Researchers Cut Quantum Resource Demand for Power Grid Islanding

A new method limits the spread of disturbances in electrical grids through controlled islanding, partitioning a compromised grid into connected, electrically sustainable islands. Classical methods face sharply growing computational costs as network size and island count increase. Quantum optimisation offers an alternative for exploring this combinatorial partition space. However, monolithic quantum formulations encode all assignment decisions in one circuit, causing qubit demand and circuit complexity to scale with network size. In this study, Yuqi Jiang of Tsinghua University and colleagues propose a qubit-bounded sequential distributed quantum approximate optimisation algorithm (QAOA) framework to tackle coherent controlled islanding under limited quantum resources. It formulates the optimal islanding strategy through a series of sequential QAOA optimisations. Distributed quantum algorithm tackles large-scale power grid partitioning with fewer qubits A five-fold reduction in qubits needed for controlled islanding has been achieved, resolving problems with 300 buses, a scale previously inaccessible to monolithic quantum approximate optimisation algorithm (QAOA) approaches. Modern power systems are undergoing a significant transformation with the increasing integration of distributed energy resources (DERs) such as solar photovoltaic arrays, wind turbines, and energy storage systems. While these DERs offer numerous benefits, including increased resilience and reduced carbon emissions, they also introduce substantial variability and uncertainty into the power grid. This variability stems from the intermittent nature of renewable energy sources and the decentralised control of these resources. During disturbances, such as faults or sudden load changes, these effects can intensify generation-load imbalances and potentially lead to cascading failures, resulting in widespread blackouts. Controlled islanding, a proactive grid management technique, aims to mitigate these risks by intelligently partitioning the compromised grid into smaller, interconnected ‘islands’ that can operate independently and maintain electrical stability. However, determining the optimal islanding strategy, identifying which lines to open and close to create these islands, is a computationally challenging combinatorial optimisation problem, particularly for large-scale power systems. Current quantum hardware imposes limitations due to qubit availability and coherence times, restricting the size of effectively solvable problems. The number of qubits represents the fundamental unit of quantum information, and their limited availability necessitates efficient algorithms that minimise qubit requirements. Coherence time refers to the duration for which a qubit can maintain its quantum state before decoherence occurs, leading to errors in computation. Researchers at Kent State University, collaborating with Real-Time Power, developed a sequential distributed QAOA framework that formulates the complex task of partitioning power grids into smaller, manageable regional subproblems. This approach leverages the principles of divide-and-conquer, breaking down the overall problem into a series of sequential optimisations, each addressing a specific region of the grid. The sequential nature of the algorithm allows for the reduction of qubit demand, as each subproblem requires fewer qubits than a monolithic formulation that attempts to solve the entire problem at once. Furthermore, distributing the computation across multiple sequential optimisations can potentially improve the algorithm’s robustness and scalability. Eleven IEEE standard power systems, ranging from nine to 300 buses, were used to evaluate the framework, representing a significant increase over previously achievable scales. A ‘bus’ in a power system represents a node where power is supplied or consumed, and the number of buses is a common metric for the size and complexity of the grid. The use of standard IEEE test systems ensures the reproducibility and comparability of the results. Larger grids did not exponentially increase demands on the quantum processor, as the computational workload increased linearly with network size. This linear scaling is a crucial advantage, as it suggests that the algorithm can handle increasingly complex grids without experiencing a prohibitive increase in computational cost. Five-fold increases in gate fidelity were observed, and the sequential distributed quantum approximate optimisation algorithm (QAOA) framework successfully recovered feasible solutions matching conventional methods, even when simulating quantum noise. Gate fidelity is a measure of the accuracy of quantum gates, and higher fidelity indicates lower error rates. The ability to recover feasible solutions even in the presence of simulated quantum noise demonstrates the algorithm’s robustness and potential for practical implementation. Although substantial improvements in qubit coherence and error correction are still needed to realise a true quantum advantage, and current results do not yet demonstrate a speed advantage over classical algorithms, the work highlights the potential for quantum computing in power grid control. The method’s durability and practical applicability are suggested by its success in simulating imperfections of real quantum hardware, while the linear scaling of computational workload offers a pathway to handling increasingly complex grids. Sequential quantum partitioning paves the way for scalable grid stability solutions Strong methods for maintaining grid stability during disturbances are increasingly necessary with the growing prevalence of distributed energy resources, and controlled islanding offers a vital defence against cascading failures. The increasing penetration of DERs necessitates advanced grid control strategies capable of adapting to dynamic and uncertain conditions. Traditional control methods often struggle to cope with the complexity and scale of modern power systems, highlighting the need for innovative approaches.

The team and Real-Time Power have demonstrated a sequential quantum approach that tackles this challenge by breaking down complex partitioning into smaller, more manageable regional problems. This decomposition allows for the application of quantum optimisation techniques to each region independently, reducing the overall computational burden and enabling scalability. The study acknowledges its current reliance on recovering solutions already achievable with classical optimisation techniques, specifically the Gurobi solver, but establishes a key pathway for scaling quantum optimisation to larger, more complex power grids. The Gurobi solver is a widely used commercial optimisation software package, and its use as a benchmark provides a valuable point of comparison for the performance of the quantum algorithm. By adopting a sequential approach, the researchers circumvented limitations inherent in traditional quantum formulations that struggle with scaling to realistic grid sizes. These limitations arise from the exponential growth in qubit requirements and circuit complexity as the problem size increases. The sequential distributed QAOA framework effectively mitigates these limitations by decomposing the problem into smaller subproblems, each of which can be solved with a manageable number of qubits. This method decomposes the complex islanding problem into smaller, regionally focused optimisations, sharply reducing the demand for quantum resources and circuit complexity. Consequently, it paves the way for future optimisation of the framework’s performance and efficiency. The reduction in required resources represents a step towards practical quantum solutions for power grid management, and further research will focus on enhancing the algorithm’s speed and robustness. This includes exploring advanced quantum algorithms, improving qubit coherence times, and developing more efficient error correction techniques. Ultimately, this work demonstrates the potential of distributed quantum algorithms to address real-world challenges in critical infrastructure, offering a promising avenue for enhancing the resilience and reliability of future power grids. The research demonstrated a method for partitioning power grids into stable ‘islands’ during disturbances using a sequential quantum approximate optimisation algorithm. This approach overcomes limitations in scaling quantum computations to larger networks by breaking down the problem into smaller, regionally focused optimisations. Across eleven IEEE test systems, ranging from 9 to 300 buses, the framework successfully recovered solutions comparable to those found using classical optimisation software. The authors intend to improve the algorithm’s speed and robustness through further development of quantum algorithms and hardware. 👉 More information🗞 SPLIT-Q: A Scalable Sequential Quantum Computing Framework for Coherent Controlled Islanding✍️ Yuqi Jiang, Zhiding Liang, Qiang Guan, Yan Li and Ganesh Kumar Venayagamoorthy🧠 ArXiv: https://arxiv.org/abs/2608.12711 Stay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: