Why Upskilling Power Utility Teams in Data Science and AI Is Mandatory to Address Today’s Challenges

In this series, we have explored the complicated relationship between emerging technologies such as AI, and the power industry that both serves and uses them. The previous articles covered some of the innovative solutions helping utilities navigate these market shifts. The insights from our work with clients in the field have highlighted the rising role of data science and AI technologies in helping this vital sector navigate uncharted energy terrain. This article particularly shines a spotlight on the skills the industry’s workforce will need to develop, implement, and forge successfully ahead with data-driven solutions.

The growth of data

If the journey toward environmental sustainability will be powered by wind and solar energy, and if the road to societal improvement will be paved with AI and other innovations, then success in meeting these challenges will be fueled by data. For energy providers, the critical task lies in identifying and harnessing data to serve their aims amidst the burgeoning availability of sources.

Examples of existing data sources that utilities can leverage range from inspection records and sensor data to satellite-based imagery, and from weather and climate data to information on consumer preferences and usage.  Further, with the advance of newer technologies, such as two-way communication smart grids, smart meters, and IoT sensor networks, energy companies can capture much larger volumes of data, both structured and unstructured.

By better leveraging this growing data ecosystem, power utilities are finding ways to manage supply to more precisely equal demand without waste, optimize network performance, and fine-tune investment decisions.  Innovative applications catalyze better, understand network status and operations, increase operational efficiency, and consequently make rate increases a less common experience for the consumer. 

The complexity challenge

Energy demand is evolving in ways that are both more dynamic and more complex than the industry has traditionally seen, and we are now well beyond the era when an engineer could design the next phase of a transmission network by instinct. We’re also seeing the use of much more data, and more types of data, in the assessment of network quality and related topics such as predictive maintenance planning.

For example, power utilities can now leverage drone-based video-monitoring data in real time to assess network health without relying on dispatching crews to remote locations, and groundbreaking work identifying and resolving network misbehavior (known in the trade as resonances or synchrophasor events) help improve the stability of the transmission network.

The explosive growth of the types and volumes of data is both part of the problem and part of the solution for organizations trying to harness the data at their fingertips.  Data-driven approaches to the industry’s challenges will require teams equipped with data skills. For many organizations, this will mean upskilling and reskilling their workforce.

Skills essential for today's energy workforce

Organizations in today’s energy sector need to expand their teams’ existing analytics and machine learning capabilities to support such functions as analyzing transmission irregularities, performing predictive maintenance, and detecting anomalies in energy buying. Some of the specific skills that will help them achieve these objectives include:

• Clustering Techniques

• Principal Component Analysis

• Classification and Supervised Machine Learning

• Logistic regression

• Decision Trees

• Ensemble Methods

• Anomaly Detection

• Neural Networks

• Feed Forward Networks

• Deep Learning

• Generative AI

The most successful utilities seek to put in place training programs to give their staff the skills they need to solve real-world business and engineering problems using data science techniques and datasets that are relevant to their work. Two models have worked well:

1. Training programs in specific data science analytical skillsets: One regional electrical ISO created a bespoke program for its Advanced Analytics Group, empowering analysts to spearhead new data science-focused investigations and analysis in regional transmission.

   Working with us, they developed a custom course curriculum that included a variety of Machine Learning and AI topics, which accelerated the client’s ability to understand its markets and more effectively address its mission of increasing market stability.

2. Complete data science academy: Another large utility that aspired to institute a widespread cultural evolution through data literacy worked through its internal data science teams to develop a nine-month program, focused on Python, that incorporated the company’s use cases and datasets.

   In addition to their assignments and classwork, participants complete a capstone project, related to their work responsibilities, that they present to their peers and managers at the end of the program. Through this program, this utility has trained over 1,000 professionals across the client’s organization.

By combining AI, machine learning, and related data science technical skills, utilities must seek ways to put in place technical solutions and the skills to leverage these solutions to their consumers’ best advantage.

About the Author:

Fred Knops a guest contributor to CDO Magazine, is a Sr. Vice President at Data Society and head of their Utility Team.  Data Society currently offers more than 140 courses in topics ranging from data science programming languages R and Python, to topics like Storytelling with Data and Ethics and Bias in Data.  The two examples, above, highlight Data Society’s client work.