⚡ electrical engineering

Resilience-Aware Multi-Objective Energy Management for Solar-Battery Microgrids Under Renewable Generation and Load Uncertainty

This study proposes a resilience-aware multi-objective energy management framework for solar-battery microgrids that optimizes the coordination of generation, storage, and flexible demand under uncertainty, demonstrating significant reductions in grid dependence and renewable curtailment while enhancing critical load service during grid outages.

Original authors: Denesh Das

Published 2026-10-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Denesh Das

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The sun does not shine with a steady hand, and the demand for electricity in our homes and businesses does not follow a predictable rhythm. When a solar panel system is paired with a battery to power a small, local network of buildings, the challenge becomes how to balance these shifting forces. If the sun is bright but no one is home to use the power, the energy goes to waste. If the sun sets and the grid fails, the battery must decide whether to keep the lights on for a hospital or a fire station, or perhaps just for a refrigerator. This balancing act is the heart of microgrid management. For years, engineers have focused on making these systems cheap to run, but a new approach asks a different question: how can we make them tough enough to survive when the main power line goes down?

A researcher at Lamar University has developed a new way to manage these solar-battery systems, one that prioritizes keeping essential services running during a blackout. The study, which uses a detailed computer simulation of a standard electrical network, proposes a framework that does not just look at the cost of electricity. Instead, it weighs the cost of using the battery, the waste of unused solar power, and the critical need to keep vital loads—like medical equipment or emergency lighting—powered when the rest of the system is in the dark. The goal is to find a schedule that keeps the lights on for the most important needs, even if it means turning off less critical devices like air conditioning or office computers.

To test this idea, the researcher built a virtual model of a solar-powered microgrid connected to a battery storage unit. The system was designed to handle the natural ups and downs of solar power and electricity demand over a full day. The simulation included a specific scenario where the connection to the main utility grid was cut off for six hours, mimicking a severe storm or infrastructure failure. During this islanded period, the system had to rely entirely on its own solar panels and battery. The researcher compared a traditional management strategy, which focuses on standard economic efficiency, against the new resilience-aware strategy. The traditional approach tries to minimize costs and keep all loads running as much as possible, while the new approach is willing to sacrifice non-essential power to protect the critical loads.

The results of the simulation revealed a clear shift in how energy is distributed. Under the new strategy, the system imported significantly less power from the main grid, dropping from nearly 12 kilowatt-hours to just under 5 kilowatt-hours over the course of the day. More importantly, the system stopped wasting solar energy. In the traditional setup, excess solar power that could not be used or stored was simply discarded, amounting to over 8 kilowatt-hours of lost energy. The new framework managed to use every bit of that solar power, either by charging the battery or exporting it to the grid before the outage, resulting in zero wasted energy.

The true test, however, came during the six-hour blackout. When the grid went down, the traditional strategy struggled to keep the most important loads running, leaving them without power for nearly 10 kilowatt-hours of energy over the outage period. The new strategy, by contrast, reduced this gap to just over 4 kilowatt-hours. This meant that the critical loads received power for about 89 percent of the time, a significant jump from the 74 percent achieved by the traditional method. To achieve this, the system made a calculated choice: it allowed non-critical loads to go without power for longer periods. The total amount of energy lost across the entire system actually increased slightly under the new strategy, but that loss was concentrated entirely on the non-essential devices, effectively shielding the vital ones.

The battery played a crucial role in this protection. In the new strategy, the battery was managed not just to smooth out daily fluctuations, but to act as a reserve tank for the emergency. Before the outage began, the system ensured the battery was nearly full, holding about 95 percent of its capacity. As the six hours of darkness passed, the battery discharged its stored energy specifically to feed the critical loads, while the flexible, non-essential loads were allowed to dim or shut off. This disciplined approach meant that even as the battery drained, the most important services remained online for as long as physically possible. The study also tested shorter outages of two and four hours, finding that the new strategy maintained 100 percent service for critical loads during the two-hour event and over 93 percent during the four-hour event, consistently outperforming the traditional method in every scenario.

This work suggests that the way we manage energy in a crisis requires a different set of priorities than the way we manage it on a normal day. By accepting that some lights might go out to save the ones that matter, a microgrid can become significantly more resilient. The study does not claim to have solved every problem of energy management, nor has it been tested on a physical power grid in the real world. It remains a controlled simulation, a rigorous test of a concept using standard network models. However, the numbers from the simulation are clear: a system designed with resilience in mind can drastically reduce the impact of a blackout on essential services, turning a total collapse into a manageable, if imperfect, survival. The path forward for this technology lies in testing these strategies with real-world data and larger, more complex networks, but the core idea—that we must be willing to make hard choices about what to power when the grid fails—has been shown to work in the virtual world.

Technical Summary: Resilience-Aware Multi-Objective Energy Management for Solar-Battery Microgrids

Problem Statement
Solar-battery microgrids face significant operational challenges due to the variability of photovoltaic (PV) generation, fluctuating electricity demand, battery limitations, and the potential for utility grid interruptions. Existing energy management strategies often fail to simultaneously address operating costs, battery degradation, renewable curtailment, and the continuity of essential loads during grid outages. Specifically, there is a gap in research that jointly considers operating cost, battery use, renewable curtailment, energy not served (ENS), critical load requirements, and outage resilience within a single optimization framework. The core problem is how to coordinate PV generation, battery storage, utility exchange, flexible loads, and critical loads under both grid-connected and islanded conditions while managing uncertainty in renewable generation and load demand.

Methodology
The study proposes a resilience-aware multi-objective energy management framework designed for solar-battery microgrids. The methodology follows a sequential process involving data preparation, uncertainty representation, microgrid modeling, multi-objective optimization, and performance assessment.

System Architecture: The framework utilizes the IEEE 33 Bus Distribution System as a standardized test network. It models five principal elements: PV generation, battery storage, utility grid exchange, flexible loads, and critical loads. Flexible loads are treated as part of noncritical demand, while critical loads are prioritized separately.
Uncertainty Representation: Instead of relying on fixed operating conditions, the framework represents renewable generation and load variations through operating scenarios. Outage conditions are simulated with durations of 2, 4, and 6 hours to evaluate system performance during temporary utility loss.
Optimization Model: The core of the method is a multi-objective optimization model that minimizes a weighted sum of five objectives:
1. Operating cost ( $C_{op}$ )
2. Battery degradation cost ( $C_{deg}$ )
3. Energy not served penalty ( $C_{ENS}$ )
4. Renewable energy curtailment ( $C_{curt}$ )
5. Resilience-related penalty ( $C_{res}$ )
The resilience penalty is specifically defined as the normalized critical load energy not served during an outage. The model is subject to power balance constraints, battery state-of-charge (SOC) limits, charging/discharging power limits, grid exchange limits, and critical load service requirements.
Operational Modes: The system operates in two modes. In grid-connected mode, the microgrid coordinates PV, battery, and flexible demand with utility import/export. In islanded mode (during an outage), grid exchange is set to zero. The optimization prioritizes critical loads, allowing noncritical flexible demand to be shed if local generation and storage are insufficient.

Key Contributions

Integrated Multi-Objective Formulation: The study introduces a unified optimization model that jointly considers economic operation, battery degradation, renewable curtailment, ENS, and resilience, addressing a gap where previous studies often treated these factors in isolation.
Resilience-Oriented Scheduling: The framework explicitly differentiates between critical and noncritical loads. During islanded operation, it dynamically allocates limited energy resources to prioritize critical demand, even if this results in higher total energy not served for noncritical loads.
Scenario-Based Evaluation: The approach evaluates performance across varying outage durations (2, 4, and 6 hours) and operating conditions, providing a controlled simulation-based assessment without requiring proprietary utility measurements.
Resilience Metrics: The paper emphasizes the use of "Critical Load ENS" and "Critical Load Service" alongside total ENS to better assess microgrid resilience, arguing that total ENS alone is insufficient when loads have different priorities.

Results
The framework was evaluated over a 24-hour horizon using an 8 kWh battery (3 kW charge/discharge limit, 95% efficiency, 50% initial SOC) with an islanding event starting at hour 18.

Grid Interaction and Renewables: Compared to a conventional dispatch strategy, the proposed strategy reduced grid import from 11.92 kWh to 4.86 kWh (a ~59% reduction) and eliminated renewable curtailment (reducing it from 8.58 kWh to 0.00 kWh). It also facilitated grid export of 8.58 kWh, indicating better utilization of surplus PV generation.
Critical Load Performance: During a 6-hour outage, the proposed strategy increased critical load service from 74.11% to 89.07% and reduced critical load ENS from 9.57 kWh to 4.04 kWh.
Trade-offs: The strategy resulted in a higher total energy not served (increasing from 17.40 kWh to 21.97 kWh) because it intentionally shed more noncritical load to protect critical demand. The relative simulation operating cost decreased from 3.10 to 1.10.
Outage Duration Sensitivity: As outage duration increased, critical load service under the proposed strategy decreased (100% for 2h, 93.53% for 4h, 89.07% for 6h), reflecting the finite battery capacity. However, it consistently outperformed the conventional strategy across all durations.
Battery Operation: The proposed strategy maintained a higher state of charge at the onset of the outage (approx. 95% SOC) compared to the conventional approach, demonstrating a reserve-oriented operation strategy.

Significance and Claims
The paper claims that the proposed framework successfully demonstrates that limited energy can be effectively allocated toward critical demand during islanded operation, thereby reducing renewable curtailment and grid dependence. The results indicate that a cost-only objective is insufficient for microgrid management; instead, a multi-objective approach that includes resilience penalties is necessary to balance economic operation with the continuity of essential services.

The author modestly notes that the study is based on controlled simulations using the IEEE 33 Bus system and synthetic inputs rather than physical microgrid measurements. They state that future work should extend the evaluation to documented public solar and demand datasets, additional IEEE distribution systems (69 and 123 Bus), and larger scenario sets to further validate the framework's performance across varied configurations.

More like this