Better demand prediction and asset management will help utilities improve profitability and navigate the coming disruption
To deliver electricity, gas, and water, utilities need to carefully match supply and demand, whilst managing a complex network of assets.
Sophisticated insights from data can make both demand forecasting, and asset management, more predictable and more profitable.
Even in the normal run of business, this can reduce costs and waste by improving network planning and optimising generators, pipes, cables, and substations. As distribution systems become more complex – disrupted by eg new energy sources (wind, solar) and changes is user demand (sustainable living, EV ownership) – data can help utilities stay ahead of changes, predicting new demand patterns, and deploying/upgrading assets strategically.
Data has the answers – but only if you know where to look
The data to do all this is increasingly available. Smart meters provide granular insights into energy and water usage. IIoT sensors and actuators throughout generation, distribution and storage give overviews and control of assets. Diverse data sources, such as weather, geological, consumer behaviour, engineering data on asset design, can add layers of value.
But data on its own means little. These are complex datasets and we must overcome many challenges if we are to get them to give up their insights.
Some of these challenges, which we discuss in our new whitepaper, include:
- Hard to find/hard to use data: A lot of legacy data hasn’t made into the IT system. Blueprints, maps and handwritten notes need to be found and translated before they can be fed into models.
- Insufficient data: Vast networks of underground infrastructure were laid before we understood the value of data, and are hard to get at to collect data on. And data collection is often not what you really want (the nearest address of a burst pipe, rather than it’s GPS coordinates). We need to find clever ways to infer data from imperfect data sources.
- Real world data: Data is not just abstract statistics; it is also measurements of physical systems. The engineers who collect this often worry that the data teams will ignore the real world context and just look for correlations. Creating conversations between the two are important to unblock this.
- Impenetrable models: When there is a risk of not keeping the lights on or the water flowing, we can’t just rely on opaque model outputs – we need to be sure the insight is valid. Complex techniques like machine learning must be explainable and understandable to those using them to make decisions.
- Data privacy: AI models are complex and hard to understand. Can we be sure that a model trained on smart meter or asset data couldn’t be reverse engineered to identify users or steal IP?
- Insufficient data experts: Utilities exist to provide electricity, gas, and water, not data. Data science tams are usually a small and overstretched and can be distant from the business itself. Teams may have expertise, but not necessarily the range of skills or remit to do everything that is needed.
Building an insight-driven organisation
The opportunities from data in utilities are huge, but so are the challenges of drawing insight from complex data. But all these can be overcome with the right approaches to data management, modelling, skills, and collaborations.
Getting all this right will allow utilities to become insight-driven. Such an organisation would have mastery of its data. It would constantly launch data projects, and deploy resulting models, to provide reliable predictions and automation, making the business ever more streamlined and responsive.
This will not happen immediately, but there are steps that can be taken now that will progress you along this journey.
In our new whitepaper, we discuss the opportunities for utilities to harness data to gather valuable business insights, the challenges of getting and working with that data, and how to start to overcome them.
About the Authors
Sector Director, Energy, Tessella.