Developing an automated industry IoT solution for predictive maintenance

The University of Sheffield and engineering manufacturer Tinsley Bridge have been pioneering new approaches to industrial Internet of Things (IoT) applications through the innovative Pitch-In project.

The suite of technologies and capabilities that come under the umbrella of the Internet of Things (IoT) have enormous potential in manufacturing. Perhaps most transformative is the ability to implement predictive maintenance of machine tools and other equipment. In this scenario, a suite of sensors can continuously monitor and report on the working condition of equipment, flagging issues before they arise. This has the potential to prevent costly unscheduled downtime.

Both academia and industry are interested in technical solutions – but there remains a sizeable gap between theory and practice, as Professor Ash Tiwari, RAEng/Airbus Research Chair in Digital Manufacturing at the University of Sheffield, explains as the academic lead for this project.

“There are a number of research studies on the development and application of state-of-the-art IoT technologies for manufacturing processes. However, there is very limited research on how these technologies could be adapted for existing/legacy processes and what resources are required to enable this in industry.”

Dr Boyang Song, Research Associate in Industrial Internet of Things (IIoT) and Machine Learning at the University of Sheffield, explains as the researcher on this project.

“There are plenty of laboratory studies and prototyping happening; but actually very few end-to-end solutions with the latest state-of-the-art technology being tested in practice, on the factory floor. If the research can get to shop floor, then we can understand problems and capture use cases in a better way. Furthermore, for industry, it can be seen as less risky to host research collaborations within their facilities.”

Creating academia-industry synergies

The University of Sheffield has a long-standing relationship with engineering manufacturer Tinsley Bridge, which makes key components for several sectors including automotive, nuclear, rail, defence and energy. Tinsley Bridge had previously investigated some limited IoT applications; but the opportunity arose to take this much further with an ambitious new project with the University of Sheffield, funded and facilitated by Pitch-In.

This involved Tinsley’s three-story paint plant, which runs almost continuously. Here product parts go through a number of different stages – including treatment with alkaline, acidic, and negatively charged mixtures, as well as numerous rinsing stages with mineralised water and, of course, paint and finally a furnace.

Traditionally, operations would be monitored by a technician going round three times a day manually inspecting the equipment with a clipboard, taking notes and readings by hand, then inputting that digitally onto the system. Then there is a subsequent process to chart and analyse the data.

The IoT system being implemented consists of a suite of sensors measuring things like chemistry (including pH), pressures, temperatures and the power consumption of various motors. These digital readings then go wirelessly to a database where some simple processing can generate alerts, for example when pressure drops below a certain threshold.

“What we’re aiming for is to automate the whole process,” says Alex Kelly, Technical IT Manager at Tinsley Bridge. “If we can collect the data more frequently we can understand the machines better. We can take readings thousands of times a second, and often it’s in those little bits of hidden data where you’ve got the really useful information that can tell you a lot about how the machines are working. It can show up faults before you even know about them.”

Exploring advanced technologies

The University and Tinsley Bridge team have also been investigating more advanced technologies, such as artificial intelligence and machine learning techniques to analyse gathered data and look for emergent patterns. These may be able to give even more insight into the behaviour and workings of the machines.

In terms of advanced hardware, they have trialled new networking and communications technologies such as LoRaWAN, which can transmit data from sensors over very long distances without the same level of interference that can plague wifi.

The team has also investigated the use of a cloud-based solution that automates data capture, analysis and maintenance alerts – all entirely remotely. This also made the collaborative research process easier, as Alex and Boyang were able to work on code together, using a GitHub repository (an online platform for software development and version control). Boyang also created a simulation of the entire factory IoT system within his own home using Raspberry Pi devices.

Adapting to unprecedented times

The Covid-19 outbreak led to the necessity to utilise new ways of working to respond to unprecedented times, with the necessity to work remotely more important than ever. This included the creation of a GitHub repository, which allowed data and code sharing.

Time in the factory was spent installing sensors, inputting code and collecting data, with Boyang setting up a simulation to simulate the factory IoT system in his own home using Raspberry PI’s. Alex explains, “It [Covid] was a major opportunity and a major challenge; before we used to have regular meetings with Boyang, we used to come and physically inspect sensors installed, but after Covid, we changed our way of working.”

Mutually beneficial outcomes

For Boyang the project has already proved successful in terms of technology experience in practice, on the factory floor and he is currently writing a journal paper on the outcomes to share with the academic and industry community.

“When we came in we had the freedom to really start from scratch and implement the most advanced IoT system and cloud platform.

“We can now share our experience with the industrial community, so they can have the confidence to go forward with digital transformation and they know how to do it. Before this project, it was mainly research in the lab, certainly not a full demonstration on the factory floor.”

For Tinsley Bridge the project has already achieved some demonstrable efficiencies as well as giving them the foundations to assess and adopt new technologies as they arise.

Specifically, by implementing a system of automated alerts, staff no longer need to manually walk around the factory floor three times a day taking readings on a clip board. This helps at weekends when fewer people are on site and was beneficial during the coronavirus pandemic, when it was necessary to limit numbers.

The use of LoRaWAN sensors and technology has also allowed Alex to quickly and easily perform ad hoc analysis, working directly with individuals or departments in the company without the need for additional support in terms of installation.

“We’re really starting to build a business case”, he says. “The cost of these systems has come down incredibly, and there’s very little back end work. Five or ten years ago this was impossible; there was so much manual intervention to get all this data out. Tinsley is building up IoT capability, working with our key collaborators and partners like the University of Sheffield.”