In the new frontier of engineering technology, collecting data is everything. Yet the path to discovery and insights is a challenging one, reports Susan Fearn Ringsell.
Whether it is Big Data or just data, companies are paying more attention to the gap between the collection of data and actionable insights.
‘We’re drowning in data, but starving for insights’, is a comment frequently heard by Michael Risse, vice president at industrial process analytics firm, Seeq.
In other words DRIP: data rich, insight poor – it appears to be too hard and takes too long to get insights from data to improve business outcomes.
Clearly, there is a disconnect between data collection and insights, with some of the most common concerns based around technical issues, costs, accessibility and security.
“First, there is an issue with data in the singular and data in the plural,” explains Risse.
The goal is to take the cost of the insight down, so that a next layer of insights will become profitable for the organisation to pursue
Michael Risse, vice president at Seeq
“It’s the silos of data at a system or even device level, and the sum of all of those things, which is hard to connect to for investigations, analysis, and insights.”
Meanwhile the proposed software architectures to bring the data and information together for the questions of today will be incomplete for questions that arise in the future.
Risse explains: “The goal is to take the cost of the insight down, so that a next layer of insights will become profitable for the organisation to pursue.”
Advances in both software and hardware mean that automation has become inherently intelligent.
Defined loosely as the computerisation of manufacturing, Industry 4.0 promises to bring the Industrial Internet of Things (IIoT) and Big Data into the realm of manufacturing.
Visually aware
The whole point of analytics and IIoT is improved production outcomes related to quality, throughput and safety.
“Logged data is collected from all devices on the process and linked with each product travelling along the product line,” says Barry Graham, Omron’s automation marketing manager.
“The entire process can be visualised from beginning to end without stopping.
“In order to compete on a global scale, manufacturers are seeking to boost productivity, improve quality, increase the flexibility of their production lines and maximise machine availability,” adds Graham.
Alan Hunt, product manager for ABB UK Measurement & Analytics, explains the key developments that have taken place in the process environment and what operators need to do to be able to maximise the value of data.
“Maximising plant assets and reducing unplanned plant shutdowns have increasingly become a focus for reducing costs and maximising productivity.
“Currently, potentially valuable information acquired by process instruments is often left stranded in the field.”
The developments in smart instrumentation and asset management systems are offering a raft of new opportunities for unlocking this data.
Providing real-time access to an expanded array of information, Hunt believes these developments can help to dramatically improve process performance.
“The goal is to take the cost of the insight down, so that a next layer of insights will become profitable for the organisation to pursue,” adds Hunt.
When insights are expensive to achieve, only large projects are taken on, but when insights can be cheaper to obtain, an increase in benefits may be achieved.”
Once the data can be accessed, cleansed and modelled, the most common approach to data aggregation lies in the hands of front-line engineers.
Big Data has existed for a long time in production lines, but the data on its own does not make sense until it is analysed and presented in a way that enables the people who support production sites to understand and use it
Barry Graham, automation marketing manager, Omron
“So just as there is an expectation gap linked to gathering data for insight,” explains Risse, “there is a software gap between modern applications for data analytics and the reality of engineers with spreadsheets.”
“Big Data has existed for a long time in production lines, but the data on its own does not make sense until it is analysed and presented in a way that enables the people who support production sites to understand and use it,” says Graham.
There is also the anticipated cost of what it actually takes to get insights delivered.
“We’re often asked how much will this cost and impact from an IT perspective, and certainly from a user perspective,” says Risse.
However, some modern approaches, machine learning and data platforms, require engineers and scientists with new skills, so there is a cost to proposed technologies.
Last but not least, security is often raised as an issue by companies and comes with a caveat, explains Risse.
“Production data is typically used locally by the group creating it, which means moving the data within the enterprise isn’t an issue for most companies.
“However, moving data to the Cloud is not a desired solution due to the security concerns.”
Data collection processes tend to differ between brownfield and greenfield plants.
At brownfield plants sites collection is largely done in-house. “There may be some outsourced data collection for a new sensor or distributed wireless networks,” says Risse.
There is a software gap between modern applications for data analytics and the reality of engineers with spreadsheets
Michael Risse, vice president at Seeq
But these are the exception on brownfield plants, just as they are the expectation for greenfield IIoT deployments.
In any plant there are as many collectors as there are data sources, but there are many more data sets – though often disconnected. “In particular, the smaller data sets, hard to access, and of unlike types, such as sampling rate, relational versus time series and so on,” says Risse.
Rapidly finding data insights is the main goal, therefore, if data can stay where it is, without being moved or copied, at low IT impact, and is made available for frontline employees with expertise, it has the ability to improve production results.
Efficiencies, insights, speed and competitive advantage are appealing in any sector. The point of analytics and IIoT is improved outcomes. “We have seen customers improve production outcomes related to quality, throughput, safety, and uptime,” says Risse.
As software companies continue to make data transport and storage bigger and faster, what does the future look like for data collection and usage?
“Our plan is to grow the number of connectors to existing and new databases and data sources,” says Risse.
Clear solutions
Meanwhile, as more data is collected in more systems, it is can be accessed by analytics software increasingly effectively in order to gain new insights.
Risse continues: “Any economic issues in 2017 are likely to have less impact than the ongoing impact of Moore’s law on computing: faster, cheaper, smaller, connected, smarter and pervasive.
“The future of manufacturing will see the evolution of a transparent, smart and interactive factory and supply chain based on three pillars: integrated, intelligent and interactive.
“Factory floor Big Data can be turned into high value information that can be used to increase overall equipment effectiveness,” adds Graham.
The volumes of data generated by the systems in a factory automation environment are enormous and will continue to grow – and with obvious benefits in terms of potential engagement and emphasising the key role of the engineer.
