The IIoT journey starts with remote telemetry

Remote telemetry units for monitoring and control can be used in intelligent power, utility, broadcast and transport, says Craig Abbott.

The IoT has a vision of reporting data from low cost sensors across ubiquitous communications links and into huge databases within the cloud for analysis. With more data, the aim is to glean additional information to help optimise process plants, cities or even everyday lives.

There is immense value in being able to optimise operations and to detect and respond faster to impending issues. For example, remote telemetry units (RTUs) have been used for decades to monitor power consumption and battery backup at telecommunications towers. In the event of a grid failure, the RTU allows operators to continue to monitor assets to ensure continued, uninterrupted operation.

More recently, they have been deployed at wind turbine towers to monitor production – rather than consumption – at each one. The RTU has to have enough capacity to manage the limited number of I/O points that are required at each power generation tower. In the manufacturing sector, industrial IoT (IIoT) provides similar benefits for a vast range of equipment such as pumps, valves, compressors and even railway lines and potable water.

RTUs work on the simple premise that, if the condition of an asset is understood, it can be managed efficiently and respond quickly to change. In a process plant, covered by a Wi-Fi or 5G network, and with servers in the cloud or in a nearby, air-conditioned control room, RTUs gather information about critical assets.

Industrial intelligence

One of the drivers for the IoT is the ease of access to communications networks to transport data to the cloud and the scalability of cloud servers to gather and manage large volumes of data for analysis. Industrial facilities, however, can be a closed shop. Cyber security concerns mean that external communications there are often highly restricted, or even non-existent. Many industrial applications concern public safety, such as the transport of people or of hazardous materials, or the production of food, beverages and medicines for human consumption.

The risk of any external access to control systems precludes the open connectivity that is a core driver for the IoT. For some people, the lack of connectivity is enough to suggest that IoT is not suitable for industry. IoT concepts, however, can be applied to an industrial setting.

This does mean using sensors that are appropriate to the industry, with any required hazardous-area or hygiene approvals, communications networks that are restricted via a physical LAN or wide-area VPN, and potentially an on-premises, rather than cloud-based, server. The result should be the same – increased monitoring of assets, collecting additional data from disparate sources and analysing data to produce actionable information.

RTUs as edge computers

A relatively recent addition to the IoT family is the concept of an edge computer. Cloud servers are physically located in a central location and are tasked with managing data updates and requests from anywhere in the region. With rising volumes of data, there are increasing loads on both the main server and the communications networks that connect to it.

Conceptually, there is a need for another type of server that sits at the edge of the cloud, perhaps at a state or city level, which takes on the role of a data pre-processor. It should filter the data stream, removing data that is not informative, and where possible, respond locally to change, rather than defer to the main cloud server for every event. These actions reduce the load on the communications network and reduce latency in any responses.

The first thing to recognise about applying industrial IoT (IIoT) concepts to supervisory control and data acquisition (Scada) is that the industrial edge computer already exists. RTUs collect data from sensors at the remote location and process it for an immediate, local response. There is no latency in sending data to a central server and waiting for it to respond. It also solves the problem of outages when there is no link to the main server. RTUs are autonomous and can maintain local control for extended periods without supervisory oversight.

They are also data concentrators but sending every data sample to a main server will quickly overload a communications link. RTUs minimise congestion by concentrating the data to only what is necessary. An RTU can sample the level of a water tank, or the voltage on a train power line, several times per second for alarm and control purposes, and only send key details to the main server. This keeps communications traffic to a minimum. For example, sending a minimum, maximum, average, total and standard deviation for a data point each hour, rather than every one second sample, reduces communications traffic by 99.9%. This allows RTUs to provide an insight into the remote systems when there is next to no available bandwidth.

As analysis tools, RTUs collect data from local sensors, analyse it and then respond to change. The typical algorithms in use today are process control related, just like a programmable logic controller (PLC).

As the IIoT evolves, more and varied functions are being developed. For example, the Kingfisher CP-35 runs a Linux operating system on a 1GHz processor. This is a significant level of processing power available in the field, dedicated to analysing data from a single location. A fleet of 100 of such RTUs across a network gives 100GHz of processing power.

The first rule of Scada is that communications will fail. In addition to being autonomous controllers, RTUs must be data loggers. While offline, an RTU will maintain a store of data that should be sent to the central server, uploading it later once the link is restored. The latest RTUs can store hundreds of thousands, and potentially millions, of events. To put this into perspective, 100,000 events is about 140 days’ worth of hourly averages from 30 remote sensors.  In the IoT, the role of an edge computer is to pre-process data and act before the data is passed on to the main server. This allows a faster, low latency response, and minimises traffic between the central server and the edge. This is exactly the role that an RTU can fulfil in IIoT.

Industrial considerations

Unlike rack-mounted computers, located in data centres, edge computers in the IIoT are where RS485 serial links and 4.0mA-20mA sensor cabling meet 4G and 5G network connections. This is where local sensors and remote communications come together.

A typical rack mounted PC is not made to survive the punishing thermal conditions, damaging electrical events and potential outages in power supply. RTUs, on the other hand, can withstand extreme temperatures from -40oC to +85oC. They have up to 5,000V of isolation between the sensor cabling and the main processor, to protect from spurious events. The power supply also manages the recharging of a local battery for continued operations during loss of primary power supply.

In these rugged, remote environments, system damage should be considered inevitable. RTUs can be fitted with dual power supplies, dual CPUs and multiple, redundant communications modules. In the event of a failure, backup components can immediately take over, ensuring that remote monitoring and control can continue, uninterrupted until a maintenance crew arrives.

A common IIoT challenge for design engineers is knowing how to start. The low risk option is to select an RTU with protection and safety features. Many engineers will already work with RTUs, have a Scada server for receiving and viewing data, and have access to engineers or an integration partner who needs no special training.

These are all the tools necessary to start asking questions about assets and to have the required data collected and analysed. Once an IIoT trial is running, then it is possible to build towards more advanced IoT concepts, like MQTT brokers, integrated HTTPS and predictive analytics.