The Grid Has More Room Than You Think

When Your Rooftop Solar Gets Wasted Because the Grid Can't Handle It

Imagine spending $15,000 on a solar panel system and a home battery. On a sunny afternoon, your panels are generating more power than you can use. Your battery is full. The grid should absorb the surplus—but it can't. Not because the wires are broken, but because the system managing those wires has no idea how much flexibility it actually has.

This is the central paradox of modern electricity grids: as homes and businesses become power producers as well as consumers—what engineers call "prosumers"—the infrastructure designed to manage one-way power flow is straining under two-way complexity. And the tools meant to fix this, called dynamic operating envelopes, or DOEs, are themselves creating new problems.

A new paper from researchers at IIT Kharagpur, the University of Queensland, and IIT Kanpur offers a solution. Their approach, published on arXiv in July 2026, doesn't just tell prosumers how much power they can push onto the grid. It gives them a range—a zone of permissible operation—while still keeping the entire network safe and stable. The key insight is that the grid has more room to maneuver than current systems assume, and capturing that room means less wasted solar, lower costs, and healthier batteries.

The Problem with Fixed Limits

Before we get into the solution, it's worth understanding why current approaches fail. Australia, like much of the world, is experiencing a quiet revolution in distributed energy resources. Rooftop solar is now on roughly one in three homes in some states. Home batteries are following. Electric vehicles are beginning to add another layer of complexity.

This is good for emissions and good for individual electricity bills. But it's a headache for distribution network operators—the utilities responsible for getting power from the high-voltage transmission system to your street. When everyone is consuming, voltage drops predictably along power lines. When everyone is exporting simultaneously, voltage can rise above safe limits. Too much power flowing through lines designed for one direction causes congestion. The result can be anything from flickering lights to damaged equipment to full network failures.

The traditional response has been to set a ceiling: a fixed export limit. In Australia, many distribution companies cap prosumer exports at 55 kilowatts. This is straightforward and safe. It's also deeply wasteful. A home with a 10-kilowatt solar array and a 13.5-kilowatt-hour battery might regularly generate 15 or 20 kilowatts on a good day. Capping that at 55 kilowatts—or 5 kilowatts, as some networks impose—means throwing away power that the grid could use and the homeowner paid for.

Dynamic operating envelopes were invented to do better. Rather than a single static limit, a DOE adapts to conditions. On a cloudy day when few prosumers are exporting, the grid might allow higher exports. During a peak demand period when imports are straining the network, it might tighten limits. The idea is to use the network's real-time capacity more efficiently while staying within safety bounds.

But there's a catch. Current DOE systems work like a negotiation: the prosumer tells the utility what they want to do, and the utility says yes, yes, or cut back. "I plan to export 8 kilowatts," says the homeowner. "The network can handle 7, so your limit is 7 kilowatts." The system optimizes around the reported value.

This sounds reasonable until you consider forecasting error. That 8 kilowatts was a prediction based on expected sunlight and household consumption. The actual value might be 10 kilowatts if clouds parted, or 3 kilowatts if they rolled in. When reality diverges from the plan, the rigid DOE can force the homeowner to curtail—throwing away 10 kilowatts of clean solar because the system already allocated network capacity elsewhere. Or it might force load shedding, turning off appliances or draining the battery to stay within a limit that no longer reflects reality.

The root problem, as the researchers frame it, is that existing approaches treat the prosumer's reported value as a ceiling rather than a reference point. "These techniques do not allow the prosumer to exchange a larger quantity of power compared to the reported value," the paper notes. "Consequently, it could lead to prosumers strategically inflating their reported power exchange values to the DNO, which is not desirable."

In other words, users game the system because the system gives them no reason not to.

A Framework Built on Flexibility

The researchers' answer is to stop thinking of the DOE as a single constraint and start thinking of it as an allowed region. Their flexible DOE provides both an upper limit and a lower limit on active and reactive power exchange. The prosumer's reported value sits somewhere inside that range—not as a hard target, but as a preferred setpoint that the system tries to honor without sacrificing network safety.

The key word is tries. The flexible framework trades off optimality with flexibility. In exchange for accepting that the system won't always achieve the mathematically ideal result, it gains the ability to absorb uncertainty. If the sun comes out and a prosumer wants to export more than they originally planned, the flexible DOE might have room to accommodate that. If clouds roll in and they need to import more, there's lower bound slack to allow that too.

The mathematical machinery behind this comes from a technique called flexible optimization, which the researchers adapt from a 2024 paper by Simonetto and others. Rather than solving for a single optimal point, the framework optimizes over a set of acceptable solutions, building in a margin of error designated by the Greek letter epsilon (ε). At higher epsilon values, the system accepts greater deviation from perfect optimality in exchange for a larger feasible region. At lower epsilon values, it stays closer to the mathematically ideal solution but has less room to maneuver when conditions change.

"By providing both upper and lower bounds on active and reactive power exchange, prosumers are not restricted to a fixed buyer or seller role within an interval and can switch their operating mode as needed," the researchers explain. This matters because the distinction between "consumer" and "producer" is increasingly meaningless. A home with solar and storage might import in the morning, export at midday, and import again in the evening. Forcing each transaction into a rigid category misses how modern energy systems actually work.

The Prosumer's Side of the Equation

What makes this paper particularly novel is that the researchers don't just design the DOE framework in isolation—they also model how prosumers decide what to do in the first place. Previous work on DOEs often treated prosumers as passive actors who simply report their intentions. This paper puts a full optimization problem inside the prosumer's decision-making process.

Each prosumer is modeled as a small system with multiple components: solar panels, a battery, a controllable inverter, and local load. The goal isn't simply to maximize self-consumption or minimize bills in isolation. It's to minimize the total cost of energy procurement plus the cost of using the battery—specifically, the degradation cost of cycling the battery through charge and discharge.

Battery degradation is the elephant in the room for home energy systems. Every charge-discharge cycle wears lithium-ion cells a little. The battery that started with 13.5 kilowatt-hours of usable capacity might only have 12 kilowatt-hours after five years of daily cycling. This degradation has a real monetary value, but most DOE studies ignore it. The researchers argue this is a mistake: ignoring battery wear distorts the prosumer's optimal strategy, leading them to cycle their batteries more aggressively than they should.

To incorporate degradation into the optimization, the researchers use a piecewise linear model of cycle depth stress. Battery aging doesn't proceed linearly—it follows a curve that the researchers approximate with five linear segments. This keeps the math tractable (the overall problem remains a mixed-integer linear program, solvable with off-the-shelf solvers) while capturing the non-linear relationship between discharge depth and capacity loss.

The prosumers solve their local optimization problem over a prediction horizon—typically a day, broken into 15 or 30-minute intervals. They generate a plan based on forecasted solar generation and load demand. Then they report their intended power exchange to the distribution network operator through an intermediary called an aggregator.

This is stage one of a two-stage process.

Computing the Envelope

Armed with the prosumers' reported intentions, the distribution network operator now needs to compute DOE limits that keep the whole network safe while giving each prosumer as much room to maneuver as possible. This is where flexible optimization enters the picture.

The researchers model the low-voltage network using a simplified power flow representation called the Distflow model. Without diving into electrical engineering details, Distflow captures the relationship between power flows and voltage along distribution lines—specifically, the tendency of voltage to rise when power is injected and fall when it's consumed. The constraint the network operator cares about most is keeping voltage within a safe band—typically 0.95 to 1.05 per unit of nominal voltage.

The flexible optimization problem the DNO solves is more complex than a standard optimal power flow. The objective isn't to minimize cost or loss—it's to find a region of feasible operation rather than a single point. The key parameters are the flexibility parameter epsilon and a preferred setpoint for each prosumer (the power exchange they originally reported). The optimizer produces DOE limits that keep all network constraints satisfied while containing the prosumer's preferred setpoint within the allowed region.

The researchers validate this on a modified Australian low-voltage distribution network—a 24-bus system representing a suburban feeder with multiple prosumers, each equipped with solar, battery storage, and smart inverters. They simulate 24 hours of operation at 15-minute resolution, with forecast errors introduced to capture the reality that solar generation and load demand cannot be predicted perfectly.

The results show the value of flexibility.

What the Numbers Say

The researchers compare their flexible DOE approach against a conventional non-flexible approach—the kind used in most current systems, where the DOE limit is set at or below the prosumer's reported value. The comparison focuses on three metrics: total operational cost, curtailment (both solar and load), and battery degradation.

On operational costs, the flexible DOE comes out ahead. By giving prosumers room to adapt when forecasts miss the mark, the system reduces the amount of costly curtailment and expensive grid imports. The paper reports that the flexible framework achieves cost reductions primarily through "significant reductions in curtailment."

Curtailment is where the difference is most stark. In the non-flexible approach, when a prosumer's actual generation exceeds their reported value and the DOE doesn't allow it, that surplus solar is wasted. The flexible approach's lower bound provides headroom to accommodate higher-than-expected exports, capturing energy that would otherwise be lost. The same logic applies to load: when actual demand exceeds forecast, the flexible lower bound provides room to import more before hitting a constraint.

Battery degradation costs also improve under the flexible approach. This might seem counterintuitive—more flexibility should mean more cycling, which should mean more degradation. But the researchers explain this through the mechanism of reduced curtailment. When curtailment drops, batteries are used differently: they absorb surplus generation that would otherwise be wasted, rather than being cycled aggressively to manage forecast errors within overly tight constraints. The net effect is healthier batteries and lower lifetime costs.

The researchers also conduct a sensitivity analysis on the flexibility parameter epsilon. Higher values of epsilon create larger DOE ranges—more room to maneuver—at the cost of straying further from the mathematically optimal operating point. The analysis shows that even moderate values of epsilon (around 0.2 to 0.3 of the rated power) yield substantial benefits. Going higher provides diminishing returns; going lower sacrifices the flexibility advantage.

This figure shows how the DOE limits change across prosumers at a single point in time (12:00 noon) under different flexibility settings. The upper panels show active and reactive power bounds; the lower panel shows the corresponding voltage at each bus. Higher flexibility (larger epsilon) pushes the bounds outward, creating more operational room, while staying within the voltage constraints.

This figure breaks down the flexibility assigned to each prosumer under the same conditions. The amount of flexibility isn't uniform—buses that are more constrained (closer to voltage limits) receive less room than those with more headroom. This is the framework working as intended: distributing flexibility based on actual network conditions rather than applying a one-size-fits-all limit.

Why Reactive Power Matters

One detail that distinguishes this work from much of the prior literature is its treatment of reactive power. Most DOE studies focus on active power—the real energy that lights your bulbs and runs your refrigerator. Reactive power is the "wattless" component that flows back and forth between inductive and capacitive elements in the grid, important for maintaining voltage but not directly measured by your electricity meter.

The researchers argue that ignoring reactive power leaves performance on the table. Inverters—the electronic devices that convert DC power from solar panels and batteries to AC power for the grid—can control both active and reactive power output simultaneously, subject to their apparent power rating. By coordinating reactive power with active power within the DOE limits, the network can use inverter capabilities more efficiently, supporting voltage regulation while still respecting active power constraints.

In their formulation, the DOE provides four bounds per prosumer per time interval: upper and lower limits on active power, and upper and lower limits on reactive power. The prosumer can choose any combination within this four-dimensional region as long as their inverter's total apparent power stays within its rating. This is more flexible than a simple active power cap because it lets the inverter trade active power capacity for reactive power support (or vice versa) depending on what the network needs most.

A Realistic Test Bed

The validation uses a modified version of the IEEE 24-bus low voltage test network, which represents a typical Australian suburban distribution feeder. The modifications include adding solar PV and battery storage at the prosumer nodes, along with smart inverters capable of reactive power control. Load profiles come from typical residential patterns: morning and evening peaks, midday dips (when occupants are away), and variation by day type.

Solar generation follows realistic irradiance patterns with cloud cover introducing forecast uncertainty. The researchers introduce errors of approximately 15-20% on solar forecasts and 10-15% on load forecasts to represent typical prediction accuracy. These errors propagate into the real-time optimization, where prosumers must decide how to operate under actual conditions rather than predicted ones.

The hierarchical structure—prosumers optimizing locally, aggregator aggregating and forwarding information, DNO computing DOEs, prosumers re-optimizing under constraints—mirrors how such a system might be implemented in practice. The researchers note that their approach is computationally tractable: the DOE optimization for a 24-bus network with 20 prosumers solves in seconds, making real-time application feasible.

They also test scalability by examining how computation time grows with network size. The results suggest the approach can handle larger networks, though the researchers acknowledge that computational complexity becomes more challenging as the number of prosumers and the prediction horizon grow. This points toward future work on decomposition methods or distributed algorithms that could handle city-scale networks.

The Bigger Picture

Dynamic operating envelopes are not just an academic topic. They're the subject of intense regulatory interest in Australia, Europe, and elsewhere. The Australian Energy Market Commission has been exploring DOEs as a way to enable higher DER penetration without expensive network upgrades. The European Union's Clean Energy Package includes provisions for smart charging and active grid participation by distributed resources.

The fundamental tension is between flexibility and control. A fully controlled system—where the grid operator directly control controls every inverter and battery—could solve these problems, but it requires extensive communication infrastructure, privacy tradeoffs, and a level of centralized coordination that most grids aren't prepared to implement. A fully passive system—where prosumers do whatever they want within simple static limits—keeps things simple but leaves enormous value on the table.

DOEs represent a middle path: network-defined constraints with prosumer-controlled optimization inside those constraints. The flexible DOE extends this middle path by recognizing that constraints can be soft rather than hard, and that the best constraint is one that leaves room for the future to differ from the past.

The researchers' inclusion of battery degradation cost is also a sign of where the field is heading. As home batteries become more common, ignoring their operational cost becomes increasingly untenable. A system that treats battery cycling as free will push batteries harder than economics would recommend, leading to earlier replacements and higher lifetime costs for consumers. By modeling degradation explicitly, this framework gives prosumers better signals about when to charge and discharge.

What Remains Uncertain

No paper solves every problem, and this one is transparent about its limitations. The validation, while realistic, is still simulation-based. Real-world performance depends on factors that simulations struggle to capture: communication delays, measurement errors, corner cases in power flow, and the behavior of large numbers of heterogeneous prosumers interacting simultaneously.

The assumption that prosumers will honestly report their intended power exchange is also worth examining. The researchers note that one motivation for their approach is preventing prosumers from inflating their reported values, and they claim the flexible framework removes the incentive for such gaming. But this depends on how the system is implemented in practice and whether the aggregator or DNO can verify reported values against actual generation.

The computational scalability, while promising, needs further testing. A 24-bus network is small by distribution system standards. Real-world implementation would need to handle thousands of nodes, millions of prosumers, and second-by-second updates. The paper acknowledges this direction for future work.

There's also the question of how flexible optimization parameters—specifically, the epsilon values—should be chosen in practice. The sensitivity analysis shows the tradeoff between flexibility and optimality, but it doesn't provide a prescription for what epsilon value to use. In a real system, this might depend on forecast skill, network conditions, and how much the DNO is willing to sacrifice optimal efficiency for robustness.

Finally, the model doesn't currently account for prosumers without batteries or for electric vehicles, which represent a growing share of distributed storage. Extending the framework to handle heterogeneous prosumers—some with large batteries, some with small ones, some with EVs, some with none—is a logical next step.

What Comes Next

The path from this research to deployed technology runs through several checkpoints. The first is experimental validation at small scale: a neighborhood or campus distribution feeder with a dozen prosumers and full metering. The second is regulatory alignment: ensuring that flexible DOEs comply with market rules and grid codes in relevant jurisdictions. The third is commercial development: software that can integrate with existing utility systems and aggregator platforms.

The researchers themselves identify several extensions. Incorporating electric vehicles would expand the framework's relevance as EV adoption accelerates. Distributed computation approaches could scale the DOE optimization to city-wide networks. Learning-based methods for setting flexibility parameters adaptively—rather than through fixed rules—could improve performance over time as the system observes patterns in forecast errors and constraint violations.

There's also potential in combining flexible DOEs with local energy markets. If prosumers have explicit price signals for flexibility, the DOE framework could coordinate not just technical constraints but economic optimization. The current work assumes a fixed tariff structure; a more sophisticated version might use flexible DOEs to enable dynamic pricing that reflects real-time network conditions.

The Quiet Revolution

The transition to distributed energy is happening whether the grid is ready or not. Rooftop solar is cheap and getting cheaper. Home batteries are following the same cost curve that made solar panels ubiquitous. Electric vehicles will add another layer of generation and storage that can be called on for grid services.

What the grid does with this revolution—how it accommodates millions of bidirectional power flows, how it keeps voltage stable and lines from overloading, how it gives prosumers fair value for the flexibility they provide—will shape the economics of clean energy for decades.

The challenge isn't technical in the narrow sense. The physics is well understood. Power flow equations, optimization algorithms, communication protocols—none of this is speculative. The challenge is integration: weaving these capabilities into a system that was designed for one-way power flow, regulated by rules written for a different era, operated by utilities with different incentive structures, and owned by consumers who just want reliable, affordable electricity.

The flexible DOE framework is a small but meaningful contribution to that integration. It acknowledges that the future won't look like the past, that forecasts will be wrong, that prosumers should be partners rather than adversaries, and that the best constraints are the ones that leave room for the unexpected.

In the short term, that might mean your solar panels don't get curtailed on a sunny afternoon when the network has more capacity than anyone realized. In the long term, it might mean a grid that's more resilient, more efficient, and more equitable—one where the benefits of the energy transition flow to the people who made it possible.