Issue link: https://fhpublishing.uberflip.com/i/1173186
Case study: Machine learning specialists In 2014 Kelvatek's parent company Camlin Group acquired a seven-year-old Italian company that specialised in using artificial intelligence to add value to commercial businesses. Henesis srl had previously counted automotive giant Toyota among its customers and was listed among the 12 most innovative Italian companies by US publication MIT Technology Review. Now fully integrated as Camlin Italy, this acquired unit has more than doubled its headcount to 30 and formed a specialist Machine Learning Group within its Parma premises. "We have software engineers, robotics experts, mathematicians, physicians – many of us with PhDs, some with professional experience," says Luca Mussi, senior research data scientist at Camlin Italy. "With so many different backgrounds and experiences, there are always different approaches available to any problem encountered." A significant focus of the Machine Learning Group's work is to take data from UK energy assets and formulate models to help the network operators in their day jobs. "We analyse very high frequency data,' says Mussi. "Almost real-time voltage information, wave forms and sensor signals." This isn't easily done by the human eye, which is where the experts use intelligent modelling to allow powerful machines to crunch the numbers and find the story hidden in the wealth of input information. In one project, the group is looking to improve Kelvatek's fault-location services by an advanced process of elimination – if engineers know exactly what type of asset is not working correctly, they can narrow down their physical search dramatically. So machines are trying to fathom which nuggets of data can identify malfunctioning asset types. "The data scientist has no black box with the answers," says Mussi. "We look at the data lifecycle and measure results." Mussi enjoys being part of the Camlin family, saying the sector expertise in the broader company aids the data science. Alongside the Machine Learning Group, Camlin Italy consists of a machine vision team, an electronic design unit and a biomedical research outfit. "We are not left alone to find the answers, we have experts who we can ask about transformers and LV networks and faults, for example. There are 400 people in the worldwide Camlin Group, many of them engineers, and they help us out a lot. "We're in a really unique position because we have access to a large historical data set on LV cables and other assets. Another advantage of being part of the Camlin group is the ability to influence the data that is collected in the first place. "We can try to change the type of data that is brought in, or the frequency or quality of that data,' says Mussi. The group aims to use artificial intelligence and machine learning to give predictive information to network operators about forthcoming asset failures. And beyond this critical insight, the arrangement can have unplanned benefits for businesses. "If you work as a contractor on a specific and limited project you may be successful and deliver good results,' says Mussi. "If instead your job is to dig into company data and work together with the business to research new opportunities and insights, there's much higher probability that something unexpected could be found." Ultimately Mussi sees a future where artificial intelligence and machine learning help create a fleet of intelligent, self- coordinating and autonomous switching devices working away to optimise network configuration and maximise operation efficiency. "We're seeking potentially disruptive findings and it looks like we have all necessary ingredients to succeed," he says. have relied on a dumb system rather than a smart system." This set-up won't be able to cope for much longer, however. "As we move into a different world there will be a lot more moving parts, so data on how they all work will become more and more crucial," says Sandys. "Data will be the new value in the electricity system." The taskforce, managed by the Energy Systems Catapult, identified gaps in the quality and visibility of data available in the sector as well as in the skills available to make the most of what information did exist. "We came up with two prin - ciples," says Sandys. "Filling the gaps and presuming open data." These underlying fundamen- tals led to five recommendations being set out in the taskforce's report earlier this year. Summa- rised, these were: digitalisation of the energy sector; maximised value of data; increased visibil- ity of data; increased visibility of infrastructure and assets; and co-ordination of asset registra- tion. Sandys says companies in the sector reacted more positively than expected to the report. "There is now acceptance that data management should be part of business as usual," she says. "It's about how rather than why." And artificial intelligence is part of this 'how', she explains. "Once we start to get real time data, distributing as - sets and delivering value, this will not be coordinated by a command and control office in Wokingham. It will need to be algorithmically managed, pre - dicted and assessed. Machine learning is really important." Improving efficiency Randolph Brazier, head of in- novation at the Energy Networks Association (ENA), agrees that computerised analysis of infor- mation can play a big role in improving the efficiency of the energy system. "Artificial intelligence can help you understand why Some members of the AI/data science team at Camlin: Federico Sassi, Roberto Ugolotti and Luca Mussi. In association with: NETWORK / 27 / OCTOBER 2019