Data deluge: Allowing manufacturing to see the wood from the trees
Michaela Muruianu, from Digital Catapult’s innovation team outlines how data and technology are transforming our manufacturing industry.
The devil is in the data
Manufacturers today have their assets fitted with multiple sensors capturing production and performance data. Yet, in spite of this abundance of data, organisations are struggling more than ever to derive valuable insights that will enhance their performance.
The Pareto principle specifies that 80% of effects come from 20% of causes. If we were to translate this theory to the data captured by manufacturers, we could imply that 80% of insight is derived from 20% of data captured. The challenge is to identify and extract the data that will deliver this insight – the effort will pay off for those who succeed.
Playing a long game
Monitoring the performance of assets over time provides a strong indicator of the overall efficiency of factory processes. By better understanding the underlying influences on tool behaviour, we can better optimise performance across tool sets.
Challenging the theory with Seagate
Digital Catapult and Seagate recently hosted an Open Innovation Pit Stop looking into the role of data in manufacturing to enhance productivity. The first event delivered in Northern Ireland brought together SMEs, industry experts and researchers to consider what manufacturing will look like in 5-10 years, which technologies will have the greatest impact on performance and what to expect from our Fourth Industrial Revolution.
We identified a range of solutions that could help address the data challenges manufacturers face, including:
- The creation of a digital library of assets alongside models of asset behaviour can help distinguish between different shades of “good and bad” behaviour and establish thresholds of acceptable deviations in tool performance.
- Machine Learning (ML) applications such as convolutional neural networks can assist in the recognition of anomalies in imagery and recurrent neural networks can identify changes in a series of events.
- Transient models that observe patterns of change over a period of time, paired with various simulations can help uncover what sequence of events leads to failure.
- Effective data triage is essential in the process of distinguishing noises from signals.
Analysing insights derived from the digital library, manufacturers are able to understand what causes identical assets to perform differently. This enables them to look into how factors such as factory infrastructure, layout and process sequence can be optimised to maximise productivity. The switch from calendar-driven to predictive, event-driven maintenance, is another data-enabled, essential capability for the factory of the future.
It’s still about the human touch
As in previous Pit Stops, the importance of the human expert was at the centre of the discussion. Despite the increasing popularity of apocalyptic scenarios about automation leading to a jobless future, participants agreed almost unanimously regarding the central role human experts will have in factories of the future.
Highly skilled professionals and their observations about tool behaviour and factory processes will remain of great importance and can accelerate the quest of identifying the 20% of data that yields 80% of insights. These individuals will of course have to be upskilled and equipped with technologies that will enable them to enhance their proficiency. Virtual Reality (VR) and Augmented Reality (AR), in particular, present great potential for training (especially in hostile environments), data modelling and visualisation. These technologies can act as a bridge between technologies like Internet of Things (IoT) and ML, visualising the information they generate in a format that is easier to digest and act on.
Accelerating into the future
It is likely that the next decade will see considerable changes in manufacturing when technologies like IoT, VR and AR, Blockchain, 3D printing, ML and Artificial Intelligence will come to fruition and find their application in manufacturing from other industries.
Technological innovation means factories will become more agile, lean and tailored to the needs of the end customer; thereby creating opportunities for new or enhanced products and/or services. Depending however on the speed of technology development and adoption, the Fourth Industrial Revolution might, in fact, be more of an evolutionary process.
Follow @DigiCatapult if you are interested in learning more about and getting involved in the Industry 4.0 movement.
To find out more about Seagate Technology Plc visit: www.seagate.com