations in-line with the needs of certain production orders.
Nevertheless, to realize the full potential of Industry 4.0, there is a need for deploying additional digital enablers on top of the baseline cloud and big data infrastructures. Some prominent examples include:
Automation applications in the cloud cannot always achieve real-time performance, given that the communication of field systems with the cloud incurs considerable network latency. To alleviate such limitations, industrial enterprises have the option of deploying devices or small-scale computing clusters close to the field as part of the edge computing paradigm. Edge computing enables real-time actuation and control operations such as real-time quality control processes or human-centric manufacturing processes that provide real-time feedback about the performance of the worker.
Many Industry 4.0 use cases ask for sophisticated data processing beyond simple rule-based analytics. For instance, the discovery identification of defected products based on labelled images requires the deployment and use of deep learning algorithms. Likewise, the calculation of the Remaining Useful Life (RUL) of an asset based on historical data about failures is usually based on machine learning algorithms. Sooner or later, industrial enterprises will therefore have to deploy machine learning techniques as part of their Industry 4.0 projects.
In most cases advanced forms of automation hinges on the deployment and use of AI-based CPPS such as drones, industrial robots, and automated guided vehicles. These CPPS systems can be classified as smart objects. They will gradually take over laborious, tedious, repetitive, and dangerous tasks on the shop floor.
Several Industry 4.0 applications such as training and remote support leverage augmented reality technology to create realistic cyber-representations of the production tasks and to break time and space boundaries when supporting workers to perform these tasks. For instance, AR technologies can be used to train workers on specific production processes by displaying relevant instructions in devices like smart glasses. Therefore, industrial enterprises may need to deployed AR technology to strengthen the human-centred part of their Industry 4.0 deployments.
Industry 4.0 use cases can be deployed over mainstream industrial networking infrastructures like PROFINET (Process Field Networks). Nevertheless, in coming years the bandwidth needs of Industry 4.0 environments are expected to explode due to the proliferation of CPPS systems, data streams and Industry 4.0 use cases. Thus, industrial enterprises will most likely consider the deployment of next-generation industrial networking infrastructures based on the emerging 5G networks.
The deployment of these digital enablers on top of cloud and BigData infrastructures can be challenging. However, the good news is that their deployment can be gradual i.e., it can be realized over various years depending on the needs of the target Industry 4.0 use cases. Hence, a smooth migration path is possible.
Considering the need for developing, deployment and operating the above-listed infrastructures, a typical deployment path for a new Industry 4.0 solution involves the following steps:
1. Process Design and Reengineering
As a first step, some concrete user stories for the target Industry 4.0 must be specified. The latter include a detailed specification of the problems be addressed, the main users involved (e.g., production managers, production engineers, maintenance workers, factory IT), as well as the proposed Industry 4.0 solution. For instance, digital data can be analysed to improve the maintenance and the operation of an asset, such as a machine. This first step may also entail a reengineering of existing processes, towards a new IT-enabled process.
2. Solution Architecture Specification
The second step focuses on the specification of technical architecture for the digital manufacturing solution. Standards-based reference architectures such as the Industrial Internet Consortium Reference Architecture (IIRA) and Reference Architecture Model Industrie 4.0 (RAMI4.0) can be consulted to align the solution to international best practices.
3. Digital Modelling and Digital Twin Implementation
In this step, the digital models to be used must be defined and implemented. This includes for example designing the data schemas of the databases entailed in the solution, including the schemas for operational and analytical databases. Moreover, any digital twins’ representations should be also defined and implemented. Data modelling can be also based on standards like AutomationML for data exchange specifications.
4. Field Connectivity
Field connectivity is essential for collecting data from the shop floor. It enables data collection by connecting CPPS systems and IoT devices to cloud infrastructures or edge servers. To implement field connectivity, manufacturers can deploy one or more standards-based connectivity benefits such as OPC-UA (Unified Architecture), MQTT (Message Queue Telemetry Transport), DDS (Data Distribution Service), and oneM2M. These protocols are efficient in acquiring and routing data from the field to the cloud.
5. Data Analytics and Knowledge Acquisition
This step focuses on the analysis of the collected data using advanced analytics techniques such as machine learning and deep learning. The outcome of data analysis can lead to new knowledge about the production processes such as the discovery of patterns of process and production parameters that lead to defects. Data analytics functionalities can be placed either at the edge (i.e., edge analytics) or at the cloud (i.e., cloud analytics) parts of an Industry 4.0 infrastructure.
6. Closing the Loop to the Field
Several Industry 4.0 use cases leverage the knowledge of the previous step to drive optimizations in production processes and shop floor operations. Hence, following the data analytics step, this step interfaces to field systems and business information systems (e.g., Enterprise Resource Planning (ERP) Systems) to drive optimizations in production activities such as field automation, machine operations, and production scheduling.
7. Pilot Testing and Workforce Training
Prior to the deployment and operation of the solution, thorough testing is required. Such testing should take place either in a lab environment or in a pilot production line. Moreover, the end-users of the solution (e.g., plant workers, maintenance technicians) must be trained on how to use the technology and on the operational aspects of the new digital manufacturing process.
8. Deployment in Production
The final step entails the deployment of the Industry 4.0 solution on the shop floor. The operation of the solution will be continually monitored to identify the need for any additional developments and fine-tuning. An agile and iterative approach to developing and deploying additional features and functionalities can therefore boost a continuous improvement approach.
In the scope of their Industry 4.0 manufacturers will be confronted with several challenges. Some of the most prominent ones include:
Overall, the emerging Industry 4.0 revolution provides unprecedented opportunities for industrial enterprises to design, develop and deploy novel production processes that shorten production times, improve production quality and optimize production costs. In this direction, industrial organizations must deploy CPPS systems in conjunction with a pool of novel digital technologies, starting from technologies that enable field data collection and data integration in the cloud. The Industry 4.0 journey is promising, yet it also entails several challenges.
Fortunately, there are already several best practices that can be adopted towards a smooth and successful migration to Industry 4.0. Our Upkip platform is built based on international standards and best practices. As such it can help you start your Industry 4.0 transformation on the right foot.
Visit this page for more details or contact us at email@example.com for a demonstration and a free consulting call.