Key technologies defining smarter factories: connectivity

At the foundation of Industry 4.0 implementations will be the means to transport data, with reliable communication technology providing digital connections to the cloud.

In what is the first of a series of blogs describing each of the main aspects of Industry 4.0, the numerous connectivity options currently out there (as well as the ones just about to be introduced) will be highlighted. In each case, we will be looking at how they balance the requirements of range, bandwidth, power consumption and security.

Wireline protocols

Ethernet has been the basis of industrial communication for many years already. As an open standard it is served by many different suppliers. There are numerous varieties of Ethernet available (depending on the data rate that needs to be supported), with structured cabling around factories providing robust and reliable connectivity, linking directly into the computing infrastructure. Also, through the use of protocol converter modules, there is scope for older equipment that is reliant on legacy serial protocols (like RS485 and RS422) to interface with Ethernet-based systems.

As most sensor data is relatively low bandwidth, often sampling once or twice a second, the output of each sensor node doesn’t represent much of a problem individually. The challenge is more one of scale, with potentially thousands of sensors all linking to central servers (or to the cloud). The increasing prevalence of cameras/image sensors, all generating high-resolution video, is also impacting on available bandwidth. Structured cabling, such as Cat5 and Cat6, supports links with 10Mbit/s, 100Mbit/s and 1Gigabit/s data rates.

 It is certified to mitigate the electromagnetic interference (EMI) effects present in most industrial environments (that can potentially lead to packet loss). Time-sensitive networking (TSN) versions mean that deterministic communication can be provided (without any uncertainty about latency) – ensuring that each step in any complex industrial process is fully synchronised.

HART is a bi-directional protocol supporting 1.2kbps data rates that uses a 4 to 20mA link commonplace in industrial settings. It provides two channels (one analog and the other digital), and offers a high degree of reliability plus strong EMI immunity. Another alternative is Modbus. This is a serial protocol that was originally developed for programmable logic controllers (PLCs), but is now an open-source and royalty-free communication platform that can be applied to all manner of industrial systems. The data types are based around the ladder logic used in PLCs.

Wireless protocols

In an industrial context wireless has several notable advantages over wireline in relation to its comparative ease of installation (and lower associated deployment costs), system flexibility and suchlike. However, this needs to be weighed against security vulnerabilities and potential interference problems.

WirelessHART is a wireless version of the HART protocol and is standardised as IEC 62591. This adds a mesh networking capability, where each node serves as a router for messages from other devices. Taking an approach to connectivity based on this protocol avoids the need for an access point or gateway. This simplifies the network architecture and means that it is more resilient to any single point of failure, as data can be sent via a number of alternative nodes.

Located in the 2.4GHz frequency unlicensed ISM band, the Bluetooth Low Energy and Bluetooth 5.0 protocols provide cost-effective, low-power point-to-point links that are highly interoperable. Terminals are easy to find, as every tablet and mobile phone handset on the market these days is Bluetooth enabled. A mesh-based connection protocol has been added to Bluetooth 5.0 to allow many nodes to link together, thereby making the factory’s network more robust.

The 802.11 family of WiFi protocols are still viable technology for Industry 4.0 deployments, mainly because of their ubiquitous nature and the incredible interoperability they offer. WiFi nodes at 2.4GHz and 5GHz are easily available and can provide high-bandwidth links to a nearby access point. However, it should be noted that they are relatively high power and also fairly weak when it comes to security (which could mean that the factory is vulnerable to industrial espionage, etc.). Channels in the 2.4GHz band get filled up quickly with general networking, but the 5GHz band is increasingly popular for certain industrial links.

Cellular technology

Cellular technologies such as GSM (2G), 3G and 4G LTE packet-based modems can provide long-range wireless connectivity with the same elevated levels of security as mobile phones. As cellular protocols are not in the ISM band then, of course, there is an operating cost that needs to be factored in – so they will not be economically viable in a fairly large proportion of use cases.

Widespread rollout of 5G will soon be underway (with some private deployments already being initiated by large industrial players). Through 5G, significantly lower latency (down from 200ms to under 5ms) will be supported, though currently there isn’t any provision within the standard for true deterministic operation (talks about how to incorporate this into the standard are just beginning).

This will enable rapid reaction to any problem within an industrial process. It will also help to reduce the power consumption of industrial wireless infrastructure, with faster setup and breakdown of data connections. A derivative of LTE, narrowband IoT (NB-IoT) brings the security of cellular links and the faster setup/breakdown times of 5G, but draws less power than a 5G link will because it has a restricted bandwidth.

The software aspect

Industry 4.0 links are not just about the hardware involved. Basic IP links are possible, but for managing data in the cloud other, markedly better-optimised, messaging protocols are currently gaining traction. Though lightweight (to address device constraints), they comprise the necessary security and quality-of-service (QoS) functions needed for effective industrial communication. MQTT and AMQP both have TCP/IP as their foundation, while CoAP utilises more streamlined UDP/IP-based transportation.

These can run over all the different hardware-based links and are all standardised, being extremely easy to feed into a cloud-based Industry 4.0 service. Each of the wired, wireless and software protocols detailed can be combined to provide the data that feeds a database (or possibly a digital twin model, as we will discuss later in this blog series). Through this data, the way in which an industrial process is configured, tested, implemented and monitored can be totally transformed.

Conclusion

Data management is vital to Industry 4.0. Real-time data can be used to identify problems in manufacturing processes as they actually happen, while large amounts of historical data (stored on local servers, or in the cloud) can also be examined in order to uncover patterns. As we will see later in this blog series, machine learning (ML) pattern recognition mechanisms can enable the health of equipment to be determined.

This can also be compared with data from the same type of equipment at other sites around the world, then aggregated together to improve the accuracy of predictive maintenance, etc. Necessary parts can be ordered in advance, and replacement work done at a point that minimises disruption to the manufacturing process and avoids downtime.

All this requires a vast array of sensor devices constantly monitoring the equipment connected to the servers – and this will be the subject of our next blog.

Upcoming blogs in this series include:

1. Key technologies defining smarter factories – connectivity: To begin the series we will be looking at the very foundation of any industrial automation system, namely the supporting communication infrastructure. The various applicable wireless, wireline and cellular protocols will each be described, and their respective advantages outlined. 
2. Key technologies defining smarter factories – sensors: In this second blog, the wealth of sensor technology now being deployed to make production and logistical activities more effective will be examined. The value of devices that can employ energy harvesting will also be covered.
3. Key technologies defining smarter factories – the rise of cobots: This blog will detail how the new generation of robotic systems will work alongside human staff, enabling collaboration that draws on the strengths of each party and in turn benefits production. 
4. Key technologies defining smarter factories – digital twinning: In this blog, the array of possibilities that the creation of digital models could enable within an industrial context will be examined. This will cover everything from improving product development activities, to optimising floorplans, and extending the operational lifespan of equipment and predictive maintenance.
5. Key technologies defining smarter factories – AI: This will give an overview of the multitude of ways that artificial intelligence (AI) is set to transform the industrial environment, by boosting operational efficiencies, enhancing quality control procedures, assuring better worker safety, bolstering security and making facilities more responsive to situations that could halt production. 
6. Key technologies defining smarter factories – data security: To conclude the series, this blog examines the issues arising from potential data breaches. It will look at what can be done to protect the data that smart factories will constantly be generating from the serious threats posed by industrial espionage, ransomware, etc.