The growing dependency on electricity networks due to the electrification of heat and transport has highlighted the importance of network visibility. The collection of the right data is key to paving the way for this and achieving the net-zero targets set out by HM Government. Distribution Network Operators (DNOs), Transmission System Operators (TSO’s), and the Electricity System Operator (TSO) are being mandated to ‘open’ the data collected on their networks, so it can be utilised by as many energy industry stakeholders as possible. They have risen to the challenge by publishing Data and Digitalisation strategies as well as making data the heart of their business plans. The challenge now is to implement these data management strategies.
In this Q&A style blog, Federico Sassi our Senior Product Manager outlines core data management principles and explores the role machine learning and artificial intelligence can play in supporting Network Operators in achieving their data and digitalisation strategies.
1. Can you tell us more about your role at Kelvatek and the work you do in developing our data solutions?
I first started my journey around 8 years ago, leading Kelvatek’s Machine Learning Group in Parma. During this time, I worked with an exceptionally talented group of Scientists and Engineers to undertake data analysis using state-of-the-art machine learning (ML) techniques.
In my current role as the Senior Product Manager for Data Products, I work with Subject Matter Experts, other Product Managers, and Technology Leaders across Kelvatek to design and develop new methods of extracting value from data gathered via monitoring devices on the network and delivering these insights to customers.
2. Data is becoming an increasingly critical asset to the entire energy industry, what are the key challenges Network Operators must overcome to gain the most value from their data?
One of the biggest challenges facing the industry is the scarcity of Data Scientists that are skilled in mathematics, statistics, databases, and programming languages. This is not only being felt in the energy industry but across many other sectors which makes it even more difficult to secure these experienced individuals. Having a team in place with the right skill set and expertise is an essential element of a successful data management strategy.
Over the last 8 years, Kelvatek has been developing specialised teams of Data Scientists and Subject Matter Experts. This approach ensures our teams are equipped with both ML domain knowledge and an in-depth understanding of the inner workings of the UK energy sector.
The availability, completeness, and accuracy of data is constantly evolving across the industry. The energy industry, today, faces the challenge of increasingly large, complex, and disparate data sets. It’s important to remember that data by itself doesn’t have any value (it is in fact a cost) until insights can be gained from it. It is therefore fundamental to identify the most valuable data to collect and transmit from devices on the network and decide what to store on cloud-based systems.
Several tools can be utilised to make this process easier. Cloud technology can be used for scalability and cybersecurity is essential in building a comprehensive security strategy that mitigates threats quickly.
As these new challenges present themselves, Kelvatek focuses on synthesising new models and creating powerful new approaches that have allowed our highly specialised team of data scientists and Ph.D. specialists to frequently learn, automate model building, and find hidden insights at an incredible speed. We share these insights with customers and work and support them in driving up network performance.
3. Can you outline best practices in terms of collecting, processing, and analysing customers' data?
It is important to collect data with a purpose in mind. In our case, the intelligent connected devices that we have deployed on the electricity network are designed with exactly this in mind. Without the correct hardware and configuration, the high resolution and high-frequency data needed for some data use cases like predicting a fault and giving a time frame prediction for that fault to manifest would not be possible. The main purpose of data collection is to gather information in a measured and systematic manner to ensure accuracy and facilitate data analysis which will assist customers with their issues and mitigate future problems in new and innovative ways. During the data-gathering stage, the first step we take is to mitigate any cybersecurity risks by using strong encryption technology on end devices to allow secure communication.
Data in its raw form cannot be utilised efficiently by any organisation and therefore needs to be consolidated in a central database. We have also adopted the use of cloud site platforms as a means of accessing our data which are built by leading technology companies.
The key to analysing data is understanding the insights you would like to derive from the data. Our team of Data Scientists, Data Engineers, and Machine Learning Engineers firstly identify the right data and the quality it needs to be using advanced algorithms. We take this data to our Subject Matter Experts who advise us on how it should be applied to meet customer requirements. Presenting the insights to the customers is the next stage. Ensuring the insights are clear and easily understood is vital in ensuring they gain the maximum value from it. At Kelvatek, we have developed visualisation platforms and customised dashboards to give customers secure access to their data.
4. Technologies such as Machine Learning (ML) and Artificial Intelligence (AI) have been successfully applied in other industries, how are we adapting these technologies and what value have they offered our customers?
At Kelvatek, ML and AI are established technologies and we have been successfully utilising them since 2014 through our dedicated Centre of Excellence in Parma, Italy. We collaborate with specialist universities, such as the University of Parma’s Ph.D. program to develop state-of-the-art technology from academia and adopt the latest algorithms into real-life practices.
AI and ML are critical tools for data capture, analysis, and collection of information that we, at Kelvatek, use for a range of purposes, including a better understanding of day-to-day operations, making more informed business decisions, and helping customers learn more about their networks.
We use these technologies as an automation tool to allow us to look at complex tasks which involve huge volumes of data and automate them into manageable ones. This enables the rapid deployment of solutions to customers. At Kelvatek, our Data Scientists train algorithms to make classifications or predictions, uncovering key insights within data mining projects.
With the application of our tried and tested AI algorithms, ML models, extensive experience, and market-leading technology we can progressively develop new innovative solutions to overcome the rising challenges of the net-zero transition where Network Operators are seeing unpredictable load patterns that affect asset health, network resilience, network utilisation, and network planning.
Data – and the insight it provides – drives everything we do. Our advanced analytics capabilities put Network Operators firmly in control of their networks and support strategic decision-making.