Valve closed? Wireless sensors in the petrochemical industry

What can industrial IoT mean for the chemical industry? The development of low-energy electronics and wireless communication could potentially solve dozens of problems. Five years ago, Maarten Weyn’s group at the University of Antwerp received two questions from major chemical companies. BASF was looking at the possibility of locating and directing people during emergencies, whilst Fabricom wanted to be able to monitor the location and status of manual valves during a maintenance process.

Weyn was able to combine those two questions. In both cases, you need two-way wireless communication that requires little maintenance. After two years of research, he was able to define a set of problems in the chemical industry that could be solved with electronics. The most challenging issue was measuring the position of manual valves. These valves were regularly removed from the pipework, after which a subcontractor would inspect and clean them. Weyn and his group built a prototype to measure how far a valve is open without the need for external devices.

‘Almost every major player in the petrochemical industry has a project or a trial with our sensor’

Low-energy protocol

There was considerable interest in the invention. Weyn set about assembling the right team and also approached the people from industry who had raised the original questions. This led to the spin-off Aloxy. Weyn moved to part-time work at the University of Antwerp to launch the company alongside three other active founders. They now have 18 employees and have raised €6 million in funding over five years.

The first product, Manual Valve Position Indication (MVPI), can now be found in plants across Europe, the US, and Singapore. ‘Almost every major player in the petrochemical industry has a project or a trial with our sensor,’ says Weyn. As the name suggests, the MVPI measures the position of a valve on the valve itself. For a vertical valve, this is relatively straightforward, but for a horizontal valve it is more complex. ‘We had to combine several sensors.’

The final solution consists of three components: first, the sensor that measures the valve’s movement and calculates its position; then transmission via an ultra-low-energy protocol; and finally, processing the data in the cloud. ‘This allows us to detect, for example, if someone removes the sensor from the valve and replaces it incorrectly. We can also predict the battery percentage and determine how long it will last. How much a sensor needs to measure and transmit data has an impact on energy consumption, and the longer the battery lasts, the lower the maintenance costs will be.’

Filtering

As an academic, Weyn believes the product should not exist solely to generate as much profit as possible — it must also deliver maximum added value for the sector and society. Ultimately, he wants to help the sector become more sustainable, as he expects petrochemicals to remain essential for a very long time. ‘We want to offer an optimal solution across the board.’

As an example, Weyn cites a client who was able to dramatically reduce his travel costs. ‘This client in America drives 100,000 miles every week to check valve positions. If you can halve that through wireless monitoring, you’ve achieved an enormous reduction in costs — both financially and in terms of environmental impact.’ Aloxy initially focused on the European sector, where mid- and downstream petrochemicals are concentrated, but there is also a great deal of upstream activity in America. ‘That only really came home to me in the past year. The sheer scale of the surface area over which production takes place is so immense that there is a far greater need to connect and remotely monitor all of these assets. The potential for optimisation to prevent losses is even greater there.’

The real innovation lies in the way Aloxy uses the sensors to measure valve position and is able to deliver and integrate that data with the client. Weyn is proudest of the engineering solution: ‘How we can run the algorithm in an energy-efficient manner. Many installations on site vibrate constantly, and you have to filter out that noise. You can filter out the vibration during measurement, but that consumes a great deal of energy and then you’re replacing the battery after a year. Our algorithm removes the noise whilst consuming as little energy as possible, and we can continuously improve it, because new types of valves keep coming onto the market.’

‘The chemical sector is highly innovative when it comes to its processes, but it is only in recent years that it has begun taking its first steps in implementing digitalisation.’

Fewer raw materials

Aloxy has done a great deal of co-creation with its clients, Weyn explains. ‘That’s important because the applications, and especially the cultures of the companies, vary enormously. What’s more, the use of industrial IoT is still relatively new to them.’ Some clients initially viewed it as an additional cost and had concerns about cybersecurity. The sector is accustomed to wired sensors. ‘That’s why we also develop a hybrid component in consultation with clients. In that case, only the configuration and battery data is transmitted externally.’

Aloxy wants to become the chemical industry’s partner for industrial IoT. Rather than relying exclusively on its own sensors, the company aims to be the party that provides integration of sensors into a cloud-based solution. There is still considerable value to be unlocked for the chemical industry. Weyn: ‘The sector is highly innovative when it comes to chemical processes, but we are only now seeing a genuine acceleration of innovation in terms of digitalisation.’

‘We also see this within the research carried out by UAntwerp and imec within the IDLab research group. We were recently able to demonstrate, using algorithms, that a pharmaceutical company could produce the same quality of product using significantly fewer raw materials, simply by better regulating flow and temperature. But in order to integrate those algorithms more widely, you naturally need sufficient sensor data.’