Utility Week

UTILITY WEEK | MAY 2023 | 37 Pan-utility "There's no clear blueprint of what a fully synchronised intelligent energy system will look like in future so we're on a voyage of dis- covery," says Matt Webb, head of enterprise data management at UK Power Networks. "When you've got such a complex legacy system, and growing complexity in terms of all the people involved and the types of tech- nologies that need to communicate and be orchestrated, it's a real challenge." Discovery time Utilities are in the early stages of digital transformation, but now recognise the imperative to upgrade networks and systems to boost productivity and respond to press- ing customer, environmental and regulatory demands. Research by professional services com- pany Accenture found that during the pandemic utilities' adoption of certain tech- nologies was above the global average, with the majority of investments going into the Internet of Things (65%), streaming data (56%), edge computing (54%), bottom-up AI (54%), and trust-based architecture (53%). The escalating rollout of Internet of Things (IoT) sensors by the water industry, is pro- ducing data at a much higher rate and a finer granularity than previously possible, which alongside the imminent rollout of tech like smart metering, will trigger a massive influx of information. Businesses are grappling with the task of how to manage that torrent and sweat it to maximise the benefits. SES Water is running a proof of concept trial of a new data platform created to chan- nel selected data into a "data lake", where the firm can carry out analysis in combina- tion with data from other sources. Dilani Pararajasingam, head of data engineering at SES Water, tells Utility Week: "We're figuring out what the process is in terms of creating the data pipelines and refresh times, etc. The other part is the modelling work, looking at the relation- ships between datasets and what value we could add by combining them to gain more insights for specific needs and use cases." The flow of information and data in the energy sector has traditionally been uni- directional between a few large organisa- tions, including the system operator, and distribution network operators (DNOs). Looking ahead, the widespread expan- sion of local energy generation, needed to balance supply and demand and accelerate the path to net zero, places a heavy emphasis on collating and managing data from mul- tiple providers and sources including solar panels, heat pumps and electric vehicles. Webb comments: "The active network management piece is really key. Our main role in decarbonisation is to facilitate the connection of low-carbon technologies, we cannot be a blocker. As all those com- plexities come together we have to navigate through it." DNOs are already transitioning to a Dis- tribution System Operator model, which among other things will bring greater vis- ibility and control to the low-voltage end of the network and require companies to col- laborate to ensure interoperability across systems, processes and data. But the potential for disconnected data grows without appropriate standards and interoperability to communicate effectively. The water industry lags behind on data standards with none yet published. "Every- one's quite good at looking a™er their own data. It's when you want to combine data that these kinds of things need to be in play," says Pararajasingam. Energy companies are coming together through organisations like the Energy Net- works Association, in a drive to understand the common enablers, foundational com- ponents and standards needed to align on to support future investments in systems and data. According to Paul Linnane, director of data and development at Electralink, the operator of Britain's central energy data transfer function, an important goal is to achieve a balance between enabling tech- nology innovation and creating standards needed to build a resilient network. "Interoperability standards will enable organisations to invest with certainty when building interoperating systems to support the network," he tells Utility Week. "How- ever, producing standards too early can add significant delay and cost to innovation pro- jects without any benefit. Balancing these elements to enable innovation in the early stages, but providing certainty in the later stages of the process, is key." Electralink's Paul Linnane will speak about whole system challenges for building out digital infrastructure at utility Week Live on 17 May. Data best practice guidance, written principally by the Energy Systems Catapult, and now being taken forward by Ofgem, is expected to provide more direction on areas such as data licensing, metadata, standards, and data modelling. The regulator's rules are built into licence conditions for the cur- rent price control period, ED-2, which runs to March 2028. Ensuring data is managed appropriately and securely on networks is also contingent on gaining the requisite funding for digi- tal infrastructure, says Linnane, who also underlines the importance of machine learn- ing and ElectraLink's Project FAIR, which helps collate data from many data providers and sources into a single dataset. Clean start The emerging fields of data science and AI have shown huge potential to generate insights but proofs of concept have o™en faced issues scaling up due to the high degree of manual data cleansing required before algorithms can be set to work. "O™en you have highly skilled SMEs, such as process modellers, employed to do analy- sis, but they can spend up to 80% of their time collecting data and data wrangling and cleansing before they even get to start on the value-add of modelling," explains Jethro Yates, head of data science at Thames Water. According to Yates, a smarter approach means separating out roles and responsibili- ties and employing teams of dedicated data scientists and platform engineers to carry out data acquisition and cleansing ready for SMEs to pick up the baton and focus on what they do best. Thames Water's Jethro Yates will speak about smart water networks at Utility Week Live on 17 May. "Shiny, exciting models and machine learning represent the tip of the pyramid. There's a whole base of work that needs to happen before they can function effectively," says Yates. The data cleansing challenge highlights the additional skills and third party exper- tise needed by energy and water compa- nies as they venture deeper into digital transformation. A recent report by the Energy Systems Cat- apult found energy companies are struggling to recruit people with the skills to match a data-led future. Some 40% of companies surveyed described difficulties hiring data scientists with the requisite skills. The report points out that sectors such as FinTech and social media have been more proactive in recruiting the most skilled candidates. "Other sectors have been ahead of the water sector on recruitment. We're learning we need a substantial amount of new skills across the industry and new types of part- nerships with third parties to bring in exper- tise," says Yates. Becoming "data-ready" also demands culture change and upskilling employees 1617 MAY, NEC, BIRMINGHAM See this content brought to life at Utility Week Live. Register free at: utilityweeklive.co.uk continued overleaf

• Cover