This paper addresses the problem of protecting network information from privacy system identification (SI) attacks when sharing cyber-physical system simulations. We model analyst observations of networked states as time-series outputs of a graph filter driven by differentially private (DP) nodal excitations, with the analyst aiming to infer the underlying graph shift operator (GSO). Unlike traditional SI, which estimates system parameters, we study the inverse problem: what assumptions prevent adversaries from identifying the GSO while preserving utility for legitimate analysis. We show that applying DP mechanisms to inputs provides formal privacy guarantees for the GSO, linking the (ϵ, δ)-DP bound to the spectral properties of the graph filter and noise covariance. More precisely, for DP Gaussian signals, the spectral characteristics of both the filter and noise covariance determine the privacy bound, with smooth filters and low-condition-number covariance yielding greater privacy.

This work addresses a fundamental challenge in applying deep learning to network topology-dependent power system applications: developing neural network models that transfer across significant system changes, including networks with entirely different topologies and dimensionalities, without requiring training data from unseen reconfigurations. We focus on situational awareness tasks where predictions inherently depend on network-wide spatial relationships through electrical connectivity. Despite extensive research, most ML-based approaches remain system-specific, limiting real-world deployment. This limitation stems from a dual barrier. First, topology changes shift feature distributions and alter input dimensions due to power flow physics. Second, reconfigurations redefine output semantics and dimensionality, requiring models to handle configuration-specific outputs while maintaining transferable feature extraction. Our core insight is treating topology as a learnable dimension through training on diverse reconfigurations, enabling the model to learn universal power flow physics rather than memorizing fixed networks. We achieve this through three synergistic mechanisms: scalar-weight graph convolutions with physics-aware operators handle variable input dimensions and distribution shifts, adaptive pooling maps topology-specific features to unified latent representations, and parallel transformer generates configuration-specific outputs using positional encoding. Our Universal Graph Convolutional Network (UGCN) achieves transferability to any reconfiguration or variation of existing power systems without any prior knowledge of new grid topologies or retraining during implementation. Experimental results on state forecasting and false data injection detection demonstrate that UGCN significantly outperforms state-of-the-art methods in cross-system zero-shot transferability across both transmission and distribution networks.

This work addresses a fundamental challenge in applying deep learning to power systems: developing neural network models that transfer across significant system changes, including networks with entirely different topologies and dimensionalities, without requiring training data from unseen reconfigurations. Despite extensive research, most ML-based approaches remain system-specific, limiting real-world deployment. This limitation stems from a dual barrier. First, topology changes shift feature distributions and alter input dimensions due to power flow physics. Second, reconfigurations redefine output semantics and dimensionality, requiring models to handle configuration-specific outputs while maintaining transferable feature extraction. To overcome this challenge, we introduce a Universal Graph Convolutional Network (UGCN) that achieves transferability to any reconfiguration or variation of existing power systems without any prior knowledge of new grid topologies or retraining during implementation. Our approach applies to both transmission and distribution networks and demonstrates generalization capability to completely unseen system reconfigurations, such as network restructuring and major grid expansions. Experimental results across power system applications, including false data injection detection and state forecasting, show that UGCN significantly outperforms state-of-the-art methods in cross-system zero-shot transferability of new reconfigurations.

Internet of Things (IoT)-enabled electric vehicles (IoEVs) enable intelligent charging coordination that accounts for grid congestion. However, increased data exchange raises privacy concerns, as charging patterns can reveal sensitive driver behavior to grid operators. We propose a differentially private (DP) EV charging framework that enables coordinated control while protecting driver data with theoretical privacy guarantees. Nevertheless, integrating DP inevitably introduces uncertainty into the control strategy for EVs, which can lead to infeasible solutions. To tackle this challenge, we develop a feasible and scalable control algorithm based on constrained reinforcement learning (CRL) and convex hulls. While our framework is designed to handle the uncertainty introduced by DP, it is general and also applicable to other sources of uncertainty in EV charging, such as the stochastic nature of driver behavior and renewable variability. This ensures feasible and privacy-preserving coordination of EV charging at scale. Our method constructs convex hulls within the action space to guarantee feasibility under stochastic constraints and incorporates constraint reduction techniques to improve scalability. Case studies based on IEEE benchmark systems demonstrate that the proposed approach effectively balances feasibility under uncertainty, scalability, and privacy in large-scale EV charging control.

Modern power systems face security threats from cyber-physical attacks targeting smart inverters and False Data Injection (FDI) attacks manipulating measurement data. These attacks have different objectives and require distinct mitigation strategies, making accurate attack classification critical. Traditional detection methods struggle to differentiate between attack types and benign measurement noise, causing high false alarm rates and dangerous misclassifications. This work proposes a unified detection framework that simultaneously addresses both attack types under noisy conditions. The framework leverages a robust Graph Convolutional Network (GCN) to exploit power grid topology, enabling effective discrimination between benign noise, FDI attacks, and cyber-physical attacks. Experimental validation on OpenDSS three-phase unbalanced distribution systems demonstrates superior performance compared with other benchmarks, achieving AUC scores exceeding 0.93 for all attack scenarios, significantly reduced false positive rates, and enhanced robustness to measurement noise.

The effective representation, precessing, analysis, and visualization of large-scale structured data over graphs are gaining a lot of attention. So far most of the literature has focused on real-valued signals. However, signals are often sparse in the Fourier domain, and more informative and compact representations for them can be obtained using the complex envelope of their spectral components, as opposed to the original real-valued signals. Motivated by this fact, in this work we generalize graph convolutional neural networks (GCN) to the complex domain, deriving the theory that allows to incorporate a complex-valued graph shift operators (GSO) in the definition of graph filters (GF) and process complex-valued graph signals (GS). The theory developed can handle spatio-temporal complex network processes. We prove that complex-valued GCNs are stable with respect to perturbations of the underlying graph support, the bound of the transfer error and the bound of error propagation through multiply layers. Then we apply complex GCN to power grid state forecasting, power grid cyber-attack detection and localization.

Contemporary neural network (NN) detectors for power systems face two primary challenges. First, each power system requires individual training of NN detectors to accommodate its unique configuration and base demands. Second, significant changes within the power system, such as the introduction of new substations or new generators, necessitate retraining. To overcome these issues, we introduce a novel architecture, the Nodal Graph Convolutional Neural Network (NGCN), which utilizes graph convolutions at each bus and its neighborhoods. This approach allows the training process to encompass multiple power systems and include all buses, thereby enhancing the transferability of the method across different power systems. The NGCN is particularly effective for detection tasks, such as cyber-attacks on smart inverters and false data injection attacks. Our tests demonstrate that the NGCN significantly improves performance over traditional NNs, boosting detection accuracy from approximately 85% to around 97% for the aforementioned task. Furthermore, the transferable NGCN, which is trained by samples from multiple power systems, performs considerably better

The increasing integration of converter-interfaced generation within large-scale synchronous power systems is presenting both new challenges and opportunities in how we operate these dynamical networks. One of these challenges is designing and parameterizing the digital control loops that dictate the dynamical behavior of these fast-acting resources. Improperly tuned gains, or unexpected controller couplings through the network, can lead to poorly damped oscillations during disturbance conditions and/or decrease the system stability margin. Within this work, we present an approach for adaptively tuning a damping controller to improve the dynamic response of converter-interfaced generation, based only on local measurements. We show that, with low-amplitude probing, we can identify a subset set of observable system modes. We then propose a linear MISO state-feedback controller to improve the damping ratio and stability margin of the system. We show the effectiveness of the proposed controller through both eigenvalue analysis and time-domain simulations. We first demonstrate the proposed approach on a simple 3-bus system followed by a larger test case, the IEEE 14 bus system, with multiple devices simultaneously probing the network.

This paper presents a comprehensive stochastic optimization model that seamlessly integrates aggregate electric vehicle (EV) charging demand response with power grid system operations, leveraging the inherent flexibility of EV charging. Our main novel contribution is tackling the problem of uncertainty in the demand characteristics. In our stochastic model, we capture not only unknown user charging patterns but also the effect of a pseudo-randomized mechanism applied to provide differential privacy (DP) guarantees to users whose charging patterns are not disclosed. From a control perspective, the intrinsic randomness of the users charging needs, compounded with randomness introduced by the DP mechanism can easily result in infeasible solutions. To overcome this challenge, we adopt a robust optimal control strategy that encompasses the intersection of potential sampling scenario-based constraints. In addition, to manage the high-dimension of the control action space, we approximate the intersections of the feasible regions with a reduced set of polyhedron constraints. In conclusion, our case studies based on IEEE standard systems demonstrate that the proposed algorithm effectively addresses robustness, scalability, and differential privacy for EV users by dynamically adapting to control the demand response for renewable energy integration while consistently ensuring the privacy of EV drivers.

Deep Reinforcement Learning (DRL) algorithms have become popular for solving power system control problems. In conventional DRL, the agent is incentivized to explore all policies that can be encoded in a neural network (NN) with the sole objective of maximizing the reward function. In doing so existing DRL algorithms can produce infeasible solutions, i.e. recommend actions that do not satisfy the power flow equations, voltage limits as well as dynamic constraints. Power systems, are safety critical systems and ensuring the aforementioned physical constraints are met is essential. Often the problem remedied by projecting the actions in the feasible set, which is suboptimum. In this paper, we propose a primal-dual approach to learn optimal constrained DRL policies for dynamic optimal power flow problems, aiming at controlling power generations and battery outputs. Case studies on the IEEE standard systems validate the superior performance of dynamically adapting the policy to the environment and the required safety levels.

Distributed energy resources (DER) and control assets on the grid provide mechanisms to ensure voltage support and power quality, and can be used as a means to maintain voltages close to nominal values. In this work, we study the security region of a system in the presence of devices with. We focus our efforts on the dynamics of devices that apply discrete changes, such as voltage regulators or capacitor banks, and DERs that change the power injections with Volt/Var/Watt functionality. The slow dynamics of these devices, coupled through the power flow equations, are modeled as a switching dynamical system where the switching condition depends on the operating point of the system. We identify the feasible set of DER power injections that prevent the switching of devices and oscillations in the voltage profile, and formulate an optimization problem that finds a robust operating point for voltage control. This operating point is the center of the largest Euclidean ball inscribed in the security region, and the radius of the ball is the security margin of the system. We showcase our models using the IEEE test cases. Our work has applications in voltage control and distribution planning

This paper proposes novel architectures for spatio-temporal graph convolutional and recurrent neural networks whose structure is inspired by the physics of power systems. The key insight behind our design consists in deriving the so-called graph shift operator (GSO), which is the cornerstone of Graph Convolutional Neural Network (GCN) and Graph Recursive Neural Network (GRN) designs, from the power flow equations. We demonstrate the effectiveness of the proposed architectures in two applications: in forecasting the power grid state and in finding a stochastic policy for foresighted voltage control using deep reinforcement learning. Since our design can be adopted in single-phase as well as three-phase unbalanced systems, we test our architecture in both environments. For state forecasting experiments we consider the single phase IEEE 118-bus case systems; for voltage regulation, we illustrate the performance of deep reinforcement learning policy on the unbalanced three-phase IEEE 123-bus feeder system. In both cases the physics based GCN and GRN learning algorithms we propose outperform the state of the art.

In this work, we introduce Log(v) 3LPF, a linear power flow solver for unbalanced three-phase distribution systems. Log(v) 3LPF uses a logarithmic transform of the voltage phasor to linearize the AC power flow equations around the balanced case. We incorporate the modeling of ZIP loads, transformers, capacitor banks, switches and their corresponding controls and express the network equations in matrix-vector form. With scalability in mind, special attention is given to the computation of the inverse of the system admittance matrix, Ybus. We use the Sherman-Morrison-Woodbury identity for an efficient computation of the inverse of a rank-k corrected matrix and compare the performance of this method with traditional LU decomposition methods in terms of FLOPS. We showcase the solver for a variety of network sizes, ranging from tens to thousands of nodes, and compare the Log(v) 3LPF with commercial-grade software, such as OpenDSS.

Volt-VAR and Volt-Watt functionality in photovoltaic (PV) smart inverters provide mechanisms to ensure system voltage magnitudes and power factors remain within acceptable limits.However, these control functions can become unstable, introducing oscillations in system voltages when not appropriately configured or maliciously altered during a cyberattack.In the event that Volt-VAR and Volt-Watt control functions in a portion of PV smart inverters in a distribution grid are unstable, the proposed adaptation scheme utilizes the remaining and stablybehaving PV smart inverters and other Distributed Energy Resources to mitigate the effect of the instability.The adaptation mechanism is entirely decentralized, model-free, communicationfree, and requires virtually no external configuration.We provide a derivation of the adaptive control approach and validate the algorithm in experiments on the IEEE 37 and 8500 node test feeders.

The integration of converter-interfaced generation into our power systems is changing how we control and operate these networks. While these fast-acting resources are more controllable than conventional synchronous machines, this additional controllability presents some challenges. One of these challenges is the increased cyber-physical attack surface arising from interactions among the numerous digital control loops of these devices. In this work, we present a supervisory adaptive controller that temporarily increases the outer-loop controller bandwidth of these devices in the event of sustained oscillatory behavior. We design this controller to inherently remain inactive during normal operation and only become active during sustained abnormal operating conditions. We show how this proposed controller can mitigate a cyber-physical attack, even when the attacker has full knowledge of the network model and access to real-time state information for state-feedback control.

This paper introduces a comprehensive control model that integrates aggregate electric vehicle (EV) charging demand with power grid systems operations, capitalizing on the flexible nature of EV charging. This innovative approach allows us to model and manage electrical loads in a scalable manner. The main contribution is the study of a constrained reinforcement learning (CRL) method for the predictive control of optimal power flow, paired with EV charging control. The CRL-based control method operates with the understanding that future EV arrivals are uncertain, while ensuring the feasibility of control actions. Our case studies, conducted on IEEE standard systems, highlight the superior performance of our approach that dynamically adapts to the evolving EV environment while consistently upholding safety constraints.

In this work, we propose a physics inspired Graph Convolutional Neural Network (GCN)-Reinforcement Learning (RL) architecture to train online controllers policies for the optimal selection of Distributed Energy Resources (DER) set-points. While the use of GCN is compatible with any DRL scheme, we test it in combination with the popular proximal policy optimization (PPO) algorithm and, as application, we consider the selection of set-points for Volt/Var and Volt/Watt control logic of smart inverters as the case study for DER control. We are able to show numerically that the GCN scheme is more effective than various benchmarks in regulating voltage and mitigating undesirable voltage dynamics generated by cyber-attacks. In addition to exploring the performance of GCN for a given network, we investigate the case of grids that are dynamically changing due to topology or line parameters variations. We test the robustness of GCN-RL policies against small perturbations and evaluate the scheme so called "transfer learning" capabilities.

Volt-VAR and Volt-Watt control functions are mechanisms that are included in distributed energy resource (DER) power electronic inverters to mitigate excessively high or low voltages in distribution systems. In the event that a subset of DER have had their Volt-VAR and Volt-Watt settings compromised as part of a cyber-attack, we propose a mechanism to control the remaining set of non-compromised DER to ameliorate large oscillations in system voltages and large voltage imbalances in real time. To do so, we construct control policies for individual non-compromised DER, directly searching the policy space using an Adam-based augmented random search (ARS). In this paper we show that, compared to previous efforts aimed at training policies for DER cybersecurity using deep reinforcement learning (DRL), the proposed approach is able to learn optimal (and sometimes linear) policies an order of magnitude faster than conventional DRL techniques (e.g., Proximal Policy Optimization).

Electronic waste carries energetic costs and an environmental burden rivaling that of plastic waste due to the rarity and toxicity of the heavy-metal components. Recyclable conductive composites are introduced for printed circuits formulated with polycaprolactone (PCL), conductive fillers, and enzyme/protectant nanoclusters. Circuits can be printed with flexibility (breaking strain ≈80%) and conductivity (≈2.1 × 104 S m-1 ). These composites are degraded at the end of life by immersion in warm water with programmable latency. Approximately 94% of the functional fillers can be recycled and reused with similar device performance. The printed circuits remain functional and degradable after shelf storage for at least 7 months at room temperature and one month of continuous operation under electrical voltage. The present studies provide composite design toward recyclable and easily disposable printed electronics for applications such as wearable electronics, biosensors, and soft robotics.