The mission of this group is to bring together utility professionals in the power industry who are in the thick of the digital utility transformation. 

Post

Making AI work for Low Carbon Future Grid

image credit: © Andranik Hakobyan | Dreamstime.com
Jasjeet Singh's picture
Director- OT/AT Oracle Utilities

Jasjeet is trusted advisor for your digital transformation journey & creating new energy services pathways. He has consulted global utilities in defining value-driven Corp Strategy, Future...

  • Member since 2011
  • 15 items added with 6,853 views
  • Dec 8, 2021
  • 639 views

This item is part of the Special Issue - 2021-12 - Data Analytics & Intelligence, click here for more

Disclaimer: The views expressed in this article are my personal and does not necessarily reflect my employer views. Any statement/thoughts shared are only author’s personal experience and not binding for any purposes.

As we know, future of energy is bright, clean, autonomous and yet unpredictable in many ways. At COP 26, global economies have signed up for low carbon/ net-zero energy transition and electricity grid’s decarbonization underpins that future ambitious plan. While some countries are making great progress to introduce innovative policies and execute performance based regulatory models (RIIO-2, REV, p2025 etc.) and bold decisions to ensure clean energy gets the front seat, others are contemplating actively to introduce similar laws, policies and regulatory guidelines. It is often said, that technology has been leading in many ways but its potential is limited by regulatory policies or lack of will, but it is changing quickly in some parts of world where regulator are allowing digital technologies to be fully leveraged.

Your access to Member Features is limited.

Now, the key question for technology providers is- Are we really there yet? Have we found the right ways of “applying digital” to solve complex problems? Because by simply implementing new shiny systems, does not necessarily translate to meaningful business outcomes. Advancement in technology is happening, and it is certainly available to be leveraged. The major hurdle, however, is “how to apply” these advancements to ensure that energy transition is seamless/ less bumpy, and is not overly expensive for consumers to deliver the benefits in timely manner.

How does AI and Machine Learning can help to accelerate net-zero transition?

There are range of digital technologies available like IoT, Design thinking, RPA, Cloud, AI & Machine learning, etc.  From all these technologies, “Science of Data” has truly opened up new opportunities, and new ways of doing the business activities in an effective & insightful manner. Almost every utility has this fuel “raw data” which can lead to “valuable insights” by effectively applying AI & machine learning algorithms in an “engineer’s way”. When executed properly, AI/ML models can help to maximize the potential of DER’s and reduce the wasteful curtailments to achieve net-zero.

Many offerings in marketplace offer AI-based solutions to solve multiple use-cases across Network planning, Grid operations, Customer engagement, DER integration, Asset optimization etc. However, the use of AI models is usually not enough to deliver the outcomes that are reliable, beneficial and trustworthy to be considered by business users. For instance, while developing predictive models based, say ANN- Artificial Neural Networks, one must be aware of the AI modeling know-how, and engineering concepts, in order to deliver meaningful outcomes. Every ANN based model has fundamental ways of working, which is based on the “layers” and “activation functions”. We need to use right “layer” based on the suitability & applicability to business problem/use-case and then activate with models with relevant function (such as Relu, Sigmoid etc.). For instance, below are some common examples of utilities use-cases mapped to ANN models:

Types of AI/ML Models

Relevance to Utility Use-Case

Deep learning

To find patterns in an unsupervised way. For instance, to find pattern in network faults

Linear/ Logistic Regression

Linear models are used to predict continuous data, load forecasting, and logistic regression for categorical data, e.g. whether an outage will happen or not.

Convolutional- CNN models

To process unstructured data. E.g. LiDAR imagery data to identify low-conductor clearances, broken cross-arms, etc.

Recurrent- RNN models

To work with time-series data, such as smart meters data, historian data to predict or forecast

Xboost

Gradient boosting is typically applied to increase the effectiveness of an outcome. Applied as “combination of models” to solve complex problems, such as future network planning, predict network congestion, cable faults. Xboost is commonly known as Ensemble modelling (bagging, boosting).

So, what is the key take-away?

The use of digital technologies, specifically AI & ML models, is only as good as its applicability to use-case. After right selection, we also need to assign careful weights to each contributing “independent variable” which is going to determine the outcome “dependent variable”. Some use-cases are highly complex, so there might be need of combination of models to achieve the reliable predictions. If we can accurately predict near-real time to day-ahead/week ahead, the orchestration of DER’s would be highly effective, low-cost, and beneficial to customers & networks, to realize the net-zero future in an affordable, reliable and safe manner.

Jasjeet Singh's picture
Thank Jasjeet for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member
Discussions
Spell checking: Press the CTRL or COMMAND key then click on the underlined misspelled word.
Matthew Knott's picture
Matthew Knott on Dec 16, 2021

Thanks for the insights as to the importance of working this those that have the business/operational mindset. Understanding and prioritizing the outcomes will help establish the applicable model(s) necessary to achieve them. It would be great to read some examples/success stories of those that have taken this path and are bridging the gap to a net-zero future.

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »