The future topics of Industry 4.0, Smart Factory and Industrial IoT (Internet of Things) stand for the convergence of physical production components with information and communication technologies. As so-called Cyber Physical Systems (CPS) - or more precisely as Cyber-Physical Production Systems (CPPS) - systems, machines and production processes will work autonomously in self-coordinating networks in the future.

Two trends are flanking this development: advances in hardware technology and the widespread use of Ethernet-based technologies in all stages of production. With their open, universal architecture and built-in multi-core Intel processors, modern embedded systems are predestined to take on the task of data integration and processing. In addition, specialized IoT gateways (see box) help to transfer data from existing sensors in the machine world to the cloud environments.

To do this, the gateway collects the data in various fieldbus formats, makes it standard Ethernet-capable and communicates with the cloud services via mobile networks or a company's normal internet access. The following will show why the IPC, in cooperation with the specialized IoT gateways if necessary, is of great benefit for the breakthrough of the Industry 4.0/CPPS idea.

Valuable data combinations

The Spectra PowerBox 3000 series has sufficient power to take on the tasks of a motion control and monitoring unit, a fieldbus controller and an IoT gateway.

© Spectra

The detailed, real-time exchange of data between production and management systems can lead to greater flexibility and better utilization of production capacity thanks to smaller batch sizes. Cost structures in production can be optimized through the automated coordination of consumption and capacities. It is also conceivable that systems could use their 'knowledge' of the current values of the production parameters and initiate maintenance independently.

However, entering the age of Industry 4.0/CPPS is a thematically and technologically challenging task: Companies must carefully consider which data is actually useful for evaluation as part of cloud services? This also raises the question of how the information from sensors, actuators, processes and machines can be fed into standard Ethernet communication.

Diversity of existing infrastructures

Compact, uncomplicated and, not least, inexpensive I/O modules are generally used for the measurement tasks. Sensors such as thermocouples, resistance thermometers or strain gauges can be connected directly to the modules, which calculate the physical quantity via precise, powerful measuring amplifiers and communicate the data to the higher-level control systems via a fieldbus protocol. The data from the I/O modules or sensors is received and processed by PLC systems and triggers rule-based actions depending on the value. Several of these programmable logic controllers often exchange information about what is happening with a higher-level SCADA system or an industrial PC. Otherwise, hardly any information leaves its direct field and control level - horizontally in the direction of neighboring production islands or vertically in the direction of the MES/ERP control level.

As a result, the communication requirements of the Industry 4.0/CPPS idea literally reach their limits in established infrastructures. This is not least due to the fact that a large number of different fieldbus protocols (Profibus, Ethercat, etc.) have become established at the control level. A 1:1 connection between the proprietary systems and protocols would hardly be feasible in terms of cost and effort.

The cloud as a data hub

For this reason, from a data perspective, the cloud is playing an increasingly important role as an information and analysis hub for applications without hard real-time requirements. In addition to the physical infrastructure, every combination of data requires intelligent modeling on the software side so that the information is understandable in the literal sense. In addition to the precise definition of which data is provided in the cloud at which interval, this includes the aspect of comparable time (UTC - Coordinated Universal Time). The example of predictive maintenance illustrates why a certain amount of (processing) intelligence is required in advance in this context. The data that is important for the maintenance of a machine or system is generally not only accessible via a single channel. Instead, the operating temperature is supplied via a thermometer with a serial connection. Vibration measurements are taken via a vibration meter with an Ethernet interface and information on the rotational speed comes from the PLC. This data must first be consolidated and provided with a binding time stamp before it leaves the factory floor for the internal or external cloud. Otherwise, there is a risk of erroneous analysis results as, for example, the high vibration values are not correlated with the rotation speed prevailing at the same time, but with the lower values from the night before.

IPCs for consolidation

Industrial companies are therefore faced with the challenge of providing the 'hidden' information of their partly proprietary system world in a suitable format and in a generally understandable way - and at affordable conditions. The IPCs mentioned above are the first port of call. The ability to consolidate work and communication loads in one system helps the automation solutions to achieve greater efficiency for the vulnerable control tasks at field level. At the same time, the data from the local field and control level is available 'from a single source' in the appropriate standard format, including time stamps for Ethernet-based data traffic.

The IoT gateway NIO-100 solution architecture offers an onboard software platform for connecting both field devices and cloud services.

© Spectra

For the global visibility and usability of information from the IPC as well as local I/O data that does not have the luxury of IPC processing, a specialized IoT gateway such as the NIO 100 from Nexcom is an economical solution.

The IoT gateway serves as a 'link' between local I/O data from machines, systems or sensors on the one hand and the cloud services located on the internet for analyzing the data on the other. The programmer is responsible for the frequency and composition of the data used for analysis. A tool environment such as Nexcom IoT Studio helps them to correlate analysis function blocks with data flow lines of the production components and sensors using drag & drop.

To introduce the concept of predictive maintenance, a company now 'only' needs to create a collection of patterns over time to compile parameter values for maintenance-relevant situations. As soon as the continuous evaluation of the current values detects a comparable constellation, it would automatically initiate the associated maintenance process. In this way, the risk of machine failure is reduced and the efficiency of machine use is increased. In short: with the help of modern IPCs and IoT gateways, a company can already venture into the Industry 4.0/CPPS era without risk and benefit from the 'hidden' added value of its data.

Authors:
Eric Biank is Product Manager IPC Components/Embedded PC at Spectra;
Uwe Hollarek is Product Manager Automation/Communication at Spectra

Cloud connection via gateway

The Nexcom Nio 100 IoT gateway provides a solution for the global visibility and usability of information from the fieldbus level. The gateway, which is based on Intel's IoT platform with the low-consumption Intel Quark SoC X1021 processor, serves as the 'missing link' between local I/O data from machines, systems or sensors and the cloud services located on the Internet. It contains the necessary software components for secure cloud operation, including a secure boot function. The raw sensor data as well as the information processed by the IPC is collected via the multi-protocol interfaces of the NIO 100, while the data is communicated to the cloud either via the normal Internet connection of the system network or optionally directly via 3G mobile radio.

Data is transported to a public cloud environment, such as IBM Bluemix, via the industry standard MQTT (Message Queue Telemetry Transport). Compared to other rather heavyweight communication architectures, the IoT protocol uses the publish-subscribe principle to conserve resources. System components or programs publish or subscribe to information via a central broker instance.

The solution includes the Nexcom IoT Studio (based on IBM's Node-Red), a platform for developing your own IoT applications.

The limits of the cloud!

The cloud is increasingly becoming the focus of automation specialists. But where are the limits with regard to the real-time requirements of control systems? Uwe Hollarek, Product Manager Automation/Communication at Spectra, provides the answers.

Uwe Hollarek, Product Manager Automation/Communication at Spectra: "In principle, all applications without hard real-time requirements are suitable for operation in a cloud environment.

© Spectra

Is cloud technology unsuitable for machine control?

Yes, because I cannot see that operating the control program or obtaining control information as an Internet service could meet hard real-time conditions in the foreseeable future. After all, the laws of physics also apply to the Internet and cloud computing. At 300,000 km/s, the speed of light is a fundamental upper limit for the transmission of information. Cloud servers can therefore be a maximum of 300 km away if a response is required within 1 to 2 milliseconds.

Does the tactile internet, which is being discussed as part of the new 5G mobile communications standard, have greater potential for machine control?

First of all, the concept of a tactile internet is still purely a research project, which is admittedly exciting. There is still a lack of suitable processor types; there is a lack of experience in setting up and operating mobile edge clouds in order to keep control applications close to the machine at all times.

Only when the results of this research lead to industrially applicable technologies and mass products will the automation industry be able to get to grips with this innovation in practice. When, how and where an internet with a latency of one millisecond can be of use in production, robotics or mobile applications - such as autonomous driving - and whether it will even be financially viable for 'normal' automation applications, can only be reliably assessed in a few years' time at the earliest.

So a clear rejection of machine control in the cloud?

If you limit the statement to the point of 'hard real-time' - yes. Applications for hard real-time requirements must work close to the machine.

If the time aspect does not play such a dominant role, the integration of cloud services in production and manufacturing environments can lead to real benefits, as outlined in the article. For example, the programming and management of the control software or the calculation of new process parameters can be completely shifted to a cloud.