Global Energy Interconnection Research Institute North America

Scenario generation is an important step in the operation and planning of power systems with high renewable penetrations. In this work, we proposed a data-driven approach for scenario generation using generative adversarial networks, which is based on two interconnected deep neural networks. Compared with existing methods based on probabilistic models that are often hard to scale or sample from, our method is data-driven, and captures renewable energy production patterns in both temporal and spatial dimensions for a large number of correlated resources. For validation, we use wind and solar times-series data from NREL integration data sets. We demonstrate that the proposed method is able to generate realistic wind and photovoltaic power profiles with full diversity of behaviors. We also illustrate how to generate scenarios based on different conditions of interest by using labeled data during training. For example, scenarios can be conditioned on weather events (e.g., high wind day, intense ramp events, or large forecasts errors) or time of the year (e.g., solar generation for a day in July). Because of the feedforward nature of the neural networks, scenarios can be generated extremely efficiently without sophisticated sampling techniques.

In this letter, a novel autonomous control framework “Grid Mind” is proposed for the secure operation of power grids based on cutting-edge artificial intelligence (AI) technologies. The proposed platform provides a data-driven, model-free and closed-loop control agent trained using deep reinforcement learning (DRL) algorithms by interacting with massive simulations and/or real environment of a power grid. The proposed agent learns from scratch to master the power grid voltage control problem purely from data. It can make autonomous voltage control (AVC) strategies to support grid operators in making effective and timely control actions, according to the current system conditions detected by real-time measurements from supervisory control and data acquisition (SCADA) or phasor measurement units (PMUs). Two state-of-the-art DRL algorithms, namely deep Q-network (DQN) and deep deterministic policy gradient (DDPG), are proposed to formulate the AVC problem with performance compared. Case studies on a realistic 200-bus test system demonstrate the effectiveness and promising performance of the proposed framework.

Hydrogen is an ideal alternative energy carrier to generate power for all of society's energy demands including grid, industrial, and transportation sectors. Among the hydrogen production methods, water electrolysis is a promising method because of its zero greenhouse gas emission and its compatibility with all types of electricity sources. Alkaline electrolyzers (AELs) and proton exchange membrane electrolyzers (PEMELs) are currently used to produce hydrogen. AELs are commercially mature and are used in a variety of industrial applications, while PEMELs are still being developed and find limited application. In comparison with AELs, PEMELs have more compact structure and can achieve higher current densities. Recently, however, an alternative technology to PEMELs, hydroxide exchange membrane electrolyzers (HEMELs), has gained considerable attention due to the possibility to use platinum group metal (PGM)-free electrocatalysts and cheaper membranes, ionomers, and construction materials and its potential to achieve performance parity with PEMELs. Here, the state-of-the-art AELs and PEMELs along with the current status of HEMELs are discussed in terms of their positive and negative aspects. Additionally discussed are electrocatalyst, membrane, and ionomer development needs for HEMELs and benchmark electrocatalysts in terms of the cost-performance tradeoff.

The complexity of modern power grids keeps increasing due to the expansion of renewable energy resources and the requirement of fast demand responses, which results in a great challenge for conventional power grid control systems. Existing autonomous control approaches for the power grid requires an accurate system model and a powerful computational platform, which is difficult to scale up for the large-scale energy system with more control options and operating conditions. Facing these challenges, this article proposes a data-driven multi-agent power grid control scheme using a deep reinforcement learning (DRL) method. Specifically, the classic autonomous voltage control (AVC) problem is taken as an example and formulated as a Markov Game with a heuristic method to partition agents. Then, a multi-agent AVC (MA-AVC) algorithm based on a multi-agent deep deterministic policy gradient (MADDPG) method that features centralized training and decentralized execution is developed to solve the AVC problem. The proposed method can learn from scratch and gradually master the system operation rules by input and output data. In order to demonstrate the effectiveness of the proposed MA-AVC algorithm, comprehensive case studies are conducted on an Illinois 200-Bus system considering load/generation changes, N-1 contingencies, and weak centralized communication environment.

The analysis of blackouts, which can inevitably lead to catastrophic damage to power grids, helps to explore the nature of complex power grids but becomes difficult using conventional methods. This brief studies the vulnerability analysis and recognition of key nodes in power grids from a complex network perspective. Based on the ac power flow model and the network topology weighted with admittance, the cascading failure model is established first. The node electrical centrality is further pointed out, using complex network centrality theory, to identify the key nodes in power grids. To effectively analyze the behavior and verify the correctness of node electrical centrality, the net-ability and vulnerability index are introduced to describe the transfer ability and performance under normal operation and assess the vulnerability of the power system under cascading failures, respectively. Simulation results of IEEE 30-bus and IEEE 57-bus test cases show that the key nodes can be effectively identified with high electrical centrality, the resultant cascading failures that eventually lead to a severe decrease in net-ability, verifying the correctness and effectiveness of the analysis.

Prelithiation has important applications to convert a nonlithiated cathode or anode materials into a controllably lithiated state required for developing advanced Li-ion batteries. However, most of the prelithiation reagents developed so far are highly reactive and sensitive to oxygen and moisture and therefore difficult for practical battery application. In this work, we developed a facile prelithiation strategy using lithium naphthalenide to fully prelithiate sulfur–poly(acrylonitrile) (S-PAN) composite into a Li2S-PAN cathode and to partially prelithiate nanosilicon into a LixSi anode, which leads to a new version of silicon/sulfur Li-ion battery. This LixSi/Li2S-PAN battery can demonstrate a high specific energy of 710 Wh kg–1, with a high initial Coulombic efficiency of 93.5% and a considerable cyclability. Also, this chemical prelithiation approach is mild, efficient, and widely applicable to a large range of Li-deficient electrodes, opening up new possibilities for development of low cost, environmentally benign, and high capacity Li-ion batteries.

Multi-terminal direct current (MTDC) grids provide the possibility of meshed interconnections between regional power systems and various renewable energy resources to boost supply reliability and economy. The modular multilevel converter (MMC) has become the basic building block for MTDC and DC grids due to its salient features, i.e., modularity and scalability. Therefore, the MMC-based MTDC systems should be pervasively embedded into the present power system to improve system performance. However, several technical challenges hamper their practical applications and deployment, including modeling, control, and protection of the MMC-MTDC grids. This paper presents a comprehensive investigation and reference in modeling, control, and protection of the MMC-MTDC grids. A general overview of state-of-the-art modeling techniques of the MMC along with their performance in simulation analysis for MTDC applications is provided. A review of control strategies of the MMC-MTDC grids which provide AC system support is presented. State-of-the art protection techniques of the MMCMTDC systems are also investigated. Finally, the associated research challenges and trends are highlighted.

With the rapid growth of electric vehicle (EV) ownership, attentions have been paid to the foundation of EVs, the electric vehicle supply equipment (EVSE). Different approaches of effort, among which battery swapping and fast charging are the two most well studied, have been made to solve the tradeoff problem between the battery charging speed and battery lifetime. There has been considerable debate over development strategy between charging and battery swapping. In passenger vehicles, the EV charging mode seems to dominate. But, does it mean that the battery swap mode is a dead-end? The answer should be “No”. There are use cases showing that battery swap can have great potentials for some particular uses, such as taxis and buses. This paper uses Monte Carlo simulations of vehicle behaviors to compare the service capacities and earnings of EV charging and battery swapping for both taxi and bus fleets. Stochastic models of taxis, buses, charging stations (CSs) and battery swapping systems are set up. Subsequently, service capacities of the EVSE are compared. The impact of factors on the service capacity, such as the size of the vehicle's battery, vehicle's moving speed, the power of the CS, and the price of the swapping service is investigated. Finally, possible reasons of today's less prevalence of battery swapping stations are discussed. The results of the analysis, which can be helpful to policymakers and industry investors, show that with same service capacity, an EV battery swapping station could provide significantly more financial and social benefits for the vehicle operators and EVSE service providers than that of an EV CS.

With the rapid growth of power systems measurements in terms of size and complexity, discovering statistical patterns for a large variety of real-world applications such as renewable energy prediction, demand response, energy disaggregation, and state estimation is considered a crucial challenge. In recent years, deep learning has emerged as a novel class of machine learning algorithms that represents power systems data via a large hypothesis space that leads to the state-of-the-art performance compared to most recent data-driven algorithms. This study explores the theoretical advantages of deep representation learning in power systems research. We review deep learning methodologies presented and applied in a wide range of supervised, unsupervised, and semi-supervised applications as well as reinforcement learning tasks. We discuss various settings of problems solved by discriminative deep models including stacked autoencoders and convolutional neural networks as well as generative deep architectures such as deep belief networks and variational autoencoders. The theoretical and experimental analysis of deep neural networks in this study motivates long- term research on optimizing this cutting-edge class of models to achieve significant improvements in the future power systems research.

Hard carbon (HC) is an attractive anode material for low-cost and high-energy density sodium-ion batteries (SIBs); however, its low initial Coulombic efficiency (ICE) limits its practical battery application. To overcome this problem, we reported a facile strategy to compensate the irreversible capacity loss of HC anodes simply by a chemical presodiation reaction of the HC electrode with a sodiation reagent (sodium biphenyl, Na-Bp). Benefiting from the strong sodiation ability of Na-Bp, HC anodes can be presodiated rapidly in a very short time and the presodiated HC (NaxHC) is found to have a desirable ICE of 100%. When coupled with the Na3V2(PO4)3 cathode to build a SIB full cell, the NaxHC||Na3V2(PO4)3 cell exhibits a high ICE of ∼95.0% and an elevated energy density of 218 W h kg–1, which are far superior to those of the control cell using a pristine HC anode (50% ICE and 120 W h kg–1, respectively), suggesting great advantages brought about by the chemical presodiation process. More importantly, this presodiation reaction is very mild and highly efficient and can be widely extended to a variety of Na-storage materials, offering a new route to develop high-performance Na-storage materials for practical battery applications.

Machine learning on graphs is an important and omnipresent task for a vast variety of applications including anomaly detection and dynamic network analysis. In this paper, a deep generative model is introduced to capture continuous probability densities corresponding to the nodes of an arbitrary graph. In contrast to all learning formulations in the area of discriminative pattern recognition, we propose a scalable generative optimization/algorithm theoretically proved to capture distributions at the nodes of a graph. Our model is able to generate samples from the probability densities learned at each node. This probabilistic data generation model, i.e., convolutional graph autoencoder (CGAE), is devised based on the localized first-order approximation of spectral graph convolutions, deep learning, and the variational Bayesian inference. We apply the CGAE to anew problem, the spatio-temporal probabilistic solar irradiance prediction. Multiple solar radiation measurement sites in a wide area in northern states of the U.S. are modeled as an undirected graph. Using our proposed model, the distribution of future irradiance given historical radiation observations is estimated for every site/node. Numerical results on the national solar radiation database show state-of-the-art performance for probabilistic radiation prediction on geographically distributed irradiance data in terms of reliability, sharpness, and continuous ranked probability score.

In this paper, a reinforcement-learning-based online optimal (RL-OPT) control method is proposed for the hybrid energy storage system (HESS) in ac-dc microgrids involving photovoltaic systems and diesel generators (DGs). Due to the low system inertia, conventional unregulated charging and discharging (C&D) of energy storages in microgrids may introduce disturbances that degrade the power quality and the system performance, especially in fast C&D situations. Secondary and tertiary control levels can optimize the state of charge reference of HESS; however, they are lacking the direct controllability of regulating the transient performance. Additionally, the uncertainties in practical systems greatly limit the performance of conventional model based controllers. In this study, the optimal control theory is applied to optimize the C&D profile and to suppress the disturbances caused by integrating HESS. Neural networks are devised to estimate the nonlinear dynamics of HESS based on the input/output measurements, and to learn the optimal control input for bidirectional-converter-interfaced HESS using the estimated system dynamics. Because the proposed RL-OPT method is fully decentralized, which only requires the local measurements, the plug and play capability of HESS can be easily realized. Both islanded and grid-tied modes are considered. Extensive simulations and experiments are conducted to evaluate the effectiveness of the proposed method.

Voltage source converter (VSC)‐based high‐voltage direct current (HVDC) and multi‐terminal (MT)/DC grid technologies are the new HVDC transmission technologies after ultra‐high voltage alternative current (UHVAC) and UHVDC transmission technologies which have been widely used in China. The application of the new technologies has resulted in a rapid increase in the number of schemes in construction and planning worldwide. This has been stimulated by the greater level of functionality available from the VSC technology, which makes it suitable for a wide variety of applications. These include the integration of off‐shore wind farms, embedded links within AC networks and interconnectors, especially where the AC networks are relatively ‘weak’. VSC technology has renewed interest in MT DC systems, which may ultimately lead to wide area DC grids. This study outlines the research and application on MT and DC grids in China with respect to VSC‐HVDC key technologies and DC grid key technologies based on the presentation given in the International Workshop on Next Generation Power Equipment held on 23 September 2016 in Xian, China. The briefing details of the VSC‐HVDC projects constructed and to be constructed in China are summarised in this study.

Greenhouse gas (GHG) emission and excessive fuel consumption have become a pressing issue nowadays. Particularly, CO2 emissions from transportation account for approximately one-quarter of global emissions since 2016. Electric vehicle (EV) is considered an appealing option to address the aforementioned concerns. However, with the growing EV market, issues such as insufficient charging infrastructure to support such ever-increasing demand emerge as well. Effectively forecasting the commercial EV charging demand ensures the reliability and robustness of grid utility in the short term and helps with investment planning and resource allocation for charging infrastructures in the long run. To this end, this article presents a time-series forecasting of the monthly commercial EV charging demand using a deep learning approach-Sequence to Sequence (Seq2Seq). The proposed model is validated by real-world datasets from the State of Utah and the City of Los Angeles. Two prediction targets, namely one-step ahead prediction and multi-step ahead prediction, are tested. Further, the model is benchmarked and compared against other time series and machine learning models. Experiments show that both Seq2seq and long short-term memory (LSTM) generate satisfactory prediction performance for one-step prediction. However, when performing the multi-step prediction, Seq2Seq significantly outperforms other models in terms of various performance metrics, indicating the model’s strong capability for sequential data predictions.

In this paper, a novel long-distance wireless power transfer (WPT) system using repeater coils is proposed to provide power supplies for the driver circuits in high-voltage applications, such as flexible alternative current transmission systems. Different from most of the existing wireless repeater systems where the load is only connected to the last coil and the repeater coils function solely as power relays, in the proposed system, multiple loads are powered by the repeaters. The repeater coils transfer power not only to the subsequent coils but also to the loads connected to them. Dual coil design is proposed for the repeaters with which load-independent characteristics are obtained with a suitable design of coupling coefficients. As a result, the load power can be easily adjusted without affecting each other. Load current characteristics and system efficiency have been analyzed in detail. The power transfer capability of the proposed system is illustrated for different coil quality factors and coupling coefficients. An experimental setup with 10 loads has been built to validate the effectiveness of the proposed long-distance WPT system. The maximum reachable system efficiency is about 84%.

The increasing penetration of distributed energy resources poses numerous reliability issues to the urban distribution grid. The topology estimation is a critical step to ensure the robustness of distribution grid operation. However, the bus connectivity and grid topology estimation are usually hard in distribution grids. For example, it is technically challenging and costly to monitor the bus connectivity in urban grids, e.g., underground lines. It is also inappropriate to use the radial topology assumption exclusively because the grids of metropolitan cities and regions with dense loads could be with many mesh structures. To resolve these drawbacks, we propose a data-driven topology estimation method for medium voltage (MV) and low voltage (LV) distribution grids by only utilizing the historical smart meter measurements. Particularly, a probabilistic graphical model is utilized to capture the statistical dependencies amongst bus voltages. We prove that the bus connectivity and grid topology estimation problems, in radial and mesh structures, can be formulated as a linear regression with a least absolute shrinkage regularization on grouped variables (group lasso). Simulations show highly accurate results in eight MV and LV distribution networks at different sizes and 22 topology configurations using Pacific Gas and Electric Company residential smart meter data.

The concept of peer-to-peer (P2P) trading, or transactive energy (TE), is gaining momentum as a future grid restructure. It has the potentials to utilize distributed energy resources (DERs), proactive demand side management (DSM), and the infusion in information and communication technologies (e.g., blockchain and Internet of Things (IoT)) for promoting the technical and economic efficiency of the system in its entirety. An efficient market framework is vital for the successful and sustainable implementation of such a concept. This article proposes a P2P energy trading framework enabled by blockchain. It consolidates bilateral contracts, an electronic-commerce platform, a double-auction Vickrey-Clarke-Groves (VCG) mechanism, and trading functionalities with the main grid. Through these multi-layer mechanisms, various trading preferences and attributes of electricity generation and/or consumption are accommodated. Meanwhile, the VCG mechanism eliminates any potential for market power exercise via incentivizing truthful bidding of participants. Different remedies are proposed to overcome the drawback of VCG, i.e., the lack of balanced-budget property. Accordingly, the proposed trading framework is described as multi-settlement and quasi-ideal. Case studies are conducted to analyze and evaluate the proposed trading framework and demonstrate the effectiveness of the proposed remedies in handling probable market deficiencies.

Abstract The fabrication of high‐quality perovskite film highly relies on chemical composition and the synthesis method of perovskite. So far, sequentially deposited MA 0.03 FA 0.97 Pb(I 0.97 Br 0.03 ) 3 polycrystalline film is adopted to produce high‐performance perovskite solar cells with record power conversion efficiency (PCE). Fewer grain boundaries and incorporation of inorganic cation (e.g., cesium) would further increase device performance via sequential deposition. Here, cesium chloride (CsCl) is introduced into lead iodide (PbI 2 ) precursor solution that beneficially modulates the property of PbI 2 film, leading to larger grains with cesium incorporation in the resulting perovskite film. The enlarged crystal grains originate from a slower nucleation process for CsCl‐containing PbI 2 film when reacting with formamidine iodide, confirmed by in situ confocal photoluminescence imaging. Photovoltaic devices based on CsCl‐containing PbI 2 film demonstrate a higher averaging efficiency of 21.3% than 20.3% of the devices without CsCl additives for reverse scan. More importantly, the device stability is improved by CsCl additives that retain over 90% of their initial PCE value after 4000 min tracking at maximum power point under 1‐sun illumination. This work paves a way to further improve the photovoltaic performance of mixed‐cation‐halide perovskite solar cells via a sequential deposition method.

Microgrids with distributed energy resources are being favored in various communities to lower the dependence on utility-supplied energy and cut the CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> emissions from coal-based power plants. This paper presents a co-optimization strategy for distributed energy resource planning to minimize total annualized cost at the maximal fuel savings. Furthermore, the proposed scheme aids the community microgrids in satisfying the requirements of U.S. Department of Energy (DOE) and state renewable energy mandates. The method of Lagrange multipliers is employed to maximize fuel savings by satisfying Karush-Kuhn-Tucker conditions. With the Fourier transform and particle swarm optimization, the right mix of distributed energy resources is determined to decrease the annualized cost. A case study to test the proposed scheme for a community microgrid is presented. To validate its effectiveness, an economic justification of the solution and its comparison with HOMER Pro are also illustrated.

This paper studies the optimization of both the placement and controller parameters for Battery Energy Storage Systems (BESSs) to improve power system oscillation damping. For each BESS, dynamic power output characteristics of the power converter interface are modeled considering the power limit, State of Charge limit, and time constant. Then, a black-box mixed-integer optimization problem is formulated and tackled by interfacing time-domain simulation with a mixed-integer Particle Swarm Optimization algorithm. The proposed optimization approach is demonstrated on the New England 39-bus system and a Nordic test system. The optimal results are also verified by time-domain simulation. To improve the applicability and efficiency of the proposed method, seasonal load changes and the minimum number of BESS units to be placed are considered. The proposed controller is also compared to other controllers to validate its performance.