IoT in Manufacturing

The application of future internet technologies in general and of the IoT in particular, in the scope of future manufacturing, can be classified into two broad categories:

• IoT-based virtual manufacturing applications, exploit IoT and cloud technologies to connect stakeholders, products and plants in a virtual manufacturing chain. Virtual manufacturing applications enable connected supply chains, informed manufacturing plants comprising informed people, informed products, informed processes, and informed infrastructures, thus enabling the streamlining of manufacturing processes.

• IoT-based factory automation focuses on the decentralization of the factory automation pyramid towards facilitating the integration of new systems, including production stations and new technologies such as sensors, Radio Frequency Identification (RFID) and 3D printing. Such integration could greatly boost manufacturing quality and performance, while at the same time enabling increased responsiveness to external triggers and customer demands.

Levels of Manufacturing Digitisation

There are following three progressive evolutionary levels of digitalization:

• Digital Products: Driven by the development of the IoT to smart connected objects, it includes developments of markets like the connected car, wearables, or smart home appliances.

• Digital Processes: Driven by the development of IoT-enabled Cyber-Physical Systems (CPS), it includes Industry 4.0, the further spread of automation in production and the full integration of simulation and data analytics over the full cycle from product design to end of life (circular economy).

• Digital Business Models: Driven by service-oriented IoT-based business models, it includes the re-shuffling of the value chains and blurring of boundaries between products and services with the final aim to increase profitability.

The achievement of this threefold objective is enabled by Digital Platforms, i.e., initiatives aiming at combining digital technologies, notably IoT, Big Data and cloud, autonomous systems and artificial intelligence, and 3D printing, into integration platforms addressing cross-sector challenges. The Digitising Industry initiatives are aimed at a pervasive adoption of Information Technologies (IT) into Operations Technologies (OT), so they all implement the IT → OT way to do it.

Digital Factory Automation

Globalization has created a new and unprecedented landscape significantly changing the way manufacturing companies operate and compete. Fierce competition, shorter response time to market opportunities and competitor’s actions, increased product variations and rapid changes in product demand are only some challenges faced by manufacturing companies of today.

The increasing demand for new, high quality and highly customized products at low cost and minimum time-to-market delay is radically changing the way production systems are designed and deployed. Success in such a turbulent and unpredictable environment requires production systems able to rapidly respond and adapt to changing markets and costumer’s needs. To capitalize on the key markets opportunities and win the competition for markets share, manufacturing companies are caught between the growing needs for:

Implementing more and more exclusive, efficient, and sustainable production systems to assure more efficient and effective management of the resources and to produce innovative and appellative customized products as quickly as possible with reduced costs while preserving product quality. Creating new sources of value by providing new integrated product-service solutions to the customer.

In order to meet these demands, manufacturing companies are progressively understanding that they need to be internally and externally agile, i.e., agility must be spread to different and several areas of a manufacturing company from devices data management at the shop floor level rising up to business data management while going beyond the individual company boundaries to enterprises data management at the organization level. Therefore, agility implies being more than simply flexible and lean.

Flexibility refers to the ability exhibited by a company that is able to adjust itself to produce a predetermined range of solutions or products, while lean essentially means producing without waste. On the other hand, agility relates to operating efficiently in a competitive environment dominated by change and uncertainty.

Thus, an agile manufacturing company should be capable to detect the rapidly changing needs of the marketplace and propagate these needs to the lower levels of the company in order to shift quickly among products and models or between products. Therefore, it is a top-down enterprise-wide effort that supports time-to-market attributes of competitiveness.

Thus, to be agile a manufacturing company needs a totally integrated approach i.e., to integrate product and process design, engineering and manufacturing with marketing and sale in a holistic and global perspective. Such a holistic and global vision is not properly covered in the manufacturing company of today.

The vision of decentralizing the automation methods towards gaining additional flexibility in integrating new technologies and devices, while improving performance and quality is not new.

Earlier efforts towards the decentralization of the factory automation systems have focused on the adaptation and deployment of SOA (Service Oriented Architecture) architectures for IoT devices. However, SOA architectures tend to be heavyweight and rather inefficient for real-time problems, and therefore cannot be deployed on the shop floor without appropriate enhancements.

Furthermore, SOA deployments tend to focus on specific application functionalities and are not suitable for implementing shared situation awareness across all shopfloor applications. In recent years, the advent of edge computing architectures has provided a compelling value proposition for decentralizing factory automation systems, through the placement of data processing and control functions at the very edge of the network. Edge computing is one of the most prominent options for implementing IoT architectures that involve industrial automation and real-time control.

Nevertheless, the adoption of decentralized architectures (including edge computing) and IoT from manufacturers remains low for several reasons, including:

Lack of a well-defined and smooth migration path to distributing and virtualizing the automation pyramid: The vast majority of manufacturers have heavily invested in their legacy automation architectures and are quite conservative in adopting new technologies, especially given the absence of a concrete and smooth migration path from conventional centralized systems to decentralized factory automation architectures. The virtualization of the automation pyramid could greatly benefit from a phased approach, which will facilitate migration, while also ensuring that the transition accelerates production, improves product quality and results in a positive ROI (Return-on-Investment).

• IoT deployments and standards still in their infancy: IoT deployments in manufacturing are still in their infancy. They tend to be overly focused on unidirectional data collection from sensors for remote monitoring purposes while being divorced from the embedded and real-time nature of plant automation problems. At the same time, they tend to ignore the physical aspects of automation i.e., they pay limited emphasis on IT infrastructure aspects. Furthermore, despite the emergence of edge computing architecture proposals for manufacturing, their implementation is still in its infancy.

• Lack of shared situational awareness and semantic interoperability: There is a lack of semantic interoperability across the heterogeneous components, devices and systems that comprise CPS-based automation environments for manufacturing. Distributed IoT/CPS components provide non-interoperable data and services, which is a setback to creating sophisticated production automation workflows.

• Lack of open, secure, and standards-based platforms for decentralized factory automation: The distribution of automation functions on the shopfloor is usually implemented in an ad-hoc fashion, which may not comply with emerging architecture standards. There is a lack of architectural blueprints for decentralized factory automation based on future internet technologies. Furthermore, emerging future internet platforms have a horizontal nature and are not built exclusively for the manufacturing domain.

• IoT Production Workflows – Systems-of-Systems Automation: The next generation of industrial infrastructures is expected to be complex System-of-Systems (SoS) that will empower a new generation of industrial applications and associated services that are actually too hard to implement and/or too costly to deploy. There are several definitions of an SOS in the literature, however, the definition that best fits the considered application context/domain is the one provided where you are defined as:

“large-scale integrated systems that are heterogeneous and independently operable on their own but are networked together for a common goal. The goal may be cost, performance, robustness, and so on”.

The state-of-the-art industrial automation solutions are known for their plenty of heterogeneous smart equipment encompassing distinct functions, form factors, network interfaces and I/O (Input/Output) specifications supported by dissimilar software and hardware platforms. Such systems are designed, implemented, and deployed to fulfil two main objectives:

• To convert raw materials, components, or parts into finished goods that meet a customer’s expectations or specifications.
• To perform the conversion effectively and efficiently to guarantee a certain level of performance, robustness and reliability while keeping the costs low.

To do that, coordination, collaboration and, thus, integration and interoperability are extremely critical issues. Several efforts have been made toward the structural and architectural definition and characterization of a manufacturing company and its production management system

Iot applications in Manufacturing

• Proactive Maintenance: maintenance activities and procedures are always under high pressure from the top management levels of a manufacturing company to guarantee cost reduction while keeping the perfect working conditions of the machines and equipment installed in a production system and to assure a certain degree of continuity in the productive process and – at the same time – the safety of the people that are part of it. To do that, several policies and strategies for maintenance have been defined, developed and adopted, namely:
• Corrective Maintenance (CM): Corrective Maintenance also called Run-to-failure reactive maintenance is the oldest policy and envisions the repair of a failure whenever it happens. It implies that a plant using run-to-failure management does not spend any money on maintenance until a machine or system fails to operate.
• Preventive Maintenance (PM): Preventive Maintenance is a time-driven policy and envisions the advanced definition of the time of intervention in order to anticipate the failure of the complex system. In preventive maintenance management, machine repairs or rebuilds are scheduled based on the meantime to failure (MTTF) statistic.

• Predictive Maintenance (PdM): Predictive Maintenance also called condition-based maintenance is a policy that envisions the regular monitoring of machine and equipment conditions to understand their operating condition and schedule maintenance interventions only when they are really needed.

• Proactive Maintenance (PrM): Proactive Maintenance is a total policy that is not “failure” oriented like the others. As a matter of fact, proactive maintenance envisions not the minimization of the machine/equipment downtime but the continuous monitoring of the machine and equipment conditions with the main objective of identifying the root causes of a possible failure and/or machine breakdown and proactively scheduling maintenance intervention to correct the abnormal values of the root causes.

• Thus, in a proactive maintenance policy, the minimization of the downtime is only the consequence of a strategy that is aimed to improve the machine/equipment health during its lifecycle and to assure overall high production system productivity, reliability, and robustness while paradoxically reducing the number of maintenance intervention. Proactive maintenance is a necessary state in the main path to effective maintenance. It has not been thought of as an alternative to predictive maintenance but as a complementary approach to predictive maintenance in the direction of effective maintenance.

• Integration of secondary processes within the main control: IoT based technologies can be deployed to provide more data about machines and equipment during their operation. Such information can be used to model the machine/equipment behaviour for the sake of failures/breakdowns detection.

• Modernization of low-tech production systems: IoT based technologies can be deployed in low-tech production processes, i.e., production processes that are not natively ready for industry 4.0 and make them industry 4.0 compatible.

• IT Integration: IoT technologies can easily provide data to all the layers of the automation pyramid enabling a true cross-layer integration.

• Maintenance engagement: IoT technologies can enable a better engagement of the maintenance department in the health of the overall production system.

Mass Customisation

The deployment of IoT technologies in virtual manufacturing chains and decentralized factory automation systems enables the reduction of the production batch side and facilitates mass customization. IoT devices can be deployed across the supply chain (e.g., even at retail locations) to obtain insights into customers’ preferences. At the same time, the flexible integration of new technologies (such as stations, sensors, and devices) facilitates the reduction of the batch size. Overall, IoT supports mass customization across all points of the supply chain.

• Human-centred production scheduling (notably in terms of workforce allocation): IoT technologies (such as RFID tags) can be deployed to provide access to the users’ profile and context, thus enhancing conventional techniques for distributing tasks among workers to take into account the (evolving) profile and capabilities of the worker, including his/her knowledge, skills, age, disabilities and more.

• Workplace Adaptation: IoT devices such as sensors and PLCs (Programmable Logic Controllers) can provide the means for adapting the factory workplace operation and physical configuration (i.e., in terms of automation levels and physical world devices’ configuration) to the characteristics, needs and capabilities of the workers, to maximize their performance and the overall productivity of the plant, while also maximising worker satisfaction.

• Worker’s engagement in the adaptation process: IoT technologies can enable the comparison of the performance of a worker in a given task with the corresponding performance of skilled workers, to finetune the task distribution and workplace adaptation processes. Feedback on the performance of workers will be derived based on RFID tags and wearable devices, which can provide information about the worker’s stress, fatigue, sleepiness, and more.

• Enhanced Workers’ Safety and Well-Being: The deployment and use of IoT wearables (such as Fitbit devices) can enable the tracking of the workers’ activity levels. Fitbit data can be accordingly used to enhance workers’ safety and reduce healthcare and insurance costs for the manufacturers.

Outlook and Directions for Future Research and Pilots

Despite the development of these technologies, there are still technological challenges, especially in the following areas:

• Security and Privacy: IoT data on the shop floor varies in terms of volume and velocity while including structured, unstructured, and semi-structured data sources. At the same time, IoT deployments in manufacturing comprise multiple devices, which must be secured on the network. Holistic multi-layer approaches to security are therefore required to ensure the safeguarding of personal data and control over the flow and exchange of sensitive information across the manufacturing chains and/or the shop floor industrial network.

• Big Data Analytics: Manufacturers need to convert data into actionable insight. Given the large volume of data, this is a significant challenge. The generation of business-critical insights based on these data is still in its infancy since data stemming from the manufacturing environment tends to be largely underutilized.

• Adoption of Edge Computing: Emerging edge computing architectures have distinct advantages for the implementation of decentralized architectures, yet they have not been widely deployed yet.

• Need for Standards-based Reference Implementations: Recently, standards-based organizations (such as the Industrial Internet Consortium) have produced reference architectures for industrial automation and the integration of digital enterprise systems in the manufacturing chain. The provision of the reference implementation of these standards will pave the way for their wider adoption and sustainable use by manufacturers.