The Strategic Role of AI for Microgrids

Share Button

Sean McEvoy, senior VP Energy at Veritone, describes AI for microgrids and explains how it provides reliable power with speed and precision.

AI and microgrids

Sean McEvoy, Senior VP Energy, Veritone

The AI challenge: Predicting the unpredictable

Advanced microgrids manage a range of variables in their effort to produce the cleanest, most efficient and most reliable energy. Even under ideal circumstances, there can be massive amounts of data for a grid-connected microgrid to sort through, including weather conditions, fluctuating patterns of energy consumption and production, and fuel and electricity pricing  —  which all can change in real time. 

On top of that, as we increasingly strive to achieve climate goals with renewable energy, microgrids are taking on the task of managing the extreme variability of renewable energy.

How can microgrids provide reliable power, while optimizing load, capturing market opportunities and balancing the variability of renewables — and do so faster and with more precision? 

That’s where artificial intelligence (AI) comes into play.

AI for microgrids: The new tool for the microgrid management toolbox

AI software offers microgrid operators a new tool in their toolbox to predict what is often, otherwise, unpredictable. 

But what exactly is AI? 

AI is the ability of a machine — via computing iterations and algorithms — to perform cognitive functions ordinarily associated with the biological human mind, such as reason, perception, deduction, problem solving and other actions. These “brain” functions are performed using data and computing.  

AI can solve problems, learn patterns and iterations, and draw inferences that would otherwise be cumbersome for a human mind to accurately compute, especially with large data sets. AI, the cloud and big data have come together to process huge volumes of data and “learn” patterns from the data to accurately draw inferences.

Used in automation, robotics and production for many industries, AI is also now available for microgrids and their supporting technologies. 

AI can be applied in the planning, deployment and operation phases of a microgrid, so it benefits microgrid developers, equipment providers and integrators, and operators.

For microgrid developers, AI offers swift computation for real-time modeling of a massive amount of data, helping them make capacity sizing decisions about grid equipment, solar and wind configurations, and use of electric vehicle charging infrastructure. 

For equipment providers, AI provides modeling and simulation for efficient use of controllers, solar inverters, battery systems and other distributed energy grid resources.

For microgrid operators, AI removes the need for grid operators and asset managers to intervene in microgrid operations. Instead, energy is automatically delivered at the right price, place and time — without human actors making any decisions or pulling any levers.

AI modeling and simulation improves microgrid operations in many ways

The outcome of AI modeling and simulation helps drive microgrid management decisions. By processing large amounts of available data, machine learning is put to work and artificial intelligence is employed to help all stakeholders who manage microgrids make informed decisions. AI helps them consider the following:

    • Manual vs. autonomous: What operations, control and equipment should be manual and what should be automated and autonomous? 
    • Distributing and storing power: How should power be distributed and stored based on production and demand?
    • Managing renewable variability: How should the microgrid respond to variability in power generation — particularly renewables?
    • Enhancing forecasting: What is the accurate supply, demand and pricing forecast?
    • Reducing energy costs: How can energy costs be reduced as well as the total cost of ownership
    • Improving resiliency: How can the microgrid provide optimal energy reliability, whether connected to the larger grid or in an island configuration?
The brave new world of microgrids

Microgrids are front and center in the transformation from traditional grid models to decentralized energy. The transformation is driven by the “Four Ds”: decarbonization, decentralization, democratization and digitization. Together, the Four Ds are setting the world’s energy stage for more autonomous and independent energy production and management.

Businesses, communities, institutions, government and the military increasingly rely on autonomous microgrids as they seek to lower carbon emissions, reduce energy costs and enhance energy reliability. Cloud-computing software, machine learning, AI and real-time optimization offer the path forward for microgrids to achieve these goals easier, quicker and more autonomously.

This article was written by Sean McEvoy, senior vice president of energy at Veritone.

Share Button


  1. I like the term “machine learning” much better. Why? AI, (true AI) would mimic the actual biological brain function and as such, would also find ecological, economical, moral bias’s into the “thought process”. Actual AI over a period of time might also find “patterns” of data push outcomes that could be defined as ‘neurotic’ in nature. IS THIS a bad thing? It will be in the micro to millisecond World of data parsing, decision and action, if a neurosis intertwines with the logic to create an unintended outcome. Say the difference of avoiding a cascading grid failure or creating or allowing a cascading grid failure. Best be careful, what you ‘wish’ for, right now what is Science Fiction can become Science Fact. Finding out one needs neurosis ‘filters’ and ‘psychosis’ filters built into their system will slow down the process to create the “shades of gray” position of indecisiveness. When you get to AI that “thinks”, “do it and let GOD sort it out”, you may find yourself in deep trouble.