AES Corporation is one of the world’s largest independent producers of renewable energy, green hydrogen, and battery storage.
This Fortune 500 company is the only US-based power company that operates on four continents. Additionally, AES owns utilities in the US and Latin America and has a technology innovation company that focuses on overcoming key market barriers with new and strategic energy innovations.
We recently interviewed Raiford Smith, Chief Utility Innovation Officer at AES Utilities. We talked about everything from edge computing to destroying hardware through optimization (not optimal!), ChatGPT, and whether cattle cause the most power outages (you read that right). Check out our Q&A session.
Question: We would love to hear more about your role as the Chief Utility Innovation Officer at AES Utilities. What are some ways you think about innovation in the utility industry that might surprise people? How do AI and analytics play a role in those innovations?
Raiford Smith: One of the funniest things I did in my career was destroy hardware … by accident. In 2012, we theorized decarbonization would require edge-based optimization for the grid and built infrastructure to test this hypothesis. Unfortunately, we quickly realized the physical infrastructure couldn’t keep up with the higher fidelity data and software optimization available at the edge.
We couldn’t just build out a super-fast digital infrastructure and declare victory. We needed physical infrastructure that could keep up with the digital infrastructure. While decarbonization has proved the need for edge-based capabilities, transformation doesn’t exist in a silo – it happens across multiple areas simultaneously, including with our assets, people, and processes.
Q: The past three years have been filled with disruption. Can you provide an example of how AES applied analytics to monitor and improve the health and performance of your physical assets?
Smith: One of the key areas where AES has used analytics to track and improve the performance of our assets has been around our wind generation units. We deployed SAS® Viya® to help us better understand and spot trends in key components and understand when those components might fail. We also wanted to find ways to improve the performance of the asset, extend its uptime, and ensure its availability for clean energy around the clock. We monitor and manage renewable and storage assets all over the world. We built some impressive analytics to spot issues early and take proactive action to keep clean energy production going round-the-clock.
Q: Can you comment on how continuous improvement intersects with and guides applications of advanced analytics to achieve improvement in process performance?
Smith: One of the most insightful things I learned early in my career was using analytics to improve our outage response time. Surprisingly, one of the leading causes of outages was “cattle.” Now that may seem unintuitive to most people – it certainly was to me. When I dug in, it turned out “cattle” was the first choice alphabetically on the outage cause code list, making me suspicious that it wasn’t really the leading cause at all.
We’d spent a lot of time and effort building new technology but apparently not enough time actually improving our processes to capture data accurately.
And so one of the most exciting and engaging things for us to do is think about not just building interesting insights and new analytics, but also ensuring that the processes, procedures and people involved are similarly upskilled so that we’re not getting unintuitive answers like cattle are rampaging, causing outages.
Q: In the past, you’ve spoken on the importance of data governance across the enterprise. How does an integrated analytics platform that covers the entire analytics life cycle help AES adapt to change?
Smith: A lot of analytics solutions are narrowly tailored to do a few things in a few spaces. One of the nice things about a platform like SAS is its ability to hook into many things and gives us a single pane of glass from which to understand our data, improve our data, and build insights. To me, that’s super critical in terms of accelerating your transformation. It helps us get there faster.
So an integrated analytics platform really is an important consideration for organizational change management.
Q: In the Internet of Things (IoT) space, we often hear about analyzing data at the edge. By that, I meant sorting through large streams of data at the source and watching for any anomalies. In your case, the data source might be windmills, solar panels or other energy sources. Can you talk about the benefit of having visibility into that data at the time it’s created?
Smith: Energy systems have to balance supply and demand in fractions of a second. So time – and being able to do all of that computational capability in the shortest amount of time – is critically important. The way the grid operates today gives us a lot of leeway because we have buffers built into the system. But as we add new, decarbonized technologies, that buffer disappears. One way to gain the speed we need is to build out analytics that work at the edge. Edge computing also helps us improve the resiliency and security of the grid because it changes the way we manage and use controls and removes some reliance on humans.
Q: We’ve heard a lot about generative AI solutions. Putting aside the pros and cons for a minute, if generative AI technology continues at its maturity rate, how do you think it will affect your industry over the next three to five years?
Smith: In the energy space, we rely on senior leadership with 25+ years of experience to make our decisions. They come up with the rules and provide insights and ideas based on the vast knowledge and expertise they’ve acquired over the decades. If we were to ask them to upskill or change the nature of their work, it would be really hard without some way to help augment their capabilities and supplement their already vast insights. Supplementing the skills of our existing workforce so they can focus on higher-level thinking and more strategic, innovative planning is where I think generative AI solutions will make an impact.