PC-SEP24-PG16.1_Layout 1 17/09/2024 12:38 Page 16
INDUSTRY 4.0/IIOT DATA STREAMING OFFERS STRATEGIC INSIGHT
As new and unforeseen applications for AI and ML emerge across the manufacturing sector, real-time data will be the foundation upon which they develop
Richard Jones, VP EMEA, Confluent, explains the value of data in factory-wide automation, and how data streaming offers immediate access to robust, reliable, and well-protected data
ecent years have seen Industry 4.0 move from buzzword to reality. From augmented reality to factory-wide automation, modern factories are a showroom for technologies that were once a pipedream Successful manufacturing is unarguably data driven; the better an understanding we have of a factory’s processes, the better placed we are to improve them. The adoption of these futuristic systems is made possible by greater access to, and control over, data, and the understanding that comes with it. Of course, that data needs to be accurate — which means the older that data is, the less value it can offer. Ensuring that data is accessible as soon as possible after it’s collected is important if manufacturers are to understand what’s happening at their company in the moment.
R
This means that many companies are looking to stream data, rather than collect it and process it afterwards. Being able to access data when it’s in motion, enables the adoption of the transformative technologies. In fact, recent research has found that almost two thirds (63%) of IT decision-makers in manufacturing agree data streaming has eased the adoption of artificial intelligence (AI), or machine learning (ML).
But how does this all take place? What does the data tell us about manufacturing operations — and how exactly does data streaming change that for the better? Smart factories are a hotbed of sensors and IoT devices, with almost every piece of equipment capable of ‘talking’ to one another — transferring data between both one another and to key decision-making systems.
16 SEPTEMBER 2024 | PROCESS & CONTROL
This gives manufacturers unprecedented strategic insight into the strengths and weaknesses of their infrastructure. Take predictive maintenance as one example. Whereas machines would have once run until they broke and then been taken out of commission for a time, AI can detect tiny deviations from the norm — even down to the sounds that they make.
These discrepancies can indicate component wear or impending failure, and the IoT network connecting that machine to the control room can flag that maintenance should be conducted ahead of time. Incremental improvements like this can massively impact overall performance. The International Society of Automation (ISA) has reported that around 5% of plant production is lost per year due to machine downtime, equivalent to just under $650 billion globally per year. It’s no surprise then, that 60% of original equipment manufacturers (OEMs) will deploy AI and real-time data to pre-empt equipment failure by 2025, according to industry analysts IDC.
We can see similar leaps forward in the robots deployed across the factory floor. Robots can interact far more organically with human employees, potentially responding to gestures, facial recognition, or voice commands.
Data doesn’t stop at the factory door, either. Just as we create a digital copy of machines to understand them better (a ‘digital twin’) the modern supply chain can be seen almost in its entirety through data. Potential bottlenecks, inventory queries, delivery times, and more can all be tracked from a single point of control,
keeping manufacturers in control. Access to those control systems from different departments is the foundation of more effective collaboration between internal teams, as well as with customers and external partners. Everyone can work from the same frame of reference, pulling towards a common goal — with changes in real time. If data is that important, we must be able to verify the source, quality, and overall reliability of it as it enters the factory ecosystem. If data is outdated, poorly formatted, not cleaned correctly, or not properly protected, it can mislead us.
Modern manufacturers do not have the time to spend hours or even days sifting through data to make sure it’s of the required quality. Legacy methods like batch processing would demand compromise here, as they simply can’t keep up with the flow of data. Real-time data streaming does not need to make such concessions. A real-time data streaming platform can integrate hundreds of data sources, creating a cohesive network of pipelines by which data can travel. Data doesn’t have to stand still; it can be prepared and analysed even on the way to its intended destination.
As a result, research from Confluent shows that three in five (59%) of manufacturing IT leaders see data streaming platforms as “important” to achieving their data- or information-related goals – 35% actually see it as “critical.” For those that have already committed to such a platform, one third (33%) have seen a fivefold return on investments into data streaming data, while almost half (46%) have seen their investments doubled. It’s worth exploring an example here from BMW Group., who captures an astonishing volume of data from 31 plants across 15 countries – as wide-reaching as overall performance, and as granular as the position of components on the assembly line. This data is made available via BMW’s Integration Platforms organisation. That platform required a foundation that was both scalable and reliable. To address this, BMW has adopted a “data streaming platform” (DSP) that allows its teams to focus on the insights and opportunities that data can offer. Thanks to the access to real-time data, BMW can order components with more foresight and context, better monitor supply and demand, take better care of its equipment, and deliver a higher quality product to the customer.
Confluent
www.confluent.io
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48
