The Data Driven Future of the Digital Utility levering the Power of AI / ML
image credit: Greenbird Integration Technology
- Jul 17, 2019 1:14 pm GMT
- 963 views
The Data Driven Future of the Digital Utility
What does the future utility look like?
According to Navigant, the future utility will utilize an Energy Cloud to share data across the utility’s ecosystem of partners, suppliers and customers. So how will this benefit utilities? How will utilities capture business value through disruptive energy platforms?
IDC reported that "by 2019, in order to support their digital transformation agendas, 25% of the top 100 utilities will cut IT costs at least 30% by migrating IT infrastructure into the public cloud."
The Transformation of DSOs
DSOs are racing to reinvent their roles and have started a profound transformation journey. The transition will be reborn in 3D: decentralized, divergent and digital. Machine Learning plays an important role in this data-driven transition. Access to data is crucial. This means systems integration will be a key factor for the evolution of Machine Learning (ML) and Artificial Intelligence (AI).
Main Benefits for DSOs
The main benefits (examples) that can be realized by DSOs on their journey towards becoming Platformed Digital Utility include:
o Outages prediction
Modeling outage risk requires AI to rapidly acquire and precisely analyze multiple geospatial statistics. Better prediction of outages minimizes their impact or even prevents them from happening. Having advance knowledge of where outages are most likely to occur enables utilities to position resources more efficiently. Data integration using advanced APIs combined with ML provides new insights from weather, smart meter data, network topology, historical events and asset data.
o Reduction of Non-technical loss (NTLs)
Large energy losses that cannot be accounted for, severely affect power system operations. Worldwide, utilities lose $96 billion annually. It’s a serious issue, especially in developing countries. In Brazil, for example, non-technical losses account for approximately 40% of the total electricity generated.
Today’s modern APIs provide integration to data and AI. These work together to detect usage patterns, access billing history and other customer data that corroborate and provide new insights into real-time and irregular energy flows and consumer behavior. One of Greenbird’s ecosystem partners, Mycroft Mind, has developed disruptive algorithm for NTLs, helping utilities avoid costly energy losses with a success rate of 30%.
o Microgrid management
The integration of Distributed Energy Resources (DERs), including microgrids, provides significant opportunities for DSOs to deploy ML algorithms. AI can optimize IoT devices and orchestrate complex connections. AI offers new opportunities to improve the integration of renewable energy resources and reduce the losses experienced in the energy grid today.
o Demand-side Management (DSM)
Most DSOs have limited monitoring of the real-time status of their low voltage network at the 400V level. ML allows DSM to be automated and optimized in a much smarter way. Moreover, DSOs are able to better utilize the existing grid capacity while maintaining service reliability. Data used for energy forecasting is more dynamic than ever before due to smart metering implementations. Thus, an adaptive ML algorithm may reduce forecast error when the inputs are highly variable.
o Predictive Maintenance
In the power grid itself, utilities can apply AI to the data from IoT devices in order to monitor events that affect grid reliability. DSO can use data from smart line sensors and inspection images collected from drones to build analytical models. In this context, deep learning algorithms can help to identify defects and predict failure. ML can anticipate grid disturbances and automatically issue mitigating controls to avoid outages.
o Yield Optimization
Energy production depends on the availability of generation assets. DSOs can use ML to optimize the start-up & shut-down process of rotating equipment, predict future maintenance needs based on performance degradations and prevent unplanned downtime. As McKinsey stated, "ML optimizes a wind turbine’s yield based on past performance, historic data, real-time communication with other wind farms and changes in wind speed and direction."
Adaption and Disruption
Are utilities acting fast enough to capitalize on the benefits gained from machine learning?
According to Thorsten Heller, “Adaptation to disruptive technologies needs to happen at a much faster pace.” It starts with having a data-driven strategy that focuses on integrating systems and managing the information flow to generate new opportunities and foster innovation.
Greenbird is building an ecosystem of partners and clients who are working towards accelerating the energy revolution.