Intelligent Integrations & ML for Industry 4.0

It's no big secret that data’s value is growing exponentially, and manufacturing is one industry that has long recognised and benefitted from data’s potential. Despite its abundance, accessing data in the right way to harness its value is sometimes a challenge. And it’s this challenge that tubular product manufacturer voestalpine Rotec turned to pioneer in intelligent integration specialist Crosser to unite its facilities’ data.

voestalpine Rotec is an international group of companies that produces tubular products from precision steel, lightweight aluminium and high strength titanium. It's particularly specialised in passive passenger safety solutions for the automotive industry and other industrial machine applications. Headquartered in Krieglach, Austria, it employs almost 1,300 people in Austria and its eight additional production sites located across the world.

Producing components that play such a crucial role in passenger safety and accurate machine performance, voestalpine Rotec places the utmost importance on ensuring the technical accuracy and superior performance of its output. As we pass the tenth year since the birth of the Fourth Industrial Revolution, and with the ongoing development of new digital technologies like machine learning, tapping into machine data is key to voestalpine Rotec’s future success. But how should it be done?

Right data, wrong place

Like many other manufacturing facilities, voestalpine Rotec produces a vast volume of data. The problem is not producing it, but accessing it in a valuable way. Some of the machines on its factory floors are cutting edge, while others have been operational for many generations, which has resulted in data collecting in siloes. Data silos create a challenging IT environment, relying on human knowledge and time-consuming processes to determine where data is located.

In the past, voestalpine Rotec implemented several vendor solutions in unison to meet the majority of its data standardisation needs. If the use case required didn't exist within the standard system offered by its multivendor environment, a small in-house coding development team would create a custom programme internally. However, this approach has a high total cost of ownership, high responsibility and it's not possible to scale across other machines or locations. Surely a smoother process existed?

With limited resources and just a small in-house coding team, voestalpine Rotec wanted a vendor-independent universal platform to connect its IT systems, access various data sources and harness the value of IT. For this reason, voestalpine Rotec turned to Crosser.

Connecting with Crosser

voestalpine Rotec’s struggles with limited resources aren’t uncommon in manufacturing. The IT skills gap has plagued industry for years — according to Skillsoft’s 2021 Global Knowledge IT Skills and Salary Report, the global unmet demand for talent has increased by 145 per cent since 2016, due to the unmatchable pace of technological development and a common struggle to recruit personnel with the right skillset.

With limited resources, voestalpine Rotec wanted to predominantly rely on standard capabilities to enable machine to machine (M2M) communication, but also had some use cases where it needed to add specific functionality through individual programming. The dual nature of Crosser’s platform fulfilled both of these requirements.

“The Crosser platform is unique in that it is designed using a low-code modular structure,” explained Kai Schwab, regional sales director for the DACH region at Crosser. “By placing prebuilt modules into a sequence, individuals from any business area can create a process without prior coding knowledge, taking pressure off small, under capacity IT teams.

“The unique thing about Crosser is that if a desired module does not exist to satisfy the use case a customer is looking for, we have the flexibility to add new custom modules specifically to provide the function. Crosser will build a specific module for any customer requirements, and it's simple to do so thanks to the modular architecture of the platform.”

The best of both worlds

In summer 2020, the initial implementation phase of Crosser’s platform across voestalpine Rotec’s facility began. At first the main use case was to gather, standardise and gain actionable insights from raw data from several siloed sources across various IT systems.

Implementing this use case required custom coding. Some of the IT infrastructure was generations old, made up of old legacy systems with no official interface available to access anymore. voestalpine Rotec worked closely with Crosser to design and integrate very specific modules to gain access to this data.

“Taking the dual approach of standard and custom capabilities was the only way to gain access to the data and achieve a fully comprehensive overview of our interfaces and machines,” explained Markus Begutter, IT manager global at voestalpine Rotec. “Once we gained full visibility of our IT system in a centralised location, it was possible to deploy further sub-use cases based on the data insights accessed via the Crosser Platform.”

Some of the use cases that voestalpine Rotec has deployed so far include the automatic communication of stoppage times to its enterprise resource planning (ERP) system any time there’s a problem in production. The Crosser platform has also enabled vision quality control by sending inline measurements and process data from a robotic vision system to the database.

In another branch of the Rotec group, a safety use case has been deployed via Crosser where automatic notifications are sent via SMS or email in the event of an emergency. Based on data collected and analysed through Crosser, a programmable logic controller (PLC) system triggers an alarm in the case of a fire or other emergency and relevant personnel are automatically notified to maximise facility safety.

An added layer

In addition to gathering and normalising raw data for real time actionable insights, voestalpine Rotec is also using the Crosser platform’s machine learning (ML) capabilities. voestalpine Rotec has built a ML model using Python script to collect information from machines to enable predictive quality control of its output.

“The ML algorithm assists our facilities in setting the right parameters for production. It gathers data on measurements taken from parts during production and suggests recommended actions to predict machine deviations and faults,” explained Begutter.

“The ML model is run through the Crosser platform. This ensures we deliver the right quality and granularity of data to ensure the model always works correctly,” added Schwab.

The ML model use case is ongoing with constant improvements being made. For the ML model and any other use case in development, Crosser holds a regular status meeting with voestalpine Rotec to determine the exact challenges they’re currently facing.

“Crosser’s customer-centric ideology stood out to me from the start,” expressed Begutter. “If we’re facing any challenge when deploying a new use case, the Crosser team always reacts quickly. They are proactive in investigating issues and make the time to deeply understand their customer’s needs. Having this regular touchpoint is important to us, and shows how Crosser provides ongoing support even once the initial implementation of their solution is complete.”

Long-term results

Now that voestalpine Rotec has worked with Crosser for almost two years, the company has achieved a good overview of its IT interfaces in a centralised location for all data to be accessed. By connecting every machine through one platform and deploying predictive ML algorithms, the company has an elevated overview of what and where operations are taking place.

“The key benefits of working with Crosser are the extensibility and scalability of its solution. It's possible to extend a use case from one machine to multiple versions of the same machine for a fast, efficient roll out. It's scalable because there is an infinite world of use cases that can be continually developed and deployed as we consider new uses of the data we collect,” explained Begutter.

“Using our data in this way helps us to improve product quality, achieve operational excellence and reduce waste production while increasing output and reducing machine downtime. There’s plenty of scope to roll out existing use cases through the Crosser platform further across the company in the next 24 months.”

If data continues to explode in such a rapid way, according to Raconteur, the world will produce 463 exabytes daily by 2025. To put that into perspective, in 2020, daily data production was just 2.5 exabytes. By implementing a flexible and scalable solution, voestalpine Rotec has futureproofed its data strategy to handle the past present and future of manufacturing — and whatever technology it may bring.