Ireland-based materials manufacturer, Kingspan has recently deployed AWS accelerator solution, AMP, from Storm Reply to help the company maximise its data. We caught up with John Neary, Automation Manager of the company's Insulated Panels Division, to find out more.
Founded in 1965, Kingspan’s building materials business has grown into a €8.3bn revenue company with over 200 manufacturing sites worldwide. In 2019 the company also opened its IKON Global Innovation Hub, dedicated to finding innovative solutions for sustainability, performance and efficiency in construction.
Following a previous spell with the company, John got a call to return to Kingspan in 2019, rejoining in 2020 just prior to the outbreak of the COVID-19 pandemic. Despite the obvious upheaval, John explained that the lull around projects and deadlines caused by the pandemic, allowed the company to dedicate time on other areas of the business.
“We still implemented projects globally, but they were far less in number,” he commented. “However, this helped us with the standardisation of operations across the business, something we’re seeing the fruits of now. As a company, we want to develop not only our products, but also the environmental element of the business, and we have a large, growing team that are enabling us to do that.
“We want to be able to manufacture and deliver equipment built and designed by us. For example, our chemical systems, from a product perspective, are all built in-house to retain IP.”
John explained that as the business has grown, so data has become increasingly integral to Kingspan’s strategy. In the past, the limited amount of data generated by the company was tied to statistical process control (SPC) systems, which helped the company harvest and store the data on-premises, but little else.
He added: “Our data was quite scattered in the sense that the plant in Ireland would be completely different, from a data perspective, to one in England. Again, with the lull caused by COVID those systems subsequently became obsolete and we ended up with a number of old operating systems running in the background. For example, Flash became a security risk.”
It was then that Kingspan had the opportunity to invest with Storm Reply on using SPC obsolescence as a tool and pipeline to start harvesting data up to the cloud. The company is currently working with Storm Reply on building out that offering. In addition, Kingspan also engages directly with Siemens and Mendix from an Amazon perspective.
From an infrastructure point of view, the company has removed on-premise servers and the requirement for their maintenance. This has been done by the deployment of a simple edge device connected to a PLC, and this can also be supported remotely through Amazon Web Services.
“By putting data into the cloud we can standardise our whole system,” continued John. “So, a transducer in Kingscourt, Ireland, will be the same as a transducer in Pennsylvania. This means we can start comparing the two. Beforehand that would have been extremely difficult and would have involved trying to use Excel spreadsheets downloaded from a database. The fact that it’s now in the cloud means our people can run a report on both transducers and compare them from two different plants; everything is a lot more interconnected and streamlined.”
Legacy historian, SCADA and statistical process control systems hold companies back from making data driven decisions, and while Kingspan has begun the journey to the cloud, it’s an issue the company is wrestling with.
The company still has legacy systems in place and is currently working through 22 ‘brownfield’ sites that have to be removed. These involve normal legacy IT issues where data is still being generated by those systems day-to-day, but running reports against that data is proving difficult.
“It’s only when we gather more data that we’ll really start to see the benefits,” John added. “And we’re still very much in the infancy stage of this journey right now. We’re starting to collect additional data from condition monitoring sensors etc that we didn’t have with SPC systems, but we’re also still running legacy servers to capture some of that data in some cases.”
With regards to data collection, Kingspan had many locations with a physical server on-site. That data had to be held for a certain amount of time which was proving a challenge in itself. The servers had to be kept stable and patched, and server downtime was taking up a lot of engineering resource.
“That resource has now been freed,” added John. “And we’ve also reduced our capital expenditure costs for those systems, because we’re not having to buy them, (the servers, licencing etc). It’s now more of an OPEX cost in terms of the data that we’re harvesting, and if the line goes down or we shut a plant. It means we haven’t made an investment in a service which has then become redundant.
“Cloud has enabled me to see into my own team. I had two people in particular dealing with SPC; one has now moved into an automation lead role while the other is working with Storm Reply and AWS. I’ve theoretically reduced headcount in that area by moving onto the cloud-based system. We’re actually recruiting more people around the cloud piece so that we can bring the Amazon system to the next level ourselves. From a maintenance point of view, it’s taken a huge headache away from us and freed up resources.”
John explained that in terms of exactly what machine data is stored in the cloud, it is machine and company dependent. Currently, Kingspan’s flow, pressure and temperature data is stored on AWS cloud but there are set points from a machine perspective, that are on a HMI-based system for setups etc.
The company’s ultimate goal is to use its Amazon systems and AI to help predict failures and quality issues etc. Although it is a number of years away from that scenario, Kingspan is already working with Storm Reply on a proof of concept (POC) and Amazon on a profile system to make sure the company is heading in the right detection.
“At Kingspan, cloud, data and seeing what it can do is new for us,” said John. “We’re very much a business that likes proof before we invest heavily in something. With those smaller investments we can start to answer those key questions around whether we implement predictive maintenance in the cloud; can we use the data from our machines to predict failures of pumps, motors, etc?
“Without Storm Reply we didn’t have the expertise or the know-how to get these projects started; we had zero experience. Without them, we wouldn’t be where we are now. They’re going to work on the next POC with us to find a solution. And it’s going to be based around the same developer master system that can then be rolled out to all of our plants without heavy investment thereafter.
“We want to start using our data in a meaningful way and want to make sure that if a piece of data goes out of spec, then we know what impact it has on our product. We’re working with Storm Reply on implementing a quality system in the cloud.
“From an engineering perspective, we would love the data to be able to tell us a product or piece of equipment is perfect, or when it needs a maintenance check. We also want it to develop into an MES system, using the data we’re collecting, and link it to SAP and NetSuite. That will mean, from a production line perspective, if a global engineer installs a production line in any area, we ensure we’re offering the full package from an installation, engineering and operations point of view for the whole line. And that’s where the cloud and Storm Reply will come in.”
For more information, click here.