Renke Huang

In demand response (DR) programs, load aggregators (LAs) coordinate DR customers to provide flexibility support for the smart grid. However, existing pricing mechanisms often struggle to simultaneously achieve incentive compatibility and operational controllability in retail-side DR. To address this challenge, this paper proposes an incentive compatible online pricing strategy for customer directrix load (CDL)-based DR of LAs. Unlike traditional price-based DR schemes, the proposed approach leverages CDL, which inherently enhances controllability by serving as both the guiding signal and the performance evaluation metric. Within the CDL framework, a DR customer response model and an LA centralized optimization model are established. Under a general discomfort function setting, the centralized response is characterized by a one-dimensional fixed-point equation, based on which a customized incentive compatible pricing strategy is derived to align individual customer responses with the LA’s desired response curve. Furthermore, the widely used quadratic discomfort model is studied as an analytically tractable special case, leading to closed-form response and pricing expressions. An online learning algorithm is then proposed to enable LAs to learn the unobservable discomfort parameters of DR customers from observable response decisions, thereby reducing information exposure and enhancing practical implementation performance. The proposed algorithm is theoretically shown to achieve logarithmic regret with respect to the operating cycle, and a problem-level regret lower bound is further established to show that the minimax cumulative regret of the underlying CDL-based online pricing problem is also logarithmic. Simulation results based on real-world data demonstrate the validity and effectiveness of the proposed method.

This paper proposes a robust transient disturbance rejection current controller (TDRCC) to improve the transient performance of doubly-fed induction generators (DFIGs) in wind turbines during grid disturbances, such as short circuits. The proposed robust TDRCC replaces the proportional-integral-derivative (PID) current controllers conventionally used in the vector control scheme of the DFIG rotor side converter (RSC). The TDRCC estimates the transient disturbances caused by grid faults or other external disturbances and provides compensation to the DFIG rotor voltage in the current control loops of the RSC to improve the robustness of the controller to disturbance. A sliding-mode current controller (SMCC) is also designed to highlight the transient performance improvement of the DFIG using the TDRCC over the state-of-the-art control methods during grid short circuit faults. Simulation studies are conducted in PSCAD/EMTDC for a 3.6-MW DFIG wind turbine with the proposed TDRCC, the conventional PI controller, and the SMCC, respectively during the most severe balanced three-phase grid short circuit fault specified by the U.S. grid code as well as an unbalance single-phase grid short circuit fault. Hardware experiments are conducted on a 200-W DFIG wind turbine emulator with the three different controllers for the same three-phase short circuit fault. Simulation and hardware experiment results show that the TDRCC reduces the peak rotor current significantly when compared with the PI and SMCC controlled DFIG system and, therefore, would help prolong the lifespan of the DFIG's power electronics and reduce maintenance costs.

The inverter-based resource (IBR) has been an important component of power generation within both transmission and distribution systems. The impacts of high IBR penetration on system transient stability have not been explored extensively due to the lack of high-fidelity power grid and inverter models. The aggregate IBR model in the transmission system is only suitable for representing bulk power system (BPS)-connected IBRs, and it cannot precisely reproduce the dynamics of distributed IBRs in the distribution system due to its oversimplified structure. In this study, we have developed a high-penetration fully-connected transmission and distribution (T&D) co-simulation platform that supports the dynamic co-simulation of a synthetic mini-WECC transmission system with multiple IEEE 8500-node distribution systems. The entire platform can accommodate 10,000+ dispersed and 40+ BPS-connected IBRs and achieve 100% power generation from IBRs. A novel iterative T&D power flow initialization technique was proposed to maintain stable transition from power flow to dynamic co-simulation mode with high penetration of IBR models. This study explores the impacts of grid-forming (GFM) and grid-following (GFL) inverter control on the dynamics of large-scale T&D network. Moreover, a co-simulation platform was developed and tested using high-performance computing (HPC) resources to facilitate parallel simulation.

This article proposes a model-based deep reinforcement learning (DRL) method to design emergency control strategies for short-term voltage stability problems in power systems. Recent advances show promising results in model-free DRL-based methods for power systems, but model-free methods suffer from poor sample efficiency and training time, both critical for making state-of-the-art DRL algorithms practically applicable. DRL-agent learns an optimal policy via a trial-and-error method while interacting with the real-world environment. And it is desirable to minimize the direct interaction of the DRL agent with the real-world power grid due to its safety-critical nature. Additionally, state-of-the-art DRL-based policies are mostly trained using a physics-based grid simulator where dynamic simulation is computationally intensive, lowering the training efficiency. We propose a novel model-based-DRL framework where a deep neural network (DNN)-based dynamic surrogate model, instead of a real-world power-grid or physics-based simulation, is utilized with the policy learning framework, making the process faster and sample efficient. However, stabilizing model-based DRL is challenging because of the complex system dynamics of large-scale power systems. We solved these issues by incorporating imitation learning to have a warm start in policy learning, reward-shaping, and multi-step surrogate loss. Finally, we achieved 97.5% sample efficiency and 87.7% training efficiency for an application to the IEEE 300-bus test system.

Accurate models of generation systems are critical for maintaining reliable and secure grid operations. In this paper, a novel and systematic approach is proposed to identify and calibrate the generator system problematic parameters using continuous wavelet transform (CWT) and advanced deep-learning technology. At the beginning, the phasor measurement unit (PMU) data are used through “event playback” to check the consistency between simulated and measured real and reactive power of the generation system. Once the results indicate a model parameter validation is required, a group of suspicious parameters will be identified as the primary problematic parameter candidates (PPCs). These primary PPCs are randomly perturbed to generate the event playback simulation data, which are used by the CWT and convolutional neural networks (CNNs) to further narrow down the primary PPCs into a smaller set of candidates. Then, the identified candidates are perturbed again to generate massive event playback simulation data for training a parameter calibration neural network. We designed a multi-output neural network structure to find the mappings between the perturbed parameters and the simulation data using both CNN and long short-term memory (LSTM) models. Finally, the well-trained and tested CNN-LSTM model is used to estimate the accurate value of the suspicious parameters with actual PMU measurements. The proposed CNN-LSTM network can accurately and reliably estimate the generation-system problematic parameters, and has better performance when compared to other machine-learning methods, such as the multilayer perceptron network and the conditional variational autoencoder method. The accuracy and effectiveness of the proposed approach have been validated through simulation and real-world data.

Static Var Compensator (SVC) has often been used in the power system industries to regulate the bus voltage and reactive power. With an increasing concern of the power system transient stability, the SVC has also been used to dynamically regulate the power flow for oscillation dampings in recent years. A guided surrogate-gradient-based evolutionary strategy (GSES) based SVC control approach is proposed in this paper to damp power system inter-area oscillations. The GSES algorithm is used to train a reinforcement learning agent to learn the best decisions that control the SVC, with an objective of quickly damping inter-area oscillations. The GSES algorithm brings many benefits such as easier, more robust and efficient training with less required hyper-parameters. Because the GSES can engage with multiple individual workers during the training process, parallel computation techniques are implemented in this research to accelerate the computational speed. To test the performance of the proposed GSES-based SVC controller, we compared this novel method with a supplementary SVC damping controller on two different test systems. Results indicate that the proposed GSES-based SVC control approach can effectively damp power system inter-area oscillations. Especially, in some case studies the proposed approach successfully saved the test system from a transient-instability-caused breaking down.

This letter discusses performance degradation of traditional supervisory control and data acquisition (SCADA)-based static state estimation (SSE) caused by violation of the static assumption. The sensitivity of nonsynchronized sampling errors against oscillations is tested through real measurement data and grid model from a subsystem of the North American power gird. The test results show that as an important reason leading to the failure of SCADA-based SSE, the non-synchronized sampling error of SCADA is sensitive to oscillations in that the error rapidly increases when oscillations occur and vanishes when the oscillations cease.

As power systems are undergoing a significant transformation with more uncertainties, less inertia and closer to operation limits, there is increasing risk of large outages. Thus, there is an imperative need to enhance grid emergency control to maintain system reliability and security. Towards this end, great progress has been made in developing deep reinforcement learning (DRL) based grid control solutions in recent years. However, existing DRL-based solutions have two main limitations: 1) they cannot handle well with a wide range of grid operation conditions, system parameters, and contingencies; 2) they generally lack the ability to fast adapt to new grid operation conditions, system parameters, and contingencies, limiting their applicability for real-world applications. In this paper, we mitigate these limitations by developing a novel deep meta-reinforcement learning (DMRL) algorithm. The DMRL combines the meta strategy optimization together with DRL, and trains policies modulated by a latent space that can quickly adapt to new scenarios. We test the developed DMRL algorithm on the IEEE 300-bus system. We demonstrate fast adaptation of the meta-trained DRL polices with latent variables to new operating conditions and scenarios using the proposed method, which achieves superior performance compared to the state-of-the-art DRL and model predictive control (MPC) methods.

As power systems are undergoing a significant transformation with more uncertainties, less inertia and closer to operation limits, there is increasing risk of large outages. Thus, there is an imperative need to enhance grid emergency control to maintain system reliability and security. Towards this end, great progress has been made in developing deep reinforcement learning (DRL) based grid control solutions in recent years. However, existing DRL-based solutions have two main limitations: 1) they cannot handle well with a wide range of grid operation conditions, system parameters, and contingencies; 2) they generally lack the ability to fast adapt to new grid operation conditions, system parameters, and contingencies, limiting their applicability for real-world applications. In this paper, we mitigate these limitations by developing a novel deep meta reinforcement learning (DMRL) algorithm. The DMRL combines the meta strategy optimization together with DRL, and trains policies modulated by a latent space that can quickly adapt to new scenarios. We test the developed DMRL algorithm on the IEEE 300-bus system. We demonstrate fast adaptation of the meta-trained DRL polices with latent variables to new operating conditions and scenarios using the proposed method and achieve superior performance compared to the state-of-the-art DRL and model predictive control (MPC) methods.