The Challenge of Balancing Electricity Generation and Customer Load

Ever since Edison’s Pearl Street station came online, it’s been challenging to match the amount of electricity being generation with the customer’s fluctuating load. It’s a delicate balance requiring constant monitoring and adjustment. Too much generation means wasted electricity and lost money. Not enough electricity and the system becomes unstable. It’s a condition that needs a dynamic solution, but it’s only becoming more challenging as we transition to clean energy. Also, extreme weather events aren’t helping, they’re severely impacting it.

These are constant reminders that the grid’s aging infrastructure needs upgrading, but it’s expensive and slow to happen. In the U.S. more than 70% of transmission lines are over 25-years old, with some having reached 100 years of age. Last year only about 125 miles (201 km) of transmission were added. The average age of U.S. power transformers is around 40 years, which isn’t good considering their lifespan is about 25 years. The bottom line here is that the power grid was built a long time ago, for a time and place that no longer exist. Also, it was built using standards there are no longer relevant in today’s ecosystem.

After all, in the 1960s power quality meant keeping the lights on, but in today’s interconnected world even an intermediate power fluctuation costs our customers a lot of money. The term behind-the-meter (BTM) was meaningless in the 1970s, but on our modern power system the BTM segment is an integral part of the power delivery system. In the 1980s 100-year storms were rare storms, but today they’re commonplace and it’s the 500-year that’s happening more often.

If that isn’t enough, for several decades electricity demand has been relatively flat and then hyperscale datacenters hit. Last year the EIA (U.S. Energy Information Agency) announced that electricity consumption has been increasing about 2% annually, and will continue increasing, at that rate, for the next 25 years. That may not sound like much, but a 2% annual growth rate equals a 50% consumption increase over those 25 years. Still, there are technological solutions addressing these issues.

It's Time for a New Approach

Last month, “Charging Ahead” did a high level assessment of the AI-driven digitalization that’s becoming extremely popular with utilities and grid operators. Adnan Chaudhry, Siemens Energy’s digital grid guru provided some interesting commentary on how these applications are reshaping the power delivery system. One major takeaway from our discussion was the fact, “continuing to add assets in the grid alone won’t satisfy the growing hunger for electricity.” The power grid “needs a complete end-to-end digitalization,” and artificial intelligence (AI) is the catalyst making that happen.

Adding AI’s functionality to smart grid applications like AMI (advanced metering infrastructure), community microgrids, and VPPs (virtual power plants) has proven an effective addition for the struggle of supply-demand balancing. AI tames the flood of data from millions of customers’ meters and it makes the assortment of DERs (distributed energy resources) manageable. It allows all those small-scale DERs to be aggregated into a utility-scale resource, which directly addresses the supply-demand issues we’re facing.

When 2024 ended, it was estimated that there were roughly 55 gigawatts of DERs located BTM and they’re mostly small rooftop solar installations. Fifty-five gigawatts means there’s a profusion of them. Because of EIA’s consumption alert, there is a great deal of interest in harnessing this untapped quantity of DERs, which would go a long way in addressing the supply-demand dilemma. Keep in mind, managing these massive numbers of DERs isn’t possible without AI.

Speaking of numbers, remember that old geometry theorem, “The whole is greater than the sum of the parts?” In this case, those parts are AI microgrids, AI VPPs, and AI AMI, but it requires a more advanced technology. It has to be the AI-driven genre, which is the most innovative AI. It uses AI as the predominate force behind the system’s functionality. Also, there’s another component needed for this “whole,” and that’s AI-driven distribution automation. It keeps all the others working together. Before getting deeper into AI-driven distribution automation, let’s review some of the basics associated with the standard distribution automation package.

Managing Complexity

Conventional distribution automation systems have allowed utilities to remotely control and monitor the operation of their distribution systems for years. The earliest systems were limited in their abilities compared to modern schemes. They were strained to their limits as they tried respond to changing requirements of the power system. In addition, real-time response to grid changes was sparse. The introduction of smart grid technologies brought about interconnectedness and big-data, which improved power system awareness.

As the complexity of the distribution automation systems increased, so did the quantities of big-data and there was a marked decrease in user-friendliness. Early forms of machine learning were introduced and they became a real gamechanger. Today more sophisticated forms of AI have become commonplace and distribution automation has thrived. MarkertsandMarkets, a research company, projects the global distribution automation market will grow from an estimated US$20.56 billion in 2025 to US$40.40 billion by 2030. They noted that the global distribution automation market is expanding because of the global shift toward grid modernization.

The Answers Are Out There

Looking a little deeper into AI-driven functionality as applied to distribution automation shows it has revolutionized the technology’s decision-making, adaptability, and data driven insights to name a few features. One of AI-driven distribution automation’s strong points is big-data analysis and using it for forecasts and predictions. It has also improved its ability for breaching silo databases for information that hasn’t been available.

It can give today’s utility planning group the benefit of knowing weather patterns and being able to shift through historical data. The maintenance and operations can now be linked to information about how often a distribution circuit was being overloaded in real-time. In effect, AI-driven distribution automation systems can tap into various databases available that have been commonly utilized because it’s too complicated for manual data-mining, and that only touches the surface.

One last feature before moving on. Faults happen even to the best designed distribution circuits, but AI-driven distribution automation offers fast fault detection, identification, and rapid system restoration, which is very important for customers. These and other features have been developed to improve the power delivery system and we’ll explore them more later, but for now the topic being addressed is supply-demand support. That brings us back to the plethora of gigawatts BTM that can become AI-driven VPPs.

Flexibility is Needed

Aggregating BTM resources has been popular with third parties and utilities for many years, but this approach is somewhat different. AI-driven distribution automation working with AI-driven AMI, and AI-driven VPP technology, however, provides the power grid with a flexible power supply. It provides a quick response to fluctuations in the balancing of electricity’s supply and demand. That reduces the risk of outages and lessens voltage instability and provides better real-time control and monitoring, which lets the advance VPP utilize its DERs more effectively.

AI-driven distribution automation VPP platforms are a responsive technological application that is flexible and adaptable to the fluctuations common with balancing the supply and demand of electricity. Being located on the distribution network puts them close to the load, which means they do not need transmission line additions, or much if any permitting or interconnection delays. Their configuration can be tailored to deliver the exact performance requirements based on the utility’s needs, priorities, and they’re really user-friendly.

At the beginning of the year, DOE (Department of Energy) updated their VPP applications study. It started off saying they are a critical solution to the challenges facing our power grid. It went on saying, “VPPs are cost-effective solutions for balancing the grid that can be deployed at scale within six months to maximize the use and value of existing infrastructure.” DOE felt that “most utilities can implement some form of VPPs today without any policy or regulatory changes. However, VPP deployment has been highest in areas where state regulators and policymakers have implemented VPP-supportive actions.”

An Evoloving Situation

VPPs are an effective method that’s changing the dynamic landscape of supply and demand, but as DOE indicates utilities and grid operators need support and encouragement. There were over 33 gigawatts of operation VPPs in the U.S. at the end of 2024, so the technology has been proven. Still more regulatory and policy support is needed. VPP technology is an off the shelf technology and it has been working on many power delivery systems.

These AI-driven distribution automation VPP platforms are a major shift from traditional energy systems and the energy landscape shows it. They’re cost-effective solutions that are becoming increasingly more sophisticated. VPPs are developing more autonomous features for their operation. These feature are needed for quickly equalizing the power system’s supply-demand issues. It’s getting progressively more difficult maintaining that critical balancing point in a modernizing grid. Watching this digital grid technology evolve is going to be exciting!