Power System Operation Research Articles

Distributed multi-agent reinforcement learning for multi-objective optimal dispatch of microgrids.

The distributed microgrids cooperate to accomplish economic and environmental objectives, which have a vital impact on maintaining the reliable and economic operation of power systems. Therefore a distributed multi-agent reinforcement learning (MARL) algorithm is put forward incorporating the actor-critic architecture, which learns multiple critics for subtasks and utilizes only information from neighbors to find dispatch strategy. Based on our proposed algorithm, multi-objective optimal dispatch problem of microgrids with continuous state changes and power values is dealt with. Meanwhile, the computation and communication resources requirements are greatly reduced and the privacy of each agent is protected in the process of information interaction. In addition, the convergence for the proposed algorithm is guaranteed with the adoption of linear function approximation. Simulation results validate the performance of the algorithm, demonstrating its effectiveness in achieving multi-objective optimal dispatch in microgrids.

An online estimation method of nodal inertia parameters based on steady state operation of power system

An online estimation method of nodal inertia parameters based on steady state operation of power system

Coordinated stochastic power system operation between TSO, renewable sources, energy storage and hydrogen energy equipped DSO, and HSO

Coordinated stochastic power system operation between TSO, renewable sources, energy storage and hydrogen energy equipped DSO, and HSO

Steady-State Modeling of Small Modular Reactors for Multi-Timescale Power System Operations With Temporally Coupled Sub-Models

Steady-State Modeling of Small Modular Reactors for Multi-Timescale Power System Operations With Temporally Coupled Sub-Models

Cyber-Physical Interdependence for Power System Operation and Control

Cyber-Physical Interdependence for Power System Operation and Control

Integrated model and automatically designed solver for power system restoration

Power system restoration strategies schedule the power plants to re-energize the power network and loads after a blackout. These strategies are of great significance to ensure a resilient power system. Most power system restoration models consider either certain restoration stage(s) or all stages, with each or some of the stages managed independently. They destroy the essential characteristics of the interaction among the restoration stages and lead to sub-optimal solutions. To address this issue, in this paper, we first propose an improved integrated power system restoration (IPSR) model covering the entire restoration process without decoupling and sectionalization. We then develop both mathematical and metaheuristic solvers for the proposed model. In particular, we introduce the automated solver design paradigm to the highly non-linear constrained restoration problem-solving. By the paradigm, we automatically design the metaheuristic solver that gains enhanced performance beyond handcrafted solvers without algorithmic expertise. It is suggested that the automatically designed metaheuristic solver is a promising new option in variable and complicated power system operation scenarios without algorithmic expertise. Finally, a comprehensive case study demonstrates the proposed model’s significance and the proposed solver’s efficiency. Numerical tests show that the proposed IPSR model and solver obtained over 20% lower restoration time than current restoration methods, demonstrating IPSR’s and the proposed solver’s efficiency and effectiveness.

Analisa Over Current Relay (OCR) dan Ground Fault Relay (GFR) pada Jaringan Distribusi 20 kV

In operation of the electric power system, disturbances often occur which can result in disruption of the distribution of electric power to consumers. The disturbances that often occur in the 20 kV medium voltage network distribution system are short circuit faults and overload faults. Protections used to prevent the disturbance from spreading and not disturbing healthy areas is the over current relay and ground fault relay. Based on the results of the researcht, showed that the relay coordination settings were OCR Incoming Primary Iset = 1818.65 A, Secondary Iset = 0.909 A, TMS = 0.199. In the OCR of the Primary Iset Feeder = 504 A, Secondary Iset = 1.26 A, TMS = 0.199, GFR Incoming Primary Iset = 144.19 A, Secondary Iset = 0.072 A, TMS = 0.238 and GFR of the Primary Iset Feeder = 140.19 A, Secondary Iset = 0.35 A, TMS = 0.23

The Environmental and Public Health Benefits of Offshore Wind Power Deployment in China.

With the rapid decline in the levelized cost, offshore wind power offers a new option for the clean energy transition of the power sector in China's coastal areas. Here, we develop a power system capacity expansion and operation optimization model to simulate the penetration of offshore wind power in China and quantify the associated health effects. We find that offshore wind power has great potential in mitigating the negative impacts of existing coal-fired power emissions. By deploying cost-competitive offshore wind power, it is projected that by 2050, offshore wind power would contribute 2% to 5% (306-654 TWh) of China's total electricity generation, cumulatively reducing pollutants from the electricity sector by 3% and preventing 165,000 premature deaths. Notably, with the implementation of incentive policies, the scale of offshore wind power deployment expands two to three times, yielding public health benefits that far outweigh the costs of policy implementation.

Development of an Optimized Neural Network Model using Hyperparameters Optimization for Electrical Load Prediction

Electrical load prediction has become essential to the efficient operation, control and management of modern electric power systems. Various machine learning prediction models have been developed for electrical load prediction, this includes Support Vector Regression, Fuzzy Logic and Neural Network (NN) modelling approaches. However, incorrect selection of model hyperparameters, which are parameters that affect the output of the prediction models, could result in low prediction accuracy of machine learning models. Hence, the development of an optimized NN model for electrical load prediction was presented in this study. Historical daily data of temperature, rainfall, relative humidity and windspeed for Osogbo, Nigeria was obtained from the National Aeronautics and Space Administration website; while electrical load data for the same location was collected from the Transmission Company of Nigeria. The data captured a period of five years (2017 to 2021). The NN models were developed with MATLAB R2022a software, and two hyperparameters, hidden layers and neuron counts, were optimized using the Bayesian optimization technique to enhance the quality of the models. The models were evaluated using mean absolute error (MAE), and root mean square error (RMSE). The MAE and RMSE for the non-optimized NN model were 6.5247 and 8.2725 respectively. Meanwhile, for the optimized NN model, the MAE and RMSE were 5.6571 and 7.4289 respectively. The obtained results show that the optimized NN performed better than the non-optimized NN models. Therefore, for more accurate load prediction, the method developed of this research is suggested for use by utility providers.

IISE PG&E Energy Analytics Challenge 2024: Forecasting Day-Ahead Electricity Prices

Electricity price forecasting is one of the cornerstones of modern-day power system operation. Over the last two decades, the field has seen many methodological advancements. On one end, traditional statistical methods remain widely prevalent in today’s energy industry, as they have proven consistently to “stand the test of time.” On the other end, emerging predictive techniques (e.g., machine and deep learning) have demonstrated immense potential. Yet, due to limited benchmarking studies, the magnitude of improvement realized by such emerging methods relative to their statistical counterparts is not entirely clear. In response, two technical divisions of the Institute of Industrial & Systems Engineers (IISE) partnered with the Pacific Gas and Electric Company (PG&E)—one of the largest utility companies in the United States—in order to organize an electricity price forecasting challenge in 2024. Using three years of pricing signals and exogenous information from California’s electricity market, the competition challenged teams of researchers and practitioners to design their own models and submit forecasts for day-ahead electricity prices, which were independently evaluated against a test set that has been reserved from all teams. This paper introduces the challenge, as well as an overview of the methods used by the top-performing contestants. A distilled summary of the key insights and lessons learned by the challenge organizers is then presented, together with their relevance to the topic of electricity price forecasting. This is then followed by recommendations for future similarly focused competitions. To accelerate the research and development on this important topic, all data used in the challenge have been made publicly available by the challenge organizers.

Algorithms of training programs for simulator training of ship's engine room personnel

The ship's engine room personnel (mechanics, electromechanics) ensure the technical operation and safe functioning of the ship's power systems. Professional training and retraining of marine specialists of ship's engine rooms is necessary due to the constant improvement of equipment and automation of electric power systems of modern ships. To ensure high-quality training of the ship's engine room team, simulator training is used with a set of modern simulators of ship power plants and electric power systems that meet the standards of the International Convention on Standards of Training, Certification and Watchkeeping for Seafarers. In the course of the study, block diagrams of training algorithms have been developed; program features and examples of their operation have been shown. The obtained algorithms for training on simulators make it possible to implement both the initial training of cadets and the retraining of specialists when improving their qualifications. In the course of simulator practice, the opportunity to develop competencies for working with high-voltage equipment (3–10 kV) of modern ships in accordance with the requirements of the STCW Convention and the STCW Code is of particular importance. Algorithms for training in high-voltage equipment maintenance should be introduced into the educational process to improve the level of qualification of specialists.

Analysis and Enlightenment of Power System Massive Blackout Accidents

The occurrence of massive blackouts has had a severe impact on both the social-economic fabric and the daily lives of the public. This paper analyzes massive blackout accidents worldwide to explore their primary causes, including natural disasters, equipment issues, imperfect management mechanisms, and human factors. The research indicates that natural disasters are the leading cause of massive blackouts, accounting for over 52% of cases. Equipment issues and human factors also contribute to a significant proportion. To ensure the safe and stable operation of power systems, this paper proposes targeted recommendations, including enhancing the resilience of power systems, improving management mechanisms, strengthening protections against human-induced incidents, and bolstering cybersecurity defenses. These measures aim to draw on lessons from past accidents while improving the overall risk resistance ability of power systems to better address potential future power crises.

How Will Concrete Piles for Offshore Wind Power Be Damaged Under Seawater Erosion? Insights from a Chemical-Damage Coupling Meshless Method.

Based on the background of the continuously rising global demand for clean energy, offshore wind power, as an important form of renewable energy utilization, is booming. However, the pile foundations of offshore wind turbines are subject to long-term erosion in the harsh marine environment, and the problem of corrosion damage is prominent, which seriously threatens the safe and stable operation of the wind power system. In view of this, a meshless numerical simulation method based on smoothed particle hydrodynamics (SPH) and a method for generating the concrete meso-structures are developed. Concrete pile foundation models with different aggregate contents, particle sizes, and ion concentration diffusion coefficients are established to simulate the corrosion damage processes under various conditions. The rationality of the numerical algorithm is verified by a typical example. The results show that the increase in the aggregate percentage gradually reduces the diffusion rate of chemical ions, and the early damage development also slows down. However, as time goes, the damage will still accumulate continuously; when the aggregate particle size increases, the ion diffusion becomes more difficult, the damage initiation is delayed, and the early damage is concentrated around the large aggregates. The increase in the ion diffusion coefficient significantly accelerates the ion diffusion process, promotes the earlier and faster development of damage, and significantly deepens the damage degree. The research results contribute to a deeper understanding of the corrosion damage mechanisms of pile foundations and providing important theoretical support for optimizing the durability design of pile foundations. It is of great significance for ensuring the safe operation of offshore wind power facilities, prolonging the service life, reducing maintenance costs, and promoting the sustainable development of offshore wind power.

Construction of a Fast Monitoring System for Electric Energy Equipment Status Based on Data Mining

In modern power system operation, it is crucial to achieve fast and accurate monitoring of the electrical equipment status. To achieve this fast and accurate detection, this study proposes a generative adversarial network that combines edge features to amplify and recognize infrared images of devices, aiming to improve the model’s training effect. This model extracted edge features from infrared images to eliminate background noise in infrared images to achieve the goal of improving the accurate monitoring of the status of electrical equipment. The results showed that on the balanced dataset, the recognition accuracy of the model could reach about 96%, and the recognition effect of the model was relatively stable. On imbalanced datasets, the highest model recognition accuracy was around 89%, and the model recognition accuracy fluctuated greatly. The constructed model effectively improves the accuracy of monitoring the operating status of electric energy equipment, achieving fast and accurate monitoring of this state. This study can achieve rapid monitoring of the operating status of electric energy equipment, effectively reducing the operation and maintenance costs of the power system.

Energy Storage Configuration and Benefit Evaluation Method for New Energy Power Plants Based on Game Theory

AbstractIn the context of increasing renewable energy penetration, energy storage configuration plays a critical role in mitigating output volatility, enhancing absorption rates, and ensuring the stable operation of power systems. This paper proposes a benefit evaluation method for self-built, leased, and shared energy storage modes in renewable energy power plants. First, energy storage configuration models for each mode are developed, and the actual benefits are calculated from technical, economic, environmental, and social perspectives. Then, the CRITIC method is applied to determine the weights of benefit indicators, and the TOPSIS method is used to rank the overall benefits of each mode. Simulation results validate the effectiveness of the proposed method and compare the benefits of the three modes, showing that the leased mode provides the highest overall benefit. This study provides a quantitative reference for the rational selection of energy storage modes in renewable energy projects.

Attentional Convolutional Neural Network Based on Distinction Enhancement and Information Fusion for FDIA Detection in Power Systems

A false data injection attack (FDIA) is one of the major threats to power systems, and identifying false data is critical to the stable operation of power systems. However, false data that closely resemble normal data hinder the accuracy of existing detection methods, and their performance further declines when exposed to ambient noise. To address these challenges, this paper proposes an attentional convolutional neural network based on distinction enhancement and information fusion (DEIF-ACNN) for FDIA detection. Firstly, by minimizing the loss of reconstruction and discrimination, this paper designed an autoencoder with a discriminator for normal data (NAE), which had the characteristic of producing a small loss for normal data. Secondly, the trained NAE is utilized to compute the feature correlation matrix between the original and reconstructed data to enhance the distinction between normal and false data. Finally, to enhance feature extraction and suppress ambient noise interference in detection, DEIF-ACNN incorporates a convolutional block attention module (CBAM) to emphasize key feature channels and highlight crucial regions in the feature matrix. Experimental results show that DEIF-ACNN outperforms other FDIA detection methods on IEEE-9, IEEE-14, and IEEE-118 bus power systems, achieving an accuracy of 99.22%, 99.83%, and 100.00%, respectively. In addition, the method exhibits the best robustness under different noise environments, and its accuracy is maintained at about 80%.

A blockchain consensus mechanism for real-time regulation of renewable energy power systems

With the ongoing development of renewable energy sources, information technologies and physical energy systems are further integrated, which leads to challenges in ensuring the secure and stable operation of renewable energy power systems in the face of potential cyber threats. The strengths of blockchain in cybersecurity make it a promising solution to these challenges. However, existing blockchains are not well-suited for control tasks due to their low real-time performance. Here, we present a consensus mechanism that enables real-time security control of systems, called Proof of Task. Instead of solving meaningless hash puzzles in Proof of Work, Proof of Task addresses problems closely related to the stable operation and control performance of these systems. With the proposed verification mechanism, Proof of Task significantly enhances the real-time performance of blockchain while mines its computational resources for tasks of interest. To demonstrate the effectiveness and necessity of Proof of Task, it is deployed across three renewable energy power systems. The results show that Proof of Task markedly fortifies the security and computing capability of these systems, ensuring their reliable and stable operation. This work highlights the promise of blockchain to facilitate security control and trusted computing of large-scale, complex-dynamic systems.

A multi-scale and multi-modal convolutional neural network for condition monitoring of transmission line

Abstract Monitoring the fatigue damage of transmission lines is crucial for stable power system operation. However, existing model-driven methods face challenges such as high computational complexity and reliance on expert knowledge, while data-driven methods require large amounts of abnormal state data. To address these issues, a multi-scale and multi-modal convolutional neural network (CNN) is proposed for real-time condition monitoring of transmission lines. Key steps include: firstly, empirical Fourier decomposition is used to decompose the original signals, extracting multi-scale state information at different frequency scales. Then, time-domain, frequency-domain, and time–frequency domain analyses are performed on the decomposed signals to capture multi-modal information. Based on this, a multi-modal fusion network is proposed based on a CNN to extract shallow and deep features, with a fully connected layer used for multi-modal feature fusion. Notably, the algorithm is implemented on a microprocessor for practical application. Experimental results show that the proposed model achieves a diagnostic accuracy of 93.06%, outperforming classical networks. It also surpasses models trained solely on time, frequency, or time–frequency features by 25.18%, 21.8%, and 19.3%, respectively.

Dispatch model of park-level integrated energy system with photovoltaic/thermal hydrogen generation equipment: A scenario analysis method based on improved K-means clustering

As the carbon emission problem brought by the traditional energy consumption pattern is becoming more and more prominent, it has become a common strategic choice for all countries to realize the carbon-neutral goal and to drive energy conversion to clean and renewable energy. Meanwhile, the significant uncertainty of renewable energy sources and their large-scale access to the grid poses new challenges to power system operation and dispatch. Therefore, this paper proposes a dispatch model for the park-level integrated energy system (PIES) with photovoltaic/thermal (PV/T) hydrogen generation equipments based on improved stepped carbon trading and adopts a scenario analysis method based on improved K-means clustering to cope with the uncertainties of the multi-energy loads and PV/T output. First, the PV/T model is constructed through five energy balance equations of the glass protection screen, solar panel, heat pipe, water inside the heat pipe, and heat storage tank, which is further combined with the hydrogen ion exchange membrane electrolyzer (HIEME) to analyze the energy flow conversion in the process of hydrogen production by the coupled equipment. Second, an improved stepped carbon trading price model is proposed by constructing a continuously monotonically increasing trading price function. Finally, based on the improved stepped carbon trading mechanism, the day-ahead dispatching model aims to optimize the sum of PIES operation cost and carbon trading cost, and the example simulation is carried out by the scenario analysis method based on improved K-means clustering. The simulation results show that the proposed model has certain significance in hydrogen production efficiency and carbon emission reduction.

Novel variable charging pricing strategy applied to the multi-objective planning of integrated fast and slow electric vehicle charging stations and distributed energy resources

Electric vehicle charging stations are increasingly being deployed to accommodate the rising demand for electric vehicles. However, this additional load on the distribution system can lead to significant challenges, including increased power losses and undervolage issues. Therefore, this paper proposes the optimized planning of electric vehicle charging stations, along with the integration of distributed energy resources considering the power system operator’s perspective. The approach focuses on minimizing the costs associated with public charging stations and distributed generators, as well as reducing power losses and voltage deviation of the system. Additionally, this study incorporates both fast and slow chargers to address user requirements, while promoting the use of slow charging to help reduce the peak load demand caused by the electric vehicle charging stations. A novel variable pricing strategy for slow charging is also proposed to decrease the operator’s expenses while promoting this mode of charging. The proposed methodology is validated through the integration of a 33-bus distribution test system with a 25-node traffic system. The analysis was conducted using three multi-objective optimization methods: Multi-Objective Cuckoo Search, Multi-Objective Flower Pollination Algorithm, and Non-dominated Sorting Genetic Algorithm II. From the results, it is possible to conclude that the proposed approach provides benefits not only to the system operator but also to EV users, given the reduction in their travel costs. Despite the increased charging demand, the optimization results show a reduction in voltage deviation by 11.57%–16.94% and power losses by 19.71%–24.20% compared to the original system.