1 Introduction

Since its first appearance less than ten years ago the digital transformation has evolved into a driver regarding both, business and private life. It has become a major buzzword and a hype which itself was driven by technology and technical improvements in computerization, increasing processor capabilities and network capacities at decreasing prices at the same time. Thus, methods even well-known from the past have experienced a revival on the base of these new technical capabilities. They are coming along with new methods being developed in the context of digital business.

The German term “Industry 4.0” invented some seven years ago characterizes the fourth industrial revolution [1, 2] and in the meantime it is recognized worldwide as a synonym of industrial internet of things (IIoT).

Industry 4.0 in a literal sense focuses the industry’s digital transformation, however, it is widely used outside industry as well. Regarding manufacturing the industrial internet of things covers the overall manufacturing ecosystem (Fig. 1) including external partners like suppliers, service providers and customers as well as processes such as procurement, logistics and intralogistics, industrial engineering, quality management and assets services and maintenance etc.

Fig. 1.
figure 1

Manufacturing ecosystem.

Full size image

Even the product itself becomes part of the manufacturing ecosystem since its intelligence – either intrinsic by integrated processors or extrinsic e.g. by intelligent workpiece carriers – enables the product to communicate with the production line thus upscaling the product to a “smart product”.

2 IoT Impact on Manufacturing Ecosystem

2.1 Production Logistics

In the area of intralogistics and materials handling it is obvious that smart assistants and robots help increasing efficiency and improving process quality. Robot carriers help moving raw parts from stock to the production line and finished parts vice versa. These driverless robots are equipped with networked sensors and operate autonomously without any track-guiding [3]. It is even possible to move the complete shelve from stock to the production line, in order to keep local stock in the production as low as possible [4].

If the raw material is available at the line mounting and feeding the machine is assisted using augmented reality methods by either projecting the “how-to information” directly to the working space or using augmented reality devices for projection of indicators, signs, or entire holograms. Thus, by intelligent assistants even unskilled workers will be enabled to perform mounting and loading [5, 6]. By motion capturing the operator can be controlled and guided. Sometimes, this even will be cheaper than installing a fully automated robot solution.

2.2 Manufacturing

When investigating IoT solutions in manufacturing the first ideas that come to mind are those of increasing process efficiency, agility and flexibility. “Lot-size 1 by methods of mass production” is the embodiment for industry 4.0 [7]. It leads to agility and flexibility in production to be prepared to react promptly to customer’s demands. However, it induces a highly intelligent and automated manufacturing processes with minimized manual work and tooling times. Nowadays this can be compared to the crowning of the smart factory initiatives as described in detail by Soder [8]. Obviously, it cannot be substitute to it but the prosecution and it will only deliver best output if it is applied to lean production processes. First of all muda has to be eliminated from the process before intelligent automation and IoT methods and tools are introduced. Usually, cyber physical production systems (CPPS) serve to integrate both, physical processes in combination with computational intelligence and networking capabilities (CPS) on the one hand side and manufacturing science and technology on the other. Together with comprehensive process and system data these CPPS set up the digital twin of the system, production line, or plant. Digital twins serve to simulate processes, systems’ operation, or e.g. the cooperation of humans and machines.

However, setting up of industry 4.0 environments in existing plants may raise unexpected problems, e.g. when the existing machines are not suitable for integration and can only be upgraded in the operation context, i.e. the overall control and operating system including its programming framework has to be upgraded [9]. Larger amounts of effort and money must be invested in these cases.

2.3 Distributed Manufacturing

Not only since discussing the fourth industrial revolution, distributed manufacturing is a current method in industry. However, within the context of industry 4.0 it reveals its full potential. E.g. 3D printing offers the opportunity to optimize the use of production capacities around the globe within an internet-based network of 3D printers. In addition, not only the manufacturer’s printers can be used but printers at the customer’s or a partner’s facility as well. This principle is similar to the early grid computing ideas in IT [10], when distributed (private) computing capacities are used by foreigners during times of non-usage (e.g. at night). It can be enhanced by an online video monitoring system for customers to monitor the production [11]. The monitoring installation can even include some intelligence like masking the printing space or shape/progress control in order to only show released content [12]. This approach becomes most powerful, since the company not only develops, produces and sells 3D printers but in addition provides 3D printing services at multiple company locations. Thus, company’s and customer’s printers can be made available to multiply the production (printing) capacity [13].

2.4 Quality Management and Predictive Maintenance

Regarding quality management in a smart factory a variety of applications and parameters can be used for quality control and documentation. Acoustic emissions (structure-borne sound) [14] can be detected during assembly, machining or bonding processes, e.g. when connectors are snapping in or when friction welding is used. In the latter case even measuring the machining forces can be used as criteria. By optical shape and deformation monitoring the turning and milling processes can be monitored and documented. Measuring micro forces and temperature in glass grinding processes will deliver parameters for the process stability.

In addition, these and further physical signals are the base for predictive maintenance within a plant. Since by means of smart sensor equipment machine states are monitored and analyzed upcoming failures can be predicted and fixed before outages occur. In this respect, especially acoustic emission is well known since the seventies of last century (e.g. [14]). Its application to monitor tool wear and fraction during turning and drilling under shop floor conditions was investigated by C. Scheer at ETH Zürich in 2000 [15]. Furthermore, commercial tools already exist for monitoring and analyzing vibrations of motors in order to predict abrasion or failure of bearings [16]. In the famous German movie “Das Boot” (1981) the engineer used an ear trumpet for monitoring the machine sound. He used his long-lasting experience to interpret what he has heard. Today’s integrated IoT solutions in some cases are even based on the same principles, however using modern analyzers as shown in the requirements section of this article. Especially for monitoring motor states further parameters can be investigated, e.g. power consumption and temperature of the motor coil etc. (see Fig. 2).

Fig. 2.
figure 2

Direct and indirect production (process) data and parameters.

Full size image

Within a smart factory there are many more examples for smart applications helping to increase plant availability and uptime. When moving parts or goods within the production line conveyor belts are most widely used with clamps for fixing the parts safely. A reliable clamp operation therefore is obligatory for process safety as well as safety at work. The lock and release times of these clamps can be used to predict clamp failure due to contamination [17]. This example shows, that a single IoT application not necessarily needs to be very complex, however, the overall complexity is increased by each and every application.

3 Product Related Aspects

As mentioned earlier the product itself will become part of the digital transformation and its manufacturing ecosystem. It is expected that in future the product will control the production process autonomously [18]. In addition the product will carry a lot of digital information regarding manufacturing process and history, quality and certification, material and properties, etc. If the product does not have any internal data storage the information is stored outside the product body, e.g. in the cloud with a unique identifier (e.g. IP address, QR code, RFID) as a reference.

Many decades before the IIoT has been postulated to be the fourth industrial revolution, material properties as well as material models have been investigated in order to simulate manufacturing processes. Kraft et al. simulated the die pressing and sintering of ceramics to optimize the die by means of finite element (FE) analysis [19]. The author of this article himself used combined experiments and FE simulations to determine material properties [20, 21]. As in-process methods combined with IoT sensor signals the FEM analysis will increase the model’s accuracy significantly. However, even todays commercial computers are not powerful enough, to deliver results within acceptable response times.

4 Requirements

The industrial internet of things not only changes the world of manufacturing. Technical prospects on the one hand side and technical contraints on the other are defining the requirements of the manufacturing ecosystem. Since smart devices and cyber physical (production) systems are available from multiple suppliers, it becomes obvious, that a common integration space is required for setting up the communication within the system. And it needs data: meta data from the machines in the production line, oftenly already generated when the machine itself is developed and produced (machine lifecycle data), and data from the production process (field data). This for sure includes a common semantics of information and data collected from the field. Furthermore the data acquisition has to be absolutely robust and reliable in the industrial environment. However, solely collecting data must not be the goal of the overall effort. The data have to be analyzed by appropriate means in order to enable further actions in the system. This sometimes requires local processing power, e.g. by edge computing, and decentralized data storage capacities, e.g. in the cloud. When analyzing field data for process monitoring and predictive maintenance thresholds need to be defined. These can either be calculated analytically or by model assumtions an iterative solutions or by recording reference values from the field. The latter represents the “expert’s experience” and is mostly used in the field.

In addition to integrating and interfacing systems and machines the ecosystems is extended to business partners like suppliers and customers. Therefore several more interfaces have to be set up, further increasing the complexity of the manufacturing ecosystem. Inevitably this leads to the demand for a platform approach with interfaces to business partners, machines and systems, and software and applications (Fig. 3).

Fig. 3.
figure 3

Manufacturing ecosystem with data integration platform approach.

Full size image

Of course, availability and all aspects of cyber security are of utmost importance for a reliable manufacturing framework in industry 4.0. The progress in this area is encouraging, however, the security issue will never be solved a full 100%.

5 Conclusion and Future Scope

The business transformation driven by the internet of things and services already has a deep impact on the manufacturing processes. Manufacturing like many other industrial disciplines is continuously growing together with information technology with many chances and opportunities but with restrictions and consequences as well.

If data collected from both, the machinery and the field level is significant and reliable it enables analyses and measures to improve process quality and efficiency in production. Nevertheless, the chances arising have not been completely implemented up to now. Even in spite of the awesome progress in computer technology some areas still lack technical capabilities. Here artificial intelligence and quantum computing will provide future enhancement.