Search

This paper presents, for the first time, a framework for Kolmogorov-Arnold Networks (KANs) in power system applications. Inspired by the recently proposed KAN architecture, this paper proposes physics-informed Kolmogorov-Arnold Networks (PIKANs), a novel KAN-based physics-informed neural network (PINN) tailored to efficiently and accurately learn dynamics within power systems. The PIKANs present a promising alternative to conventional Multi-Layer Perceptrons (MLPs) based PINNs, achieving superior accuracy in predicting power system dynamics while employing a smaller network size. Simulation results on a single-machine infinite bus system and a 4-bus 2- generator system underscore the accuracy of the PIKANs in predicting rotor angle and frequency with fewer learnable parameters than conventional PINNs. Furthermore, the simulation results demonstrate PIKANs capability to accurately identify uncertain inertia and damping coefficients. This work opens up a range of opportunities for the application of KANs in power systems, enabling efficient determination of grid dynamics and precise parameter identification., Comment: 10 pages, 12 figures

The United States has witnessed a growing prevalence of droughts in recent years, posing significant challenges to water supplies and power generation. The resulting impacts on power systems, including reduced capacity and the potential for power outages, underscore the need for accurate assessment methods to ensure the reliable operation of the nation's energy infrastructure. A critical step is to evaluate the usable capacity of a regional power system's generation fleet, which is a complex undertaking and requires precise modeling of the effects of hydrological and meteorological conditions on diverse generating technologies. This paper proposes a systematic, analytical approach for assessing the impacts of extreme summer drought events on the available capacity of hydro, thermal, and renewable energy generators. More specifically, the systematic framework provides plant-level capacity derating models for hydroelectric, once-through cooling thermoelectric, recirculating cooling thermoelectric, combustion turbine, solar PV, and wind turbine systems. Application of the proposed impact assessment framework to the 2025 generation fleet of the real-world power system in the PJM and SERC regions yields insightful results. By examining the daily usable capacity of 6,055 at-risk generators throughout the study region, we find that in the event of the recurrence of the 2007 southeastern summer drought in the near future, the usable capacity of all at-risk power plants may experience a substantial decrease compared to a typical summer, falling within the range of 71% to 81%. The sensitivity analysis reveals that the usable capacity would experience a more pronounced decline under more severe drought conditions. The findings of this study offer valuable insights, enabling stakeholders to enhance the resilience of power systems against the potential effects of extreme drought in the future., Comment: 15 pages, 16 figures

The uncertainties from distributed energy resources (DERs) bring significant challenges to the real-time operation of microgrids. In addition, due to the nonlinear constraints in the AC power flow equation and the nonlinearity of the battery storage model, etc., the optimization of the microgrid is a mixed-integer nonlinear programming (MINLP) problem. It is challenging to solve this kind of stochastic nonlinear optimization problem. To address the challenge, this paper proposes a deep reinforcement learning (DRL) based optimization strategy for the real-time operation of the microgrid. Specifically, we construct the detailed operation model for the microgrid and formulate the real-time optimization problem as a Markov Decision Process (MDP). Then, a double deep Q network (DDQN) based architecture is designed to solve the MINLP problem. The proposed approach can learn a near-optimal strategy only from the historical data. The effectiveness of the proposed algorithm is validated by the simulations on a 10-bus microgrid system and a modified IEEE 69-bus microgrid system. The numerical simulation results demonstrate that the proposed approach outperforms several existing methods., Comment: 13 pages, 14 figures. Submitted to IEEE Transactions on Systems, Man, and Cybernetics in Aug. 2019

This letter investigates a Branching Dueling Q-Network (BDQ) based online operation strategy for a microgrid with distributed battery energy storage systems (BESSs) operating under uncertainties. The developed deep reinforcement learning (DRL) based microgrid online optimization strategy can achieve a linear increase in the number of neural network outputs with the number of distributed BESSs, which overcomes the curse of dimensionality caused by the charge and discharge decisions of multiple BESSs. Numerical simulations validate the effectiveness of the proposed method., Comment: Submitted to IEEE Journal and under review, 4 pages, 3 figures

Natural disasters such as storms usually bring significant damages to distribution grids. This paper investigates the optimal routing of utility vehicles to restore outages in the distribution grid as fast as possible after a storm. First, the poststorm repair crew dispatch task with multiple utility vehicles is formulated as a sequential stochastic optimization problem. In the formulated optimization model, the belief state of the power grid is updated according to the phone calls from customers and the information collected by utility vehicles. Second, an AlphaZero[1] based utility vehicle routing (AlphaZero-UVR) approach is developed to achieve the real-time dispatching of the repair crews. The proposed AlphaZero-UVR approach combines deep neural networks with stochastic Monte-Carlo tree search (MCTS) to give a lookahead search decisions, which can learn to navigate repair crews without human guidance. Simulation results show that the proposed approach can efficiently navigate crews to repair all outages., Comment: 10 pages, 8 figures, submitted to IEEE PES Journal

The uncertainty of distributed renewable energy brings significant challenges to economic operation of microgrids. Conventional online optimization approaches require a forecast model. However, accurately forecasting the renewable power generations is still a tough task. To achieve online scheduling of a residential microgrid (RM) that does not need a forecast model to predict the future PV/wind and load power sequences, this paper investigates the usage of reinforcement learning (RL) approach to tackle this challenge. Specifically, based on the recent development of Model-Based Reinforcement Learning, MuZero, we investigate its application to the RM scheduling problem. To accommodate the characteristics of the RM scheduling application, a optimization framework that combines the modelbased RL agent with the mathematical optimization technique is designed, and long short-term memory (LSTM) units are adopted to extract features from the past renewable generation and load sequences. At each time step, the optimal decision is obtained by conducting Monte-Carlo tree search (MCTS) with a learned model and solving an optimal power flow sub-problem. In this way, this approach can sequentially make operational decisions online without relying on a forecast model. The numerical simulation results demonstrate the effectiveness of the proposed algorithm., Comment: 14 pages, 17 figures, submitted to TSG