Artificial Intelligence is no longer Science Fiction! Even though the type that actually mimics human behavior is still quite far away – at least for the time being – we have been surrounded by so-called “narrow AI” for a while. Narrow AI handles a single task type or a domain-specific problem and here we have passed a tipping point. What has become possible with data capture, storage and processing has put AI on steroids!
Cloud environments can now handle huge data volumes efficiently, enabling cost-effective AI solutions that will not only support utility operations but also enable better quality customer services. Machine learning on collated metering data is used to enhance revenue assurance, meet energy efficiency goals and strengthen the retailer value proposition in highly competitive energy services markets.
Potential and challenges of AI
AI has real potential whenever large enough data sets can be captured to effectively train the underlying machine learning algorithms. On the grid-op’s side, the transition from classic grid monitoring at the substation level to full residential smart metering has scaled available data volumes and resolution by orders of magnitude in terms of what the utility can leverage. On the retail and generation side, traders and originators have always been at the forefront of technology adoption, exploiting whatever data crunching can give an edge in their deals.
At Landis+Gyr we have had many conversations with energy companies looking to apply AI and machine learning for the optimal management and dispatch of assets. The low-hanging fruits for AI in utilities are mainly open-loop, or human-in-the-loop applications such as predictive analytics, asset management, fault detection, etc. It’s when you move to full closed-loop automation that you face the bigger challenges. Much of the “subject matter expertise” remains with a generation of seasoned utility professionals and has resisted codification for closed-loop algorithmic grid management. However, it must be expected that even some of the most intractable edge cases will fall to the advances of AI as these utility applications of AI mature.
Catalyst for disruption
Our industry faces disruption with or without AI. We have already seen how climate change has driven a backlash against hydrocarbon fuel sources, which in turn has driven a rise in intermittent renewable generation, storage technologies and electric vehicles, right across the globe. There is every reason to believe that the next decade will see an acceleration of this trend, as well as policy- and regulatory-driven incentives to experiment with new business models for grid resilience, security of supply and transactive energy.
This blog article is an extract of a round table discussion with industry experts as part of our recently released pathway magazine on the power of Artificial Intelligence. Interested to learn more?