The company has joined forces with engineers at the AMRC’s Factory 2050, who are using AI to learn what machine utilisation looks like on the shop floor of the Sheffield-based SME, which specialises in large-scale engineering for challenging environments. A demonstrator will be created to highlight the accessibility of Industry 4.0 technologies to other SMEs in the region, and the potential positive impact on shop-floor productivity.
In actual fact, the demonstrator will be the first created under an emerging AI strategy being produced at Factory 2050, which seeks to harness the work being done with AI and machine learning techniques across the AMRC, and provide real use-cases for these techniques in industrial environments.
“Using edge-computing devices retrofitted to the company’s CNC machines, we have collected power consumption data during the production of automotive suspension components,” says Rikki Coles, AI project engineer for the AMRC’s Integrated Manufacturing Group at Factory 2050. “It isn’t a complicated parameter to measure on a CNC machine, but using AI and machine learning, we can actually do a lot with such simple data.”
Data from edge-computing devices at Tinsley Bridge was sent to the AMRC’s cloud computing services and, using the latest data science techniques, ran through an AI algorithm to provide new insights for the control and monitoring of manufacturing processes. Analysing the power signatures from the data, the algorithm looked for repeat patterns or anomalies, working out how many components were machined, and deducing that three different types of part were manufactured.
“The project demonstrates to industry that, with a low-cost device collating quite simple data, AI and machine learning can be used to create valuable insights from this data for the manufacturer,” says Coles.
Director of engineering at Tinsley Bridge, Russell Crow, adds: “Interrogating our machine utilisation rates means we have better visibility of what was being manufactured and when, and the ability to assess if we are scheduling effectively. This data will allow us to look at boosting our productivity on the shop floor. Rather than investing in significant cost and time for new, digitally integrated smart machining centres, we were able to work with the AMRC to retrofit our existing capabilities to achieve the same results and enhance what data we were collecting by fitting a simple current clamp to our machines; an unobtrusive solution that caused no disruption or downtime.”
Aiden Lockwood, chief enterprise architect for the AMRC Group, says the project demonstrator will show other SMEs how easily and cheaply Industry 4.0 technologies can be accessed: “Traditionally, these tools were built into commercial packages which could be out of reach for some SMEs, so there is a misunderstanding that Industry 4.0 manufacturing techniques are only for big players who handle incredibly complex data collected over a long period of time. However, AI is evolving and these techniques now give smaller businesses the ability to do so much more with their data. In this project we are using a simple data set, collected over a short period of time, to provide real benefits for the company.
“The formation of our AI strategy will allow us to lead the way in developing new capabilities and bringing the academic, technology and business communities of the region together to educate and demonstrate AI technologies for manufacturing industries,” he adds.
The next phase of the project will see the engineers at the AMRC train the system further, so the algorithm can detect non-conforming components while in production, or identify a problem when a machine requires intervention, such as when a tool is showing inconsistent wear and impacting component quality.
“Alongside the power consumption data, the plan is to feed the algorithm with available data about which of the manufactured components were non-conforming,” explains Coles. “So, as well as providing clarity around machine utilisation, the algorithm will essentially learn what a ‘good’ manufacturing process looks like and be able to actively monitor ongoing manufacturing processes.”
Crow adds: “We started out with the aim of finding out about retrofitting our machines for data capture, to give us an idea of machine utilisation so that we could optimise production schedules. However, as the project moves forward, we expect to gain even more visibility around our production processes so we can begin to move towards a predictive maintenance model in our machine shop that will provide improvements in quality and process capability.
“The insights provided by AI will allow us to identify when our machines will require intervention for tool changes, or how long we can run them without intervention,” he adds. “Predicting tool failure and prolonging tool life will also affect our right-first-time rates, reducing non-confirming parts and boosting productivity.
“In future, data analysis will permit us to run our machines more effectively, so we can free time for our engineers to work on value-added tasks, such as programming for multiple jobs or machines, creating a smarter factory that will help us manufacture technically advanced products.”