This website uses cookies primarily for visitor analytics. Certain pages will ask you to fill in contact details to receive additional information. On these pages you have the option of having the site log your details for future visits. Indicating you want the site to remember your details will place a cookie on your device. To view our full cookie policy, please click here. You can also view it at any time by going to our Contact Us page.

TSN: tomorrow’s information superhighway?

19 November 2018

Markus Weinländer provides examples of cloud-based quality control to demonstrate the possibilities of Time Sensitive Networking (TSN).

Time-Sensitive Networking (TSN) is considered by many to be a possible successor to today's fieldbuses, as it promises low latency with high availability – optimal conditions for ‘hard’ real-time communication in automation. In addition, TSN plays an important role for the sensor-to-cloud communication in the Industrial Internet of Things (IIoT), as automation and IIoT data in the future can be transported over the same network without interfering with each other. 

The possibilities are endless when it comes to extracting more information from the sensor data at the field level than is necessary for the actual machine control. Often mentioned is predictive maintenance. Here, sensor data are utilised to prevent possible malfunctions caused by the failure of individual components. The sensor data are collected and evaluated over time to early on detect anomalies that arise, such as an increased inrush current or excessive vibration during operation. 

Another example is cloud-based quality control. Components in production are already being measured to ensure they are still within the defined tolerances. For complex goods, though, other relationships can play a role in ultimately determining the usability of a product. On the one hand, certain unfavourable correlations may arise from the various measured values and require a refinement of the workpiece – even though the individual values are still within the tolerances. On the other hand, other parameters can be consulted, for example from an empirical database of returned products. Finally, the automatic evaluation of workpiece photos can assist in the decision of whether a product is allowed to be released to the customer or not. 
Large amounts of data from a variety of sources and the search for new, unknown correlations between parameters are scenarios that are predestined for a cloud-based big-data analysis and the use of artificial intelligence (AI). As a platform, an on-premise solution presents itself – the databases and algorithms are installed on a server (automation data center) close to the factory. After all, the demands on the speed are high, as the workpiece – if the worst comes to the worst – needs to be discharged at the end of the production line and, if necessary, be reworked. In addition, this solution requires a high-performance network.

In the past, so as to not jeopardise the actual automation task and still provide sufficient bandwidth for large amounts of data, such as camera images, separate networks would have been set up to transmit the real-time data between the PLC and the Ethernet-based sensors/actuators as well as other control units, and also to transport the sensor data and photos to the cloud. However, this approach is complicated. Many sensors continue to be classically connected to I/O modules or via special sensor protocols, such as IO-Link, which do offer important advantages; only complex devices, such as cameras, possess Ethernet-based interfaces. Moreover, the automation system also supplies relevant information for the apps, about the current work progress for example. 

A common network
So, it makes sense to employ a common network based on TSN. In the future, I envisage that the field level will be connected via PROFINET on a TSN basis to a SIMATIC S7 controller, which first of all will complete the immediate automation task. Via a communication module (communication processor, CP), the relevant data for the cloud-based processing are extracted and – in a TSN-capable network – sent to the analysis apps. Such TSN networks offer varying degrees of Quality of Service (QoS), depending on the application requirements. 

Communication towards the cloud is carried out via a so-called ‘best effort’ channel of the network. Sensors possessing their own Ethernet interface ( such as cameras) likewise send their data directly to the cloud – a TSN-capable SCALANCE switch from Siemens makes this possible. For the identification of the workpieces by 2D data matrix code, optical code readers, such as SIMATIC MV540, are used. This enables the creation of a complete, bundled data set from the wide range of information. 

At the same time, the TSN network is also utilised for the ‘hot’ automation, more precisely for the communication of controllers to the peripheral devices, or the communication between controllers. This is where the real-time capability of TSN comes into play: The controllers use a protected, high-priority TSN stream in the network, whic , in case of doubt, has priority over other communication connections. In doing this, the network guarantees the required low latency as well as the demanded bandwidth. ‘Above’ the automation level, the transmission takes place via a common Internet connection; as transition, a TSN-capable SCALANCE switch is used, which directs the TSN streams in the automation network, and also realises the transition for the vertical ‘best effort’ communication to the cloud.

Conclusion
The advantage of such an architecture is that new, digital quality assurance procedures can be implemented without the need for new automation concepts with complex, extra networking. TSN-capable networks therefore create the prerequisite for a high-performance IIoT at low cost.

Markus Weinländer is head of Product Management for SIMATIC NET at Siemens.


Contact Details and Archive...

Print this page | E-mail this page