The rapid development of various renewable energy sources such as wind, solar, and energy storage technologies, as well as the integration of demand response, combined cooling, heating and power gas turbine, etc, have greatly changed the current power and energy production and consumption. In the meantime, advanced sensors, communication and control technologies have increased the potential for the economic and reliable operation of modern power and energy systems with the integration of the fluctuating renewable energies. A large amount of data can be collected from the generation side, the grid side and the demand side. How to effectively utilize the big data of all the entities to ensure energy adequacy, system security, reliability and stability, remains a great challenge for power and energy system researchers. In the mean time, the implementation of various advanced machine learning and artificial intelligence technologies are essential to solving the planning, scheduling and control problems in the whole process of energy supply and demand. Data-driven techniques have been successfully developed and widely used to solve the classification, regression and clustering problems in the power and energy area, including generation and load forecasting, power and energy planning, economic dispatch, state estimation, situational awareness, stability evaluation, etc. Such problems are generally complicated and difficult to solve with traditional model-based techniques, such that advanced big data, machine learning and artificial intelligence techniques are required. This Research Topic welcomes Original Research and Review articles about advanced data-driven methods and applications applied to power and energy systems. Target readers include both academic researchers and power and energy industry engineers. This Research Topic also aims to provide a platform to promote state-of-the-art research methods and results in related fields. All submitted papers must be written in excellent English and contain only original works, which have not been published by or are currently under review for any other journals or conferences. Potential topics include but are not limited to the Data-Driven Techniques in the following fields: • Renewable energy generation, multi-energy load consumption, and energy price forecast • Demand response • Energy storage applications • Integrated energy system applications • Smart grid and microgrid technologies • Smart home, smart building and smart city applications • Cyber physical system security applications • Power and energy system planning, operation, protection and controls • Energy economic dispatch • Electricity markets and energy markets

Xi'an Jiaotong University

Northwestern University

Wichita State University

Views

Article Views

Article Downloads

Frontiers in Energy Research

Smart Grids

Advanced Data-Driven Methods and Applications for Smart Power and Energy Systems

With the rapid development of renewable wind and solar energy, energy storage, and energy markets, as well as with the integration of advanced measurement, communication and control technologies in modern power and energy systems, a large amount of data can be collected from the generation side, the grid side and also the demand side, which provides great inspiration for a sustainable and green future. <a href="#B1">Chu and Majumdar, (2012)</a> provided a snapshot of the current energy landscape and discussed several research and development opportunities and pathways that could lead to a prosperous, sustainable and secure energy future for the world. <a href="#B2">Gijzen, (2013)</a> advocated more “big data” at the United Nations General Assembly in New York to help secure a sustainable future. <a href="#B3">Kusiak, (2016)</a> developed a data-and-knowledge sharing platform for renewable energy, which allows wind farms to be optimized through data mining. However, how to effectively utilize the big data of various entities to ensure energy adequacy, energy efficiency and energy security remains a great challenge in the field of power and energy system research.The implementation of various advanced machine learning and artificial intelligence technologies is essential to solving the planning, operation and control problems in the whole process of energy supply and demand. Data-driven techniques have been successfully developed and widely used to address the classification, regression and clustering issues in the power and energy area. Such problems are often complex and difficult to solve with traditional model-based techniques, therefore, the advanced big data, machine learning and artificial intelligence techniques are required.This Research Topic aims to provide a platform to promote state-of-the-art research methods and results in data-driven methods and applications for smart power and energy systems, such as: Modelling and Control, Energy Forecasting, Energy Efficiency, Fault Detection, Stability Assessment, etc. It is hoped that these articles can provide valuable references for academic researchers and power industry engineers in related fields.As more and more distributed renewable energy resources being integrated into modern power and energy systems, it is important to accurately model the electromagnetic and electromechanic transient process of the entire system, so as to better understand the fluctuation characterisics of the renewables and conduct appropriate controls. In this sector, four papers dedicate to improving the modeling and control issues of the multiphase converters, renewable energy uncertainties, inverter controls and adaptive maximum power tracking, etc. And high-accuracy modelling and efficient control methods can generate more operation data for subsequent forecasting and analysis, and also ensure a more reliable and secure operation of power energy systems.<a href="https://www.frontiersin.org/articles/10.3389/fenrg.2021.829382/full">Hao et al.</a> proposed an efficient real-time simulation modeling method of multiphase converters based on high-order approximation in a micro-grid scale, in which simulation data were used to test the efficiency and accuracy of the model; it turns out that the proposed model outperforms the traditional multiphase frequency converter models. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.830833/full">Zhou et al.</a> developed a model-data hybrid-driven modeling approach based on Digit Twin technology, the hybrid-driven model can accurately describe the unmodeled and uncertain dynamic characteristics of the AC/DC power system with high-penetration of renewable energy. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.879385/full">Huang et al.</a> of National Chin-Yi University of Technology, presented an online control of smart inverter for photovoltaic power generation systems in a smart grid, they have demonstrated that the dynamic response and steady-state performance of photovoltaic power generation systems were better than the traditional perturb and observed (P&amp;O) method. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.900120/full">Yin et al.</a> provided a nonlinear control method of photovoltaic power generation LVRT based on adaptive maximum power tracking, and this method can quickly and effectively control the power output of photovoltaic (PV) cells by providing appropriate reactive power support.Energy forecasting is a generally accepted area for data-driven applications. In this sector, 4 novel renewable energy and load forecasting approaches are proposed, including PV forecasting, wind speed forecasting, multi-energy load forecasting, and data-driven demand response profiling. Accurate prediction of renewable energy and load curves can provide better information for subsequent energy opmizations so as to improve the energy utilization efficiency.<a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.903998/full">Liang and Bai</a> presented a data-driven ultra short-term PV forecasting method with sky images, in which the complete dataset was extracted from sky images, historical data, and hardware systems; and a Bi-LSTM neural network was contructed to predict the output power of PV systems with different feature data extracted. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.927260/full">Sun et al.</a> developed an interval prediction method for wind speed based on ARQEA optimized by Beta distribution and SWLSTM; the interval prediction was crucial for the economic and secure operation of wind farms, which provides a robust perspective for energy optimization rather than the common-used point prediction of wind speed. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.955851/full">Yao et al.</a> proposed a novel data-driven multi-energy load forecasting model in integrated energy systems, using a uniform manifold approximation and projection method for feature data extraction and a combined TCN-NBeats model for high-accuracy joint prediction of multi-energy loads. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.968368/full">Chen et al.</a> proposed a multi-dimension refined profiling method for demand-response adjustment capabilities, based on the load characteristic data analysis and social-survey-driven user’s subjective consciousness data mining.Intelligent energy management frames and methods can help improve energy efficiency, reduce energy consumption and save energy. In this sector, eight contributors provide insightful thoughts for different aspects, such as: energy system economic dispatch, transmission loss monitoring, transmission limit optimization, distribution system monitoring, etc.Firstly, four papers are focused on the energy system economic dispatch issues. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.931890/full">Wang et al.</a> proposed an optimal dispatching model for regional power-heat-gas interconnection based on multi-type load adjustability, according to the predicted daily load curve and the output forecast of wind turbine and PV units. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.953573/full">Ma et al.</a> also proposed a dispatch method for energy efficiency improvement of an integrated energy system, which considered multiple types of low carbon factors and demand response data. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.953602/full">Wang et al.</a> presented a configuration-dispatch dual-layer optimization model for multi-microgrid-integrated energy systems, which took into account the energy storage and the demand response data. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.843536/full">Cheng et al.</a> developed a power balance partition control method based on topology characteristics of multi-source energy storage nodes, using both standard and real-world test systems.In addition, <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.811745/full">Ding et al.</a> proposed a loss prediction of ultra-high voltage transmission lines based on EEMD–LSTM–SVR algorithm, with the active power, reactive power, line current and voltage, and regional meteorological data as the inputs. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.948595/full">Ma et al.</a> performed a data-driven optimization and calculation method of the transmission section limits and their adaptability for power systems with high penetration of renewable energy; in which, a data-driven typical demand mode extraction method based on K-means clustering was proposed according to the load forecast, renewable energy forecast and market transaction data. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.891867/full">Yuan et al.</a> studied a multi-source data processing and fusion method for power distribution systems with the internet of things technology based on edge intelligence, to solve the problems of confusing storage and insufficient fusion computing performance of multi-source heterogeneous distribution data. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.932072/full">Liang et al.</a> proposed a μPMU configuration optimization method considering multiple operation modes of distribution network, which solves the problem of existing measurement devices that cannot cover the needs of distribution network condition assessment.Besides of the energy forecasting and energy optimization, power and energy systems often face a variety of faults or disturbances. Fast and accurate fault detection is an easy and effective way to isolate the faulted components or areas to reduce the impact of the faults. In this sector, six contributors provide different data-driven approaches for the power system fault detections, including transmission line defect detection, insulator string defect detection, and distribution network fault identification, etc.Firstly, two papers focus on the system level. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.826915/full">Netsanet et al.</a> proposed a cognitive edge computing-based fault detection and location strategy for active distribution networks, in which the fault detection, location, and isolation were accomplished through a combination of virtual mode decomposition (VMD), support vector machine (SVM), and long short-term memory (LSTM)-type deep machine learning algorithms. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.967592/full">Zhang et al.</a> in Technical University of Berlin, studied the permanent fault identification for distribution network based on characteristic frequency signal injection, which was also within the distribution system level.Another four papers specially focus on the transmission line and insulator string defect detections, which are essential power system equipments. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.813945/full">Li et al.</a> proposed a pin bolt state identification using the unmanned aerial vehicle (UAV)-based transmission line photos, and the performances of multiple state-of-the-art object detect methods were compared in the context of UAV inspection photos. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.839273/full">Yin et al.</a> also investigated a classification model of point cloud along the transmission line based on group normalization, the model was verified to have a smaller model size and faster identification time without losing much accuracy. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.928162/full">Xu et al.</a> presented an end-to-end insulator string defect detection in a complex background based on a deep learning model, the authors introduced a transformer structure in the insulator defect detection area, and the final results shown that the self-attention mechanism and encoder–decoder structure had great potential in computer vision tasks. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.928164/full">Ding et al.</a> proposed a high accuracy real-time insulator string defect detection method; they used YOLOv5 as the backbone model, and developed a series of advanced techniques to improve the performance, including EIoU, AFK-MC2, and Cluster-NMS.It has been widely recognized that power system transient stability assessment can be modelled as a typical binary classification problem, using most of the popular data-driven machine learning algorithms. Fast and accurate transient stability assessment can save time for the subsequent stability control measures, so that the system can recover from severe faults and disturbances. In this sector, four oustanding papers propose different data-driven approaches for the power system transient stability assessment issues; and the other four papers mainly study the system security, electromechanical oscillation, voltage stability, and commutation failures.For the transient stability assessment issues, <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.822729/full">Lin et al.</a> built a new data-driven transient stability assessment model based on fault severity assignment; firstly, an improved cost-sensitive coefficient assignment method was proposed based on fault severity; then, the fault severity of each unstable sample was calculated accordingly; finally the correction coefficient of the loss function for the data-driven machine learning models of the unstable sample can be linearized with different fault severities. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.925126/full">Zheng et al.</a> investigated the significance of samples in deep learning-based transient stability assessment, they combined the stability scores with the significance index to provide an auxiliary criterion for the degree of stability. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.932770/full">Wang et al.</a> proposed a sparse dictionary learning model for transient stability assessment, in which the stable and unstable dictionaries were developed based on the K-SVD approach. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.885673/full">Liu et al.</a> presented a time-adaptive transient stability assessment based on the gating spatiotemporal graph neural network and gated recurrent unit, and an improved weighted cross entropy loss function with the K-nearest neighbor (KNN) idea was used to deal with the unbalanced training samples.<a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.885351/full">Yuan et al.</a> proposed a quantitative method for security situation of the power information network based on the evolutionary neural network, and network vulnerability, network reliability and network threat indicators were constructed accordingly. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.908937/full">Cai et al.</a> proposed a data-driven electromechanical oscillation mode parameter identification model, through a quasi-real-time method based on randomized-DMD-multilayer artificial neural networks. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.902861/full">Zhu et al.</a> proposed a data-driven short-term voltage stability assessment model, in which a cascaded LightGBM model was constructed to mine STVS informatization, and focal loss (FL) was embedded into both offline and online phases of the model to mitigate the loss of accuracy caused by sample imbalance and overlapping. <a href="https://www.frontiersin.org/articles/10.3389/fenrg.2022.960656/full">Wang et al.</a> proposed a spatio-temporal convolutional network based identification of voltage-coupling commutation failures in multi-infeed HVDC systems; and they extracted the crucial factors that lead to the cascading relationship of multiple HVDCs, by classifying them into time and space signals, and then used the data generated in case of faults with diverse severity to train a STCN machine learning model.In conclusion, this Research Topic provides multidisciplinary data-driven studies in the field of power and energy systems, focusing on relevant data-driven modelling and control, energy forecasting, energy optimization, fault detection and stability assessment issues. Data-driven machine learning technology has become a necessity to supplement the disadvanges of traditional model-based approaches in the power and energy system area. Nevertheless, there are still challenges in data cleaning, feature engineering, hyperparameter optimization, and model generalization, etc. Hopefully, the model-data dual-driven techinique will bring us to a more sustainable and green energy future.All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.This study was supported by the National Natural Science Foundation of China (No. 52177111).We deeply thank all the authors and reviewers who have participated in this Research Topic.The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Editorial: Advanced data-driven methods and applications for smart power and energy systems

With the urgent demand for energy revolution and consumption under China’s “30–60” dual carbon target, a configuration-scheduling dual-layer optimization model considering energy storage and demand response for the multi-microgrid–integrated energy system is proposed to improve new energy consumption and reduce carbon emissions. First, a demand response model of different users and loads in the integrated energy system is established. Second, the upper energy storage configuration model is constructed by introducing shared energy storage in the multi-microgrid–integrated energy system to improve the system’s flexibility, with the optimization goal of the maximum annual profitability of shared energy storage. A carbon trading mechanism considering the dynamic reward coefficient is designed. A low-carbon economic dispatch model of a multi-microgrid–integrated energy system is constructed based on the upper energy storage capacity, charge and discharge power, and user-side demand response with the lowest annual operating cost as the optimization goal. Finally, the effectiveness of the proposed model is verified by case studies in various scenarios. The results illustrate that the proposed model can fully use demand-side controllable resources to improve system energy utilization, effectively reduce carbon emissions, and further improve the operation economy of the multi-microgrid–integrated energy system.

With the urgent demand for energy revolution and consumption under China&#x2019;s &#x201c;30&#x2013;60&#x201d; dual carbon target, a configuration-scheduling dual-layer optimization model considering energy storage and demand response for the multi-microgrid&#x2013;integrated energy system is proposed to improve new energy consumption and reduce carbon emissions. First, a demand response model of different users and loads in the integrated energy system is established. Second, the upper energy storage configuration model is constructed by introducing shared energy storage in the multi-microgrid&#x2013;integrated energy system to improve the system&#x2019;s flexibility, with the optimization goal of the maximum annual profitability of shared energy storage. A carbon trading mechanism considering the dynamic reward coefficient is designed. A low-carbon economic dispatch model of a multi-microgrid&#x2013;integrated energy system is constructed based on the upper energy storage capacity, charge and discharge power, and user-side demand response with the lowest annual operating cost as the optimization goal. Finally, the effectiveness of the proposed model is verified by case studies in various scenarios. The results illustrate that the proposed model can fully use demand-side controllable resources to improve system energy utilization, effectively reduce carbon emissions, and further improve the operation economy of the multi-microgrid&#x2013;integrated energy system.

Configuration-dispatch dual-layer optimization of multi-microgrid–integrated energy systems considering energy storage and demand response

Adjustable resources on the demand side of power system plays a vital role to improve operational flexibility of future low-carbon power system integrated with high-penetration renewable generations. While, these demand-side resources may underperform their expected potentials, due to the lack of understanding on consumers’ refined behaviors. Facing the flexibility improvement of future power system, refined portrait structure of single user combining load characteristics and subjective behavior, is constructed with multi-dimension label system from 4 aspects, including energy consumption and load characteristics, adjustable potential, behavioral awareness and user’s nature. Aiming at supplying demand response service, several key indexes are selected and further evaluated here, via data-driven load character analysis and social-survey-driven user’s subjective consciousness mining based on comprehensive evaluation with combination weighting approach. For practical application to demand response decision making, large-scale user adjustable resource is evaluated and classified based on multivariate density-based clustering algorithm. Numerical results show the feasibility and rationality of the proposed assessment method.

Adjustable resources on the demand side of power system plays a vital role to improve operational flexibility of future low-carbon power system integrated with high-penetration renewable generations. While, these demand-side resources may underperform their expected potentials, due to the lack of understanding on consumers&#x2019; refined behaviors. Facing the flexibility improvement of future power system, refined portrait structure of single user combining load characteristics and subjective behavior, is constructed with multi-dimension label system from 4 aspects, including energy consumption and load characteristics, adjustable potential, behavioral awareness and user&#x2019;s nature. Aiming at supplying demand response service, several key indexes are selected and further evaluated here, via data-driven load character analysis and social-survey-driven user&#x2019;s subjective consciousness mining based on comprehensive evaluation with combination weighting approach. For practical application to demand response decision making, large-scale user adjustable resource is evaluated and classified based on multivariate density-based clustering algorithm. Numerical results show the feasibility and rationality of the proposed assessment method.

Demand-response oriented multi-dimension refined portrait of adjustable resources based on load and survey data fusion

This article proposes a fault detection and location strategy based on cognitive edge computing to harvest the benefits of cognitive edge computing and address the special needs of active distribution networks (ADNs). In the proposed strategy, an ADN smart gateway is used to compile data in a central repository where it will be processed and analyzed. The intermediary smart gateway includes a protection unit where the fault detection, location, and isolation are accomplished through a combination of virtual mode decomposition (VMD), support vector machine (SVM,) and long short-term memory (LSTM)–type deep machine learning tools. The local measurements of branch currents and bus voltages are processed through VMD, and the informative decomposed components are provided as inputs to the SVM-based fault detection unit and LSTM-based fault location unit. The smart digital relay passes trip commands to the respective circuit breaker/s and submits compiled data regarding the history of faults and protection actions to the upper-level units. The findings from simulation results demonstrate the effectiveness of the proposed strategy to provide fast and accurate fault detection and protection against all types of faults and locations in the ADN.

This article proposes a fault detection and location strategy based on cognitive edge computing to harvest the benefits of cognitive edge computing and address the special needs of active distribution networks (ADNs). In the proposed strategy, an ADN smart gateway is used to compile data in a central repository where it will be processed and analyzed. The intermediary smart gateway includes a protection unit where the fault detection, location, and isolation are accomplished through a combination of virtual mode decomposition (VMD), support vector machine (SVM,) and long short-term memory (LSTM)&#x2013;type deep machine learning tools. The local measurements of branch currents and bus voltages are processed through VMD, and the informative decomposed components are provided as inputs to the SVM-based fault detection unit and LSTM-based fault location unit. The smart digital relay passes trip commands to the respective circuit breaker/s and submits compiled data regarding the history of faults and protection actions to the upper-level units. The findings from simulation results demonstrate the effectiveness of the proposed strategy to provide fast and accurate fault detection and protection against all types of faults and locations in the ADN.

Cognitive Edge Computing–Based Fault Detection and Location Strategy for Active Distribution Networks

In order to identify the fault nature when phase-to-phase short circuit fault occurs in distribution lines and realize adaptive reclosing of distribution lines more accurately, a fault identification scheme using an inverter-based power electronic technology and power supply from stored energy capacitor is proposed in this paper. The corresponding device is designed which is composed of rectification, energy storage and inverters. Supplied by power from high voltage side of distribution transformers, three-phase symmetrical characteristic frequency voltage signals are injected into the de-energized network. Then the fault type can be determined according to the positive sequence resistance of the equivalent network under different fault properties. In this paper, the overall structure diagram of the system scheme is proposed, and the operation sequence of the system is analyzed. The control strategy is designed to maintain the stability of the output voltage. The feasibility and rationality of the proposed scheme are verified by theoretical analysis and simulation results.

In order to identify the fault nature when phase-to-phase short circuit fault occurs in distribution lines and realize adaptive reclosing of distribution lines more accurately, a fault identification scheme using an inverter-based power electronic technology and power supply from stored energy capacitor is proposed in this paper. The corresponding device is designed which is composed of rectification, energy storage and inverters. Supplied by power from high voltage side of distribution transformers, three-phase symmetrical characteristic frequency voltage signals are injected into the de-energized network. Then the fault type can be determined according to the positive sequence resistance of the equivalent network under different fault properties. In this paper, the overall structure diagram of the system scheme is proposed, and the operation sequence of the system is analyzed. The control strategy is designed to maintain the stability of the output voltage. The feasibility and rationality of the proposed scheme are verified by theoretical analysis and simulation results.

Permanent fault identification for distribution network based on characteristic frequency signal injection using power electronic technology

Deep learning-based transient stability assessment has achieved big success in power system analyses. However, it is still unclear how much of the data is superfluous and which samples are important for training. In this work, we introduce the latest technique from the artificial intelligence community to evaluate the significance of the samples used in deep learning model for the transient stability assessment. From empirical experiments, it is found that nearly 50% of the low-significance samples can be pruned without affecting the testing performance at the early training stages, thus saving much computational time and effort. We also observe that the samples with the fault-clearing time close to the critical clearing time often have higher significance indexes, indicating that the decision boundary learned by the deep network is highly related to the transient stability boundary. This is intuitive, but to the best of our knowledge, this work is the first to analyze the connection from sample significance aspects. In addition, we combine the stability scores with the significance index to provide an auxiliary criterion for the degree of stability, indicating the distance between a sample and the stability boundary. The ultimate goal of the study is to create a tool to generate and evaluate some benchmark datasets for the power system transient stability assessment analysis, so that various algorithms can be tested in a unified and standard platform like computer vision or natural language-processing fields.

Deep learning-based transient stability assessment has achieved big success in power system analyses. However, it is still unclear how much of the data is superfluous and which samples are important for training. In this work, we introduce the latest technique from the artificial intelligence community to evaluate the significance of the samples used in deep learning model for the transient stability assessment. From empirical experiments, it is found that nearly 50% of the low-significance samples can be pruned without affecting the testing performance at the early training stages, thus saving much computational time and effort. We also observe that the samples with the fault-clearing time close to the critical clearing time often have higher significance indexes, indicating that the decision boundary learned by the deep network is highly related to the transient stability boundary. This is intuitive, but to the best of our knowledge, this work is the first to analyze the connection from sample significance aspects. In addition, we combine the stability scores with the significance index to provide an auxiliary criterion for the degree of stability, indicating the distance between a sample and the stability boundary. The ultimate goal of the study is to create a tool to generate and evaluate some benchmark datasets for the power system transient stability assessment analysis, so that various algorithms can be tested in a unified and standard platform like computer vision or natural language-processing fields.

Evaluating the significance of samples in deep learning-based transient stability assessment

Faced with the problem of fossil energy depletion and the power industry’s low-carbon requirements, low-carbon technologies in collaboration with market mechanisms, supplemented by flexible resources, are critical to achieving the low-carbon operation of integrated energy systems (IES). This paper establishes an IES considering multiple types of low-carbon factors and demand response. Firstly, the IES is deemed to participate in the carbon trading market and introduce a ladder-type carbon trading mechanism at the low-carbon policy level. Then, at the low-carbon technology level, carbon capture power plants and power-to-gas equipment are introduced to refine the modeling of the power-to-gas process. Secondly, the integrated energy system includes pluralistic energy storage technology and demand response to increase the IES’s flexibility. Based on multiple types of low carbon factors and demand response mechanisms, the scheduling model for energy efficiency improvement is constructed with the lowest sum of the cost of wind abandonment penalty, the cost of purchasing energy, the cost of equipment operation and maintenance, and the cost of carbon trading as the optimization objective, and solved by a two-stage optimization method. Five energy efficiency indicators are presented to efficiently evaluate dispatching results: wind power consumption rate, carbon trading cost, actual carbon emissions, total cost, and load fluctuation. Finally, according to an arithmetic test system based on various operation scenarios, the proposed model may increase the IES’s comprehensive energy efficiency under the coupling effect of multiple types of low-carbon factors and demand response.

Faced with the problem of fossil energy depletion and the power industry&#x2019;s low-carbon requirements, low-carbon technologies in collaboration with market mechanisms, supplemented by flexible resources, are critical to achieving the low-carbon operation of integrated energy systems (IES). This paper establishes an IES considering multiple types of low-carbon factors and demand response. Firstly, the IES is deemed to participate in the carbon trading market and introduce a ladder-type carbon trading mechanism at the low-carbon policy level. Then, at the low-carbon technology level, carbon capture power plants and power-to-gas equipment are introduced to refine the modeling of the power-to-gas process. Secondly, the integrated energy system includes pluralistic energy storage technology and demand response to increase the IES&#x2019;s flexibility. Based on multiple types of low carbon factors and demand response mechanisms, the scheduling model for energy efficiency improvement is constructed with the lowest sum of the cost of wind abandonment penalty, the cost of purchasing energy, the cost of equipment operation and maintenance, and the cost of carbon trading as the optimization objective, and solved by a two-stage optimization method. Five energy efficiency indicators are presented to efficiently evaluate dispatching results: wind power consumption rate, carbon trading cost, actual carbon emissions, total cost, and load fluctuation. Finally, according to an arithmetic test system based on various operation scenarios, the proposed model may increase the IES&#x2019;s comprehensive energy efficiency under the coupling effect of multiple types of low-carbon factors and demand response.

Dispatch for energy efficiency improvement of an integrated energy system considering multiple types of low carbon factors and demand response

A micro-phasor measurement unit (μPMU) configuration optimization approach is proposed in this article, considering the numerous operation modes of distribution network reconfiguration. The PSO algorithm with dynamic adaptability is used to optimize the setup of μPMU and improve the accuracy of state estimation for each distribution network operation mode. The configuration nodes of various operation modes are grouped and assessed by K-means according to the shortest distance, and the weights of the evaluation indexes are calculated by the AHP-CRITIC subjective and objective combination weighting method. The node with the highest comprehensive evaluation index is selected as the configuration node. The probability of multiple operation modes is then introduced. Finally, using the IEEE 118-bus distribution system as an example, the simulation demonstrates the proposed method’s effectiveness in improving distribution network state estimate.

A micro-phasor measurement unit (&#x3bc;PMU) configuration optimization approach is proposed in this article, considering the numerous operation modes of distribution network reconfiguration. The PSO algorithm with dynamic adaptability is used to optimize the setup of &#x3bc;PMU and improve the accuracy of state estimation for each distribution network operation mode. The configuration nodes of various operation modes are grouped and assessed by K-means according to the shortest distance, and the weights of the evaluation indexes are calculated by the AHP-CRITIC subjective and objective combination weighting method. The node with the highest comprehensive evaluation index is selected as the configuration node. The probability of multiple operation modes is then introduced. Finally, using the IEEE 118-bus distribution system as an example, the simulation demonstrates the proposed method&#x2019;s effectiveness in improving distribution network state estimate.

Research on μPMU Configuration Optimization Considering Multiple Operation Modes of Distribution Network

With the increasing concern on energy crisis, the coordination of multiple energy sources and low-carbon economic operation of integrated energy system (IES) have drawn more and more attention in recent years. In IES, accurate and effective multi-energy load forecasting becomes a research hotspot, especially using the high-performance data mining and machine learning algorithms. However, due to the huge difference in energy utilization between IES and traditional energy systems, the load forecasting of IES is more difficult and complex. In fact, in IES, load forecasting is not only related to external factors such as meteorological parameters and different seasons, but the correlation between energy consumption of different types of loads also plays an important role. In order to deal with the strong coupling and high uncertainty issues in IES, a novel data-driven multi-energy load forecasting model is proposed in this paper. Firstly, a feature extraction method based on Uniform Manifold Approximation and Projection (UMAP) for multi-energy load of the IES is developed, which reduces the dimension of the complex nonlinear input data. Then, considering multi-energy coupling correlation, a combined TCN-NBeats model is proposed for the joint prediction of multi-energy loads, aiming to improve the prediction accuracy through ensemble learning. Finally, the numerical case analysis using the multi-energy consumption data of an actual campus verifies the effectiveness and accuracy of the proposed data-driven multi-energy load forecasting model.

With the increasing concern on energy crisis, the coordination of multiple energy sources and low-carbon economic operation of integrated energy system (IES) have drawn more and more attention in recent years. In IES, accurate and effective multi-energy load forecasting becomes a research hotspot, especially using the high-performance data mining and machine learning algorithms. However, due to the huge difference in energy utilization between IES and traditional energy systems, the load forecasting of IES is more difficult and complex. In fact, in IES, load forecasting is not only related to external factors such as meteorological parameters and different seasons, but the correlation between energy consumption of different types of loads also plays an important role. In order to deal with the strong coupling and high uncertainty issues in IES, a novel data-driven multi-energy load forecasting model is proposed in this paper. Firstly, a feature extraction method based on Uniform Manifold Approximation and Projection (UMAP) for multi-energy load of the IES is developed, which reduces the dimension of the complex nonlinear input data. Then, considering multi-energy coupling correlation, a combined TCN-NBeats model is proposed for the joint prediction of multi-energy loads, aiming to improve the prediction accuracy through ensemble learning. Finally, the numerical case analysis using the multi-energy consumption data of an actual campus verifies the effectiveness and accuracy of the proposed data-driven multi-energy load forecasting model.

A novel data-driven multi-energy load forecasting model

The main purpose of this study is to engage in research on a grid-connected photovoltaic (PV) power generation system smart inverter. The research content includes a smart maximum power point tracking (MPPT) controller and an inverter with power regulation. First, use the PSIM software package to establish the simulation environment of the grid-connected photovoltaic power generation system and use the Sanyo HIP-186BA19 photovoltaic module to form a 744 W system for simulation. In order to enable the photovoltaic module array (PVMA) to output the maximum power under different solar insolation and ambient temperature, the architecture is based on the extension theory-based smart MPPT method to improve the dynamic response and steady-state performance of photovoltaic power generation systems compared to perturb and observed (P&O) MPPT. When the sunshine is 1,000 W/m2, the photovoltaic power generation system adopts the extension theory-based maximum power tracking method. The time required to track the maximum power point is only 0.32 s, and the steady-state ripple is only 4.127 W. However, using the traditional P&O method requires 0.741 s to track the maximum power point, and the steady-state ripple reaches 18.131 W. Thus, the dynamic response speed of the maximum power tracking method proposed is 50% faster than that of the P&O method. The steady-state performance is also better compared to the P&O method. At the same time, a simple proportional-integral (PI) controller is used to regulate the DC-link voltage, output voltage, and current of the inverter to make the voltage of the grid-connected point stable at an effective value of 220 V. Then, the voltage-power control technology is added to the photovoltaic grid-connected inverter, and a simple proportional-integral controller is used to regulate the output of the smart inverter reactive power to improve the power quality of grid voltage. Finally, simulation and experimental results are used to verify the effectiveness of the regulation performance of the developed smart inverter.

The main purpose of this study is to engage in research on a grid-connected photovoltaic (PV) power generation system smart inverter. The research content includes a smart maximum power point tracking (MPPT) controller and an inverter with power regulation. First, use the PSIM software package to establish the simulation environment of the grid-connected photovoltaic power generation system and use the Sanyo HIP-186BA19 photovoltaic module to form a 744&nbsp;W system for simulation. In order to enable the photovoltaic module array (PVMA) to output the maximum power under different solar insolation and ambient temperature, the architecture is based on the extension theory-based smart MPPT method to improve the dynamic response and steady-state performance of photovoltaic power generation systems compared to perturb and observed (P&amp;O) MPPT. When the sunshine is 1,000&nbsp;W/m2, the photovoltaic power generation system adopts the extension theory-based maximum power tracking method. The time required to track the maximum power point is only 0.32&nbsp;s, and the steady-state ripple is only 4.127 W. However, using the traditional P&amp;O method requires 0.741&nbsp;s to track the maximum power point, and the steady-state ripple reaches 18.131&nbsp;W. Thus, the dynamic response speed of the maximum power tracking method proposed is 50% faster than that of the P&amp;O method. The steady-state performance is also better compared to the P&amp;O method. At the same time, a simple proportional-integral (PI) controller is used to regulate the DC-link voltage, output voltage, and current of the inverter to make the voltage of the grid-connected point stable at an effective value of 220&nbsp;V. Then, the voltage-power control technology is added to the photovoltaic grid-connected inverter, and a simple proportional-integral controller is used to regulate the output of the smart inverter reactive power to improve the power quality of grid voltage. Finally, simulation and experimental results are used to verify the effectiveness of the regulation performance of the developed smart inverter.

Online Control of Smart Inverter for Photovoltaic Power Generation Systems in a Smart Grid

Normal power line insulators ensure the safe transmission of electricity. The defects of the insulator reduce the insulation, which may lead to the failure of power transmission systems. As unmanned aerial vehicles (UAVs) have developed rapidly, it is possible for workers to take and upload aerial images of insulators. Proposing a technology to detect insulator defects with high accuracy in a short time can be of great value. The existing methods suffer from complex backgrounds so that they have to locate and extract the insulators at first. Some of them make detection relative to some specific conditions such as angle, brightness, and object scale. This study aims to make end-to-end detections using aerial images of insulators, giving the locations of insulators and defects at the same time while overcoming the disadvantages mentioned above. A DEtection TRansformer (DETR) having an encoder–decoder architecture adopts convolutional neural network (CNN) as the backbone network, applies a self-attention mechanism for computing, and utilizes object queries instead of a hand-crafted process to give the direct predictions. We modified this for insulator detection in complex aerial images. Based on the dataset we constructed, our model can get 97.97 in mean average precision when setting the threshold of intersection over union at 0.5, which is better than Cascade R-CNN and YOLOv5. The inference speed of our model can reach 25 frames per second, which is qualified for actual use. Experimental results demonstrate that our model meets the robustness and accuracy requirements for insulator defect detection.

Normal power line insulators ensure the safe transmission of electricity. The defects of the insulator reduce the insulation, which may lead to the failure of power transmission systems. As unmanned aerial vehicles (UAVs) have developed rapidly, it is possible for workers to take and upload aerial images of insulators. Proposing a technology to detect insulator defects with high accuracy in a short time can be of great value. The existing methods suffer from complex backgrounds so that they have to locate and extract the insulators at first. Some of them make detection relative to some specific conditions such as angle, brightness, and object scale. This study aims to make end-to-end detections using aerial images of insulators, giving the locations of insulators and defects at the same time while overcoming the disadvantages mentioned above. A DEtection TRansformer (DETR) having an encoder&#x2013;decoder architecture adopts convolutional neural network (CNN) as the backbone network, applies a self-attention mechanism for computing, and utilizes object queries instead of a hand-crafted process to give the direct predictions. We modified this for insulator detection in complex aerial images. Based on the dataset we constructed, our model can get 97.97 in mean average precision when setting the threshold of intersection over union at 0.5, which is better than Cascade R-CNN and YOLOv5. The inference speed of our model can reach 25 frames per second, which is qualified for actual use. Experimental results demonstrate that our model meets the robustness and accuracy requirements for insulator defect detection.

End-to-End Insulator String Defect Detection in a Complex Background Based on a Deep Learning Model

In recent years, data-driven methods have shown great potential for the practical application of short-term voltage stability (STVS) assessment. However, most existing research works overlook the problem of sample imbalance and overlap in STVS assessment. To tackle this issue, a novel self-adaptive data-driven method for real-time STVS is proposed in this study. First, min-redundancy and max-relevance (mRMR) is employed for feature selection to reduce the computational burden. Taking the key features as inputs, a cascaded LightGBM (CasLightGBM) model is constructed to mine STVS informatization. Based on the LightGBM and cascaded structure, CasLightGBM can enhance the assessment accuracy without sacrificing the assessment earliness. Then, focal loss (FL) is embedded into both offline and online phases of the CasLightGBM to mitigate the loss of accuracy caused by sample imbalance and overlapping, thus deriving a highly comprehensive and reliable classification model for real-time STVS assessment. Extensive numerical tests are conducted on the IEEE 118-bus system, and the simulation results demonstrate that the proposed method outperforms traditional algorithms and exhibits favorable robustness to measurement noise.

In recent years, data-driven methods have shown great potential for the practical application of short-term voltage stability (STVS) assessment. However, most existing research works overlook the problem of sample imbalance and overlap in STVS assessment. To tackle this issue, a novel self-adaptive data-driven method for real-time STVS is proposed in this study. First, min-redundancy and max-relevance (mRMR) is employed for feature selection to reduce the computational burden. Taking the key features as inputs, a cascaded LightGBM (CasLightGBM) model is constructed to mine STVS informatization. Based on the LightGBM and cascaded structure, CasLightGBM can enhance the assessment accuracy without sacrificing the assessment earliness. Then, focal loss (FL) is embedded into both offline and online phases of the CasLightGBM to mitigate the loss of accuracy caused by sample imbalance and overlapping, thus deriving a highly comprehensive and reliable classification model for real-time STVS assessment. Extensive numerical tests are conducted on the IEEE 118-bus system, and the simulation results demonstrate that the proposed method outperforms traditional algorithms and exhibits favorable robustness to measurement noise.

Data-Driven Short-Term Voltage Stability Assessment Considering Sample Imbalance and Overlapping

Cascading commutation failures (CFs) pose severe risks in multi-infeed high voltage direct current (HVDC) systems. Different from the single or concurrent CF, not only the time-relevance of signals but also the spatio coupling and even control correlation of HVDCs will attribute to the cascading CFs. The conventional approaches to identify them tend to fall into a dilemma due to their complicated dynamics, wide-area coupling and vague threshold of judgement. In this paper, a deep-learning method based on the data-driven idea is proposed to recognize the cascading CFs. It analyzes the crucial factors leading to the cascading relationship of multiple HVDCs, while classifying them into time and space signals. To extract the inherent correlation between HVDCs as well as the time relevance in question, a spatio-temporal convolutional network (STCN) is formulated. The data generated in case of faults with diverse severity are applied to train STCN. Finally, the proposed framework and STCN method are validated by a customized IEEE 39 bus system and a practical power grid.

Cascading commutation failures (CFs) pose severe risks in multi-infeed high voltage direct current (HVDC) systems. Different from the single or concurrent CF, not only the time-relevance of signals but also the spatio coupling and even control correlation of HVDCs will attribute to the cascading CFs. The conventional approaches to identify them tend to fall into a dilemma due to their complicated dynamics, wide-area coupling and vague threshold of judgement. In this paper, a deep-learning method based on the data-driven idea is proposed to recognize the cascading CFs. It analyzes the crucial factors leading to the cascading relationship of multiple HVDCs, while classifying them into time and space signals. To extract the inherent correlation between HVDCs as well as the time relevance in question, a spatio-temporal convolutional network (STCN) is formulated. The data generated in case of faults with diverse severity are applied to train STCN. Finally, the proposed framework and STCN method are validated by a customized IEEE 39 bus system and a practical power grid.

Spatio-Temporal Convolutional Network Based Identification of Voltage-Coupling Commutation Failures in Multi-Infeed HVDC Systems

In traditional power systems, the transmission section limits are often set to predetermined values according to the typical peak-to-valley demand modes, which may be difficult to adapt to the complex and diverse operation modes of the new power systems with high penetration of renewable energy, and it cannot consider the coupling characteristics of AC and DC transmission sections also. In order to make better use of the efficiency of the transmission sections, a mathematical model for optimizing the adaptability of the transmission section limit is firstly proposed, in which a data-driven typical demand mode extraction method based on K-means clustering is proposed according to the load forecast, renewable energy forecast and market transaction data. Then, the adaptability assessment of the section limit and the transmission requirements are carried out. Finally, the power perturbation method is used to calculate the influence factor matrix of associated transmission section limit and stability margin of its dominant fault. And the influence factors are optimized according to section limit adaptability and stability margin influence, and the best adaptable limit of the associated transmission sections are obtained. The calculation and optimization results of the actual case of the interconnection corridor between Xinjiang and the northwest main power system of China verify the effectiveness and practicability of the proposed limit optimization model, which is able to improve the flexibility and adaptability of the actual power system operation.

In traditional power systems, the transmission section limits are often set to predetermined values according to the typical peak-to-valley demand modes, which may be difficult to adapt to the complex and diverse operation modes of the new power systems with high penetration of renewable energy, and it cannot consider the coupling characteristics of AC and DC transmission sections also. In order to make better use of the efficiency of the transmission sections, a mathematical model for optimizing the adaptability of the transmission section limit is firstly proposed, in which a data-driven typical demand mode extraction method based on K-means clustering is proposed according to the load forecast, renewable energy forecast and market transaction data. Then, the adaptability assessment of the section limit and the transmission requirements are carried out. Finally, the power perturbation method is used to calculate the influence factor matrix of associated transmission section limit and stability margin of its dominant fault. And the influence factors are optimized according to section limit adaptability and stability margin influence, and the best adaptable limit of the associated transmission sections are obtained. The calculation and optimization results of the actual case of the interconnection corridor between Xinjiang and the northwest main power system of China verify the effectiveness and practicability of the proposed limit optimization model, which is able to improve the flexibility and adaptability of the actual power system operation.

Data-Driven Optimization of Transmission Section Limits and Their Adaptability for New Power Systems With High Penetration of Renewable Energy

Transient stability assessment (TSA) has always been a fundamental and challenging problem for ensuring the security and operation of power systems. With more power electronic interface resources integrated into the grid and large renewable energies, the stability of the power system is jeopardized. Therefore, TSA of the power system should be considered in advance to keep the system running stable. In recent years, with the development of artificial intelligence (AI) technologies such as artificial neural network (ANN), support vector machine (SVM), and Markov decision process, TSA has improved dramatically. In this study, a sparse dictionary learning approach is proposed to improve the precision of the classification accuracy of transient stability assessment in power systems. Case studies of TSA using multi-layer support vector machine (ML-SVM) and long short-term memory network–based recurrent neural network (LSTM-RNN) are discussed as benchmarks to validate the proposed method. The stable and unstable dictionary learnings are designed based on datasets obtained by simulating thousands of different time-domain simulation (TDS) scenarios performed on the New-England 39-bus system in the PSAT (power system analysis toolbox) toolbox. Stable and unstable dictionaries are developed based on the K-SVD approach. The testing dataset contains both stable and unstable samples which steps into the sparse coding process to obtain the indexes. Compared with the indexes, the system’s final TSA is targeted. The proposed method exhibits satisfactory classification accuracy in transient stability prediction and provides the ability to reduce false alarms both in positives and negatives of the power system.

Transient stability assessment (TSA) has always been a fundamental and challenging problem for ensuring the security and operation of power systems. With more power electronic interface resources integrated into the grid and large renewable energies, the stability of the power system is jeopardized. Therefore, TSA of the power system should be considered in advance to keep the system running stable. In recent years, with the development of artificial intelligence (AI) technologies such as artificial neural network (ANN), support vector machine (SVM), and Markov decision process, TSA has improved dramatically. In this study, a sparse dictionary learning approach is proposed to improve the precision of the classification accuracy of transient stability assessment in power systems. Case studies of TSA using multi-layer support vector machine (ML-SVM) and long short-term memory network&#x2013;based recurrent neural network (LSTM-RNN) are discussed as benchmarks to validate the proposed method. The stable and unstable dictionary learnings are designed based on datasets obtained by simulating thousands of different time-domain simulation (TDS) scenarios performed on the New-England 39-bus system in the PSAT (power system analysis toolbox) toolbox. Stable and unstable dictionaries are developed based on the K-SVD approach. The testing dataset contains both stable and unstable samples which steps into the sparse coding process to obtain the indexes. Compared with the indexes, the system&#x2019;s final TSA is targeted. The proposed method exhibits satisfactory classification accuracy in transient stability prediction and provides the ability to reduce false alarms both in positives and negatives of the power system.

Sparse Dictionary Learning for Transient Stability Assessment

Insulator string is a special insulation component which plays an important role in overhead transmission lines. However, working outdoors for a long time, insulators often have defects because of various environmental and weather conditions, which affect the normal operation of transmission lines and even cause huge economic losses. Therefore, insulator defect recognition is a crucial issue. Traditional insulator defect identification relies on manual work, which is time-consuming and inefficient. Therefore, the use of artificial intelligence to detect the defect location and recognize its class has become a key research field. By improving the classical YOLOv5 (you only look once) model, this article proposes a new method to enable high accuracy and real-time detection. Our method has three advantages: 1) Efficient-IoU (EIoU) replaces intersection over union (IoU) to calculate the loss of box regression, which overcomes that the detection is sensitive to various scale insulators in aerial images. 2) Since YOLOv5 itself detects some natural scenes in the real world, some anchors setting by default are not suitable for defect detection, this article introduces Assumption-free K-MC2 (AFK-MC2) algorithm into YOLOv5 to modify the K-means algorithm to improve accuracy and speed. 3) The cluster non-maximum suppression (Cluster-NMS) algorithm is introduced to avoid missing detection of insulators because of mutual occlusion in images and improve the computation speed at the same time. The experiments’ results show that this model can improve detection accuracy compared with YOLOv5 and realize real-time detection.

Insulator string is a special insulation component which plays an important role in overhead transmission lines. However, working outdoors for a long time, insulators often have defects because of various environmental and weather conditions, which affect the normal operation of transmission lines and even cause huge economic losses. Therefore, insulator defect recognition is a crucial issue. Traditional insulator defect identification relies on manual work, which is time-consuming and inefficient. Therefore, the use of artificial intelligence to detect the defect location and recognize its class has become a key research field. By improving the classical YOLOv5 (you only look once) model, this article proposes a new method to enable high accuracy and real-time detection. Our method has three advantages: 1) Efficient-IoU (EIoU) replaces intersection over union (IoU) to calculate the loss of box regression, which overcomes that the detection is sensitive to various scale insulators in aerial images. 2) Since YOLOv5 itself detects some natural scenes in the real world, some anchors setting by default are not suitable for defect detection, this article introduces Assumption-free K-MC2 (AFK-MC2) algorithm into YOLOv5 to modify the K-means algorithm to improve accuracy and speed. 3) The cluster non-maximum suppression (Cluster-NMS) algorithm is introduced to avoid missing detection of insulators because of mutual occlusion in images and improve the computation speed at the same time. The experiments&#x2019; results show that this model can improve detection accuracy compared with YOLOv5 and realize real-time detection.

High Accuracy Real-Time Insulator String Defect Detection Method Based on Improved YOLOv5

The interval prediction of wind speed is crucial for the economic and safe operation of wind farms. To overcome the probability density function parameter optimization and long-term correlation of time series problems in an interval prediction method, a hybrid model based on the beta distribution of an allele real-coded quantum evolutionary algorithm (ARQEA) and a shared weight long short-term memory (SWLSTM) neural network is proposed for predicting the interval of short-term wind speed, which is beta–ARQEA–SWLSTM. Input variables are determined via autocorrelation functions, and the shape and position parameters in the beta distribution function are optimized by the ARQEA algorithm. An interval-divided multi-distribution function aggregation is proposed to deal with the fluctuation of wind speed series. Lastly, case studies are provided to demonstrate the effectiveness of the proposed method.

The interval prediction of wind speed is crucial for the economic and safe operation of wind farms. To overcome the probability density function parameter optimization and long-term correlation of time series problems in an interval prediction method, a hybrid model based on the beta distribution of an allele real-coded quantum evolutionary algorithm (ARQEA) and a shared weight long short-term memory (SWLSTM) neural network is proposed for predicting the interval of short-term wind speed, which is beta–ARQEA–SWLSTM. Input variables are determined via autocorrelation functions, and the shape and position parameters in the beta distribution function are optimized by the ARQEA algorithm. An interval-divided multi-distribution function aggregation is proposed to deal with the fluctuation of wind speed series. Lastly, case studies are provided to demonstrate the effectiveness of the proposed method.

Interval Prediction Method for Wind Speed Based on ARQEA Optimized by Beta Distribution and SWLSTM

As an important research direction of the energy system, the integrated energy system of a park plays an important role in the field of energy optimization and sustainable economic operation. In this study, a low-carbon optimal operation model of the integrated energy system of an industrial park is proposed, considering a controllable flexible load response. First, the typical structure of the integrated energy system of the park and the model of each subsystem are introduced; then, under the premise that flexible electrical and thermal loads can be used for adjustment of energy utilization, a complete dispatch scheme is constructed according to the energy consumption and system operation characteristics of the integrated energy system. Finally, an optimal scheduling model for the combined supply and demand of the integrated energy system is established with the aim of minimizing the total operating cost. For experimental results, the YALMIP toolbox and the CPLEX solver are used to verify the results of the study; simulation results show that the optimal scheduling of controllable loads can effectively reduce the comprehensive operating cost of a community integrated energy system.

As an important research direction of the energy system, the integrated energy system of a park plays an important role in the field of energy optimization and sustainable economic operation. In this study, a low-carbon optimal operation model of the integrated energy system of an industrial park is proposed, considering a controllable flexible load response. First, the typical structure of the integrated energy system of the park and the model of each subsystem are introduced; then, under the premise that flexible electrical and thermal loads can be used for adjustment of energy utilization, a complete dispatch scheme is constructed according to the energy consumption and system operation characteristics of the integrated energy system. Finally, an optimal scheduling model for the combined supply and demand of the integrated energy system is established with the aim of minimizing the total operating cost. For experimental results, the YALMIP toolbox and the CPLEX solver are used to verify the results of the study; simulation results show that the optimal scheduling of controllable loads can effectively reduce the comprehensive operating cost of a community integrated energy system.

Optimal Dispatching Modeling of Regional Power–Heat–Gas Interconnection Based on Multi-Type Load Adjustability

With increasing levels of renewable energy in power systems, the coordination of different types of dispatchable resources, such as coal-fired power plants, hydropower plants, energy storage systems, and electric vehicles, has become more important than before. To optimally dispatch these operating units, the quality of the forecasting results becomes increasingly important for the operation of power systems. In this study, an ultra-short forecasting method was proposed for photovoltaic (PV) systems. It provided a forecast of the power output for the following 5 min using sky images obtained photographically in real time. The brightness of the key area was an important factor in determining the output power of the PV system. The output power was calculated using several different features extracted from the sky images. The brightness and other key features were then processed by a bidirectional long short-term memory network. The accuracy of the proposed PV forecasting method improved the accuracy of the forecast for the total power system. A testbed system was established to capture sky images in real time and verify the effectiveness of the proposed method.

With increasing levels of renewable energy in power systems, the coordination of different types of dispatchable resources, such as coal-fired power plants, hydropower plants, energy storage systems, and electric vehicles, has become more important than before. To optimally dispatch these operating units, the quality of the forecasting results becomes increasingly important for the operation of power systems. In this study, an ultra-short forecasting method was proposed for photovoltaic (PV) systems. It provided a forecast of the power output for the following 5&#xa0;min using sky images obtained photographically in real time. The brightness of the key area was an important factor in determining the output power of the PV system. The output power was calculated using several different features extracted from the sky images. The brightness and other key features were then processed by a bidirectional long short-term memory network. The accuracy of the proposed PV forecasting method improved the accuracy of the forecast for the total power system. A testbed system was established to capture sky images in real time and verify the effectiveness of the proposed method.

A Data Driven Based Ultra Short PV Forecasting Method With Sky Images

With the increase in the power system scale, the identification of electromechanical oscillation mode parameters by traditional numerical methods can no longer meet the requirements of complete real-time analysis. Therefore, a method based on machine learning (multilayer artificial neural networks) is proposed to identify the electromechanical oscillation mode parameters of the power system. This method can take the monitorable variables of the WAMS as the input of the model and the key characteristic information such as frequency and damping ratio as the output. After processing the input and output data with randomized dynamic mode decomposition (randomized-DMD), their mapping relationship can be analyzed by using the multilayer neuron architecture. The simulation results of the 4-generator 2-area system and the IEEE 16-generator 5-area system show that this method can accurately calculate the key characteristic parameters of the system without considering the change in the control parameters and after the offline training of historical data, which shows higher accuracy, stronger robustness, and sensitive online tracking ability.

With the increase in the power system scale, the identification of electromechanical oscillation mode parameters by traditional numerical methods can no longer meet the requirements of complete real-time analysis. Therefore, a method based on machine learning (multilayer artificial neural networks) is proposed to identify the electromechanical oscillation mode parameters of the power system. This method can take the monitorable variables of the WAMS as the input of the model and the key characteristic information such as frequency and damping ratio as the output. After processing the input and output data with randomized dynamic mode decomposition (randomized-DMD), their mapping relationship can be analyzed by using the multilayer neuron architecture. The simulation results of the 4-generator 2-area system and the IEEE 16-generator 5-area system show that this method can accurately calculate the key characteristic parameters of the system without considering the change in the control parameters and after the offline training of historical data, which shows higher accuracy, stronger robustness, and sensitive online tracking ability.

Parameter Identification of Electromechanical Oscillation Mode in Power Systems Driven by Data: A Quasi-Real-Time Method Based on Randomized-DMD-Multilayer Artificial Neural Networks