Data storage solution delivers battery of benefits
28 April 2017
Keeping data secure, but ensuring it's still easy to access; maximising the huge potential of big data, but minimising the stress and burden of managing it - the ability to achieve apparently contradictory goals like these is a key challenge facing the research community.
For Dr Donal Finegan and colleagues at the UCL Electrochemical Innovation Lab, the RITS Research Data Storage (RDS) service provided the perfect means to succeed as they tackled the mystery of why some types of battery can explode or catch fire.
By delivering fast, effective, simple-to-use storage for terabytes of data produced by some of the world's most advanced X-ray imaging facilities, the service helped the team make fresh discoveries that finally revealed the deepest secrets of battery behaviour.
" "The RDS service was absolutely brilliant at listening to our specific needs and finding the best solutions" Dr Donal Finegan, UCL Electrochemical Innovation Lab
Tucked away in rural Oxfordshire, Diamond Light Source is a jewel in the UK's crown of cutting-edge science facilities. Accelerating electrons to near light speed, this synchrotron produces brilliant beams of light that can 'see' inside matter down to molecular level. Along with a synchrotron in Grenoble, France, it was Diamond that enabled electrochemistry specialists at UCL to answer a key question: why do lithium-ion batteries and similar energy storage devices sometimes fail spectacularly and (potentially) dangerously?
"Such malfunctions are rare," Donal Finegan explains. "But when they happen, the results can be serious in both safety and commercial terms. Just think of those media reports about exploding Galaxy mobile phones, or the battery fires that broke out on Dreamliner aircraft a few years ago. Our aim was to harness synchrotrons' incredible X-ray imaging capabilities to build an unprecedented description of events that occur inside a battery before and during failure, in 3D and real time. With such insights we aim to make future battery designs safer."
Right from the outset, it was clear that data management would be crucial to underpinning this pioneering EPSRC-funded work. Experiments harnessing synchrotrons' very high-speed imaging capabilities produce huge volumes of data very quickly. "The great level of detail needed to build a picture of battery failure involves a high rate of data generation, requiring large amounts of storage," says Donal. "Even a short set of experiments could produce 10 terabytes of data for us to take away and spend a year or two analysing."
Storing data on magnetic tapes or hard drives, however, simply wouldn't meet the need for:
- Certainty that vital data wouldn't be corrupted, lost or accidentally deleted.
- Capacity sufficient to hold the mountains of data that would be generated.
- Convenience in terms of ensuring that stored data could be accessed very quickly.
A superior, fit-for-purpose solution would clearly be essential to the project's viability - and the RDS service delivered it.
Success in Store
" "UCL was one of the first universities in the UK to recognise the value of de-risking the challenges involved in data storage from the researcher's perspective" Dr James A J Wilson, RITS
Designed to accommodate datasets of all shapes and sizes, the service enables UCL researchers to store and share data using a centrally provided, fully networked, fully managed facility. "UCL was one of the first universities in the UK to recognise data storage's importance to research and the value of de-risking the challenges involved from the researcher's perspective," says Dr James A J Wilson, Head of Research Data Services at RITS.
For Donal Finegan and his colleagues, the RDS service delivered three big benefits:
- Scale: The team discussed their requirements with RITS and agreed a storage allocation of 50 terabytes for the project. This eliminated any potential need to delete or compress data, or to hold any data on local PCs.
- Security: The service copied all data generated during the project to two separate secure UCL data centres, with full password protection for all stored data.
- Simplicity: For all users, interfacing with the service is quick and easy, with data and metadata up/downloadable via any computer anywhere in the world. UCL researchers registering online to use the service simply provide a list of those who will require access to the storage space allocated to their projects.
"Experimental time at synchrotrons is extremely valuable, so any loss of data would have been a massive setback for our work", says Donal Finegan. "The RDS service removed that risk. They were absolutely brilliant at listening to our specific needs and finding the best solutions. Setting up our own server system with similar storage capabilities would have cost tens of thousands of pounds, so the minimal cost of using the service was another major plus point."
For those engaged in the battery project, less time battling with data mountains and worrying about keeping them secure translated into more time and energy for pioneering research. The outcome was dramatic new insight into the process of 'thermal runaway' - the escalating chain reaction that can take hold when a lithium-ion battery degrades and fails - and specifically the interaction between battery designs and the observed failure mechanisms. This breakthrough will help everyone from manufacturers of satellites, aircraft and electric vehicles to designers of tablets, laptops and other personal computing devices avoid battery problems that can potentially cause billions of pounds' worth of brand damage.
"Contributing to this influential research project has been very satisfying for us as we move forward and keep adding to the capabilities we can provide, with long-term data archiving for completed projects also due to be in place by early 2018," James concludes. "It's an example that's really underlined the value we can add, especially as funding agencies increasingly make their awards contingent on good data management. The days are over when data storage could be dismissed as a marginal issue - it's now part and parcel of delivering successful research".