Big Data in Test and Measurement
Jon Aldred, Director of Product Management, Prenscia software at Hottinger Brüel & Kjær talkes about Big Data and the impact with the Industrial Internet of Things (IIoT).
Could you tell us what is meant by the term big data and how HBK works with this?
The term ‘big data’ has been used for a number of years, across different industries. Generally, it means combining very structured data and databases with unstructured data that exists within organizations; then trying to learn from this data. Our goal is to make better decisions based on a combination of these various different sources of data.
One example of where big data is beneficial is retail, where we use it to understand shopping preferences, such as why someone might choose one product over another, in order to develop improved marketing strategies.
HBK’s work is largely related to the engineering world, the data accumulated is most often used to drive engineering decisions. For HBK, big data’s key considerations are around gaining data from products as they’re being tested, or as they’re being used, with a view to improving product design.
Over the last few years, IoT (Internet of Things) technology has been at the forefront of new developments. How have advancements in IoT technology increased the availability of big data?
IoT, or Internet of Things includes a wide variety of devices with direct connectivity; for example a smartphone or a fridge that can be connected to the internet to provide data. Connectivity plays a big part here – more and more devices are connected to the internet, and these then become sources of information. We need to answer key questions around who gets that data, and what they do with it.
A key component of IoT is its ubiquitous connectivity. For example, any car can now be connected to the internet to provide data. One of the automotive companies that we work for told us that in 2019, their automotive OEM had 2.5 million connected public and private vehicles. All of these vehicles are generating data, and the company is collecting this data and using it to drive decisions.
IoT data tends to be comprised of small amounts of data, coming from lots of different sources. The other end of the spectrum – where HBK has typically been – is the provision of measurement technology; high precision, high quality data from one instrument, such as a single vehicle. What I see from an engineering perspective, is those two fields coming closer together – eventually streaming more, higher quality, higher precision data from lots of different sources.
From an engineering perspective, we’re looking at what can we gain and what we can learn from data coming from ordinary vehicles being driven around. A certain amount of that testing should be done upfront, so companies will implement specific tests and measurements of vehicles to confirm that everything is working as it should before they sell the vehicle. These types of tests tend to require higher precision, while involving testing several vehicles at once.
Could you tell us about some of the applications that are being used for big data analysis?
My work with HBK is largely in the areas of durability and reliability. I’m looking at how we design products with the right durability – so they don’t break – and also the right level of reliability – understanding how likely something is to break and under what conditions.
There are several key areas where big data is currently used. One, again, is product design, where big data analysis could be used for things like warranty analysis, which involves using data from warranty claims to better understand the reliability of products by looking at what is failing and when.
There’s also test and measurement, as I mentioned. We need to think about where we can gather information from and what we’ll measure with it; for example, if you’re thinking about designing a ground vehicle, you’re going to want to gather information from a fleet of test vehicles to understand how they’re being used and then adapt your design accordingly.
Another use of big data analysis is in what is often called ‘failure prediction’. This could involve operating an asset, for example, wind turbines. In the renewable energy sector, we’re looking to increase deployment of large wind farms, and these need to be cost-effectively maintained. A wind turbine is effectively a big fatigue test machine – at some point, it’s going to break because it’s constantly spinning.
Big data can help us to understand how a wind turbine is performing, and how likely it is that it’s going to fail in the next two months or in the next two weeks, for example. If we know this in advance, we can look at what we have to do in terms of maintenance to prevent that failure. The term ‘remaining useful life’ is often applied in these scenarios.
What are the benefits of using big data ecosystems compared to local machine storage data?
This is a question of scale. The kinds of problems that we’re looking to solve with big data are often too big to be solved an Excel spreadsheet on a local machine. Big data systems bring together data from different sources, and that requires robust IT infrastructure and computing power that has only become readily accessible in the last few years. For example, cloud computing is much more available now, and you can pay Amazon Web Services or Google, for on-demand computing to solve some of these problems. This isn’t always possible on local machines due to their limited capabilities.
What are the next steps for HBK in the field of big data?
We have an engineering solutions group within HBK, and we work on real world projects with companies in order to directly explore customers’ problems and develop solutions to address these.
For example, we are actively working with customers on their needs around asset performance, including companies operating large numbers of aircraft. Here, we are working with customers to understand how to maintain these more effectively while avoiding wasting cost on unnecessary maintenance. We work closely with customers to explore how they can learn holistically from what they’re doing, looking at patterns in what may not be working and how those issues can be addressed.
The next steps for HBK involve continuing to learn and work with customers on real projects, and continuing to develop our tools.
HBK has a good reputation for our measurement technology and our specific engineering expertise. As well as precision measurement, I think we’re going to increasingly see HBK being a strong force in the engineering software market.