Every day, the world becomes more connected. Artificial intelligence interprets our data and our lives become increasingly automated. It will not be long until many of our daily activities — grocery shopping, commuting or setting up an appointment — are done for us.

Without advanced programming knowledge, we have to trust that these program developers transmit our data securely and accurately. And then, what are we to do with that data? How are we to be sure that it is utilised effectively?

There are three hurdles when it comes to digital optimisation: data collection; secure transfer; and effective utilisation.

Data collection

Things are quite different from when nuclear plants first started operating in the late 1950s. Back then, trained technicians would take their knowledge to the valves or gauges and determine themselves whether everything was in order. In addition to the time and manpower burden this placed upon plant personnel, proper operation also assumed that they were all equally and properly trained.

This became an issue when an instrument would have to operate outside the recommended operating range to ensure that it would open or close. For example, positioners which typically operate at 4-20mA might be sent 0-25+mA to account for a potential miscalibration and ensure that it was 100% open or shut. This causes unnecessary wear, shortening the unit’s lifespan or making it less effective over time.

Modern distributed control systems standardise this process verification quickly and correctly at the touch of a button ensuring precise outputs. This is done by first sending a 4mA signal to the positioner (at which point no air is sent to the actuator); beginning to send air at 4.08mA and linearly increasing the signal and air until the valve is 100% open at 19.92mA. Smart positioners can offer a range of additional diagnostic capabilities such as valve monitoring, friction analysis, and air consumption testing – much of which can be done while the valve is still in service.

This can be done because smart positioners can interpret and send data in terms of air pressure and current signals.

If a 12mA signal is sent, the positioner should send the valve to 50% travel. If the valve does not go to 50% travel, you know that the valve will have to be troubleshot, just as a check-engine light in your car might turn on when some measured parameter is out of tolerance. This system of command-and-respond is one of the fundamental principles of the HART® communication protocol and it allows you first verify to that the command has been received and second to determine what has happened. This is a big improvement on the analogue method of simply sending a signal and assuming that what should have happened, did happen. Now, even a message that was interrupted by some sort of interference will return an error and try again.

Four variables — pressure, level, temperature, and flow rate — are controlled and monitored in any process. These are read, converted to a digital HART analogue input (AI), and sent to the control system that may be managed by plant asset management software.

The control system and operator in turn interpret the data and send a 4-20mA HART analogue output back to a valve positioner, which commands a valve to alter the process.

Secure transfer

As there is so much data being sent and received from units, more wiring will be needed as more monitoring devices are added to the system. This is considered a necessary evil in nuclear plants, as the network must be totally secure from outside influences.

An encrypted wireless network can be a secure solution, with series of wireless transmitters and encrypted data in accordance with industry and plant standards.

The unit itself must be robust and able to withstand vibrations and electromagnetic disturbances that it may experience out in the field. It follows IEC 62591 is the first global industry standard for wireless communication and has been used commercially for the last decade.

How is this transfer secure? The HART protocol is a frequency shift key, which sends a digital signal superimposed over the analogue signal and “hides” the signal in 16 channels over a 2.4GHz spectrum. The channel frequencies are chosen with a specific, proprietary algorithm and change periodically. There is an additional layer of physical security in the write-protect switch housed within the controller unit that makes the unit read-only. To flip this switch requires specific knowledge of the device itself, awareness of the device’s location in the plant, access to the location and passing multiple layers of security within the facility.

The distributed control system acts as a second layer of defence as it can detect potentially malicious activity through an unexpected change in the event log. In the event of critical failure, the device can be backed up and restored as a means of mitigating the problem.

Effective utilisation

The distributed control system takes all these different AIs and, through different control groups, sends analogue outputs to various devices to control the process. Emerson’s Ovation control system leverages the advanced architecture of the HART network and is scalable to be as controllable as budgets and resources allow.

The control system operates within a feedback loop where it takes a variable, compares it to the target value, and adjusts it accordingly. Over a direct, wired signal this is done very quickly, but in setting up a wireless network it is imperative to consider the path that the signal must take to be sent and processed. Each hop takes time to process and may encounter interference.

When plants first started implementing the mesh wireless network, they were underwhelmed with the speed of communication, because this processing time was not considered. Units that may have made several hops to receive a signal were not responding in a timely manner.

This may not be an issue for mostly-static control valves, but depending on the plant layout, a geographically distant process that requires tight control might run into issues in a purely wireless mesh network. A ‘star’ configuration of wired and wireless connections is much more efficient, as it minimises the number of hops and maximises the success rate of a command being received.

This is critical for some plants as they try to be more flexible in outputs. Knowing the health of valves is even more critical as more valves are frequently stroked. The distributed control system can schedule periodic, automatic tests like the previously mentioned air-consumption test to give a good idea of the health of soft parts like seals and O-rings, or a packing friction test, which verifies that nothing is wrong with the surface finish on the valve stem or that there is no galling in the valve trim.

Planning for the future

If nuclear plants are to become more economically competitive with less expensive options like solar or wind power, or if they want to become flexible, a move to digitisation is essential.

Digital instruments can read analogue signals and convert them to a digital HART signal. This signal is then encrypted and transferred upon request by the distributed control system via either wired or wireless connection. The data is interpreted by the Ovation system, which sends a signal back to a smart positioner based how the process needs to be controlled and the cycle repeats.

This continuous, automated system is the model that many industrial plants follow today. It is a simple concept but its flexibility allows it to grow as sophisticated as resources and time allow.