Artificial Intelligence (AI) has shown promising results in connected and already digital industries such as media and advertising, finance and retail. AI is entering many new fields at a high pace, and with its high impact on the environment, the manufacturing industry is one of the essential fields for the application of AI. Industrial AI, often termed as Industry 4.0, can increase production efficiency while bringing down manufacturing scrap and advancing towards a much cleaner environment.
According to McKinsey1, AI-enabled work in Germany's industrial sector could raise productivity by 0.8% to 1.4% annually and reduce overall maintenance costs by up to 10%. These productivity improvement goals require not only training AI/ML models on historical datasets but also taking corrective action items to improve the processes in real-time, which is referred to as "Intelligence on the Edge". However, in today's scenario, most of the manufacturing companies still don't have large bandwidth and high-speed connections, which are essential for making "Edge AI" the most relevant technology of the future..
Here are a few advantages of employing AI in the steel and iron industry
What is Edge AI:
"AI on the Edge" or "Edge AI" refers to the ability of machines to sense and react to their dynamic environment. With vast amounts of data being generated from hundreds of sensors on each machine, it is challenging to transfer the enormous volume of high-frequency data over the cloud. The setup is costly, has chances of data loss in the high-frequency data, and does not guarantee stability. Gartner2 states that today, around 90% of the data generated is processed in the cloud or traditional data centres, but this figure will reduce to only 25% by 2025, making edge AI the most exciting technology of the future.
Why Edge AI for Industrial Applications:
Limit Cloud and Connectivity costs
Time Critical Applications
Tvarit GmbH is a Germany based company that drives digital transformation by developing AI-based business models. Tvarit has been repeatedly awarded as the most innovative startup in Germany (TiE Germany 2018) and as one of the 8 best companies among 490 AI companies in Europe (EDI EU-Horizon 2020).
Tvarit has successfully implemented 32 industrial AI projects, ranging from scrap rate reduction, minimising the number of defective parts in the aluminium production, detecting tool wear and quality in stamping dies, monitoring the machine condition, and recommending optimal process parameters for the machine operator to improve the production quality in injection moulding.
How Tvarit adds value to Manufacturers with Beckhoff's TwinCAT 3:
From the Internet of Things to the Internet of Services, Beckhoff is driving the key technologies that underline the convergence of Internet, Information Technology (IT), and automation technologies. McKinsey1 states that players in the industrial sector should develop core analytical capabilities internally, but also leverage third-party resources for AI-based performance improvements. Tvarit GmbH joins Beckhoff in providing tailor-made AI services, using Beckhoff's TwinCAT 3 automation technology.
TwinCAT 3 Machine Learning Libraries by Beckhoff:
With TwinCAT 3 as the new, state-of-the-art software platform, Beckhoff integrates PC-based control technology with the Internet of Things (IoT) and data analytics, laying an essential foundation for Industry 4.0 and opening up a wealth of opportunities in the manufacturing industries.
TwinCAT 3 enables seamless integration of Machine Learning (ML) solutions into PC-based control applications, that run in real-time to handle highly demanding tasks like motion control. These capabilities provide product and equipment manufacturers with tailored Industry 4.0 and IoT applications, resulting in numerous advantages in terms of costs, process efficiency and quality. The cutting-edge industry-ready technology makes Beckhoff a market leader among many players in the industrial automation domain.
Industrial Motor Application:
Tvarit has also worked on the AI-based Machine Tool Health use-case with the TwinCAT 3 software. The bulky motor drives used in the printing press have busy operating schedules everyday to print millions of newspapers. A motor drive failure will cause a massive loss to the revenue of the printing press. A usual procedure during machine breakdowns is to change the motor drive losing atleast 10-15 minutes of printing time. A big challenge is to monitor the motor drive’s health and to predict the time of failure in advance avoiding unexpected failures and downtime.
The motor has close to 80 sensors connected to the TwinCAT 3’s automation platform, which log the measurements every 10 milliseconds using Beckhoff's TC3 Data Agent and TC3 Analytics Logger. The readings such as drive current, voltage, temperature, torque, etc. collected are transmitted to the storage centers via the Beckhoff's EtherCAT, for further analysis and modelling with Tvarit's Automatic Predictive Analytics (APA) platform.