Integrated Intelligent Energy

False data injection attacks detection model based on LightGBM for household load data

WANG Jin, ZHANG Xiaoyu

2024, 46(11): 1-9. doi:10.3969/j.issn.2097-0706.2024.11.001

Asbtract ( 103 )

HTML ( 8)

PDF (994KB) ( 154 )

Figures and Tables | References | Related Articles | Metrics

With the continuous development of information technology， smart meters have been widely deployed in many households， allowing power companies to better identify the socio-demographic characteristics of power consumers and provide diversified services. However， one of the threats faced by smart grids is energy theft， particularly through False Data Injection Attacks （FDIA）， which tamper with meter data for covert theft， posing a serious threat to the safe and stable operation of power systems. To address this issue， an FDIA detection model based on LightGBM was proposed. Normal user electricity consumption data was selected， and different types of FDIA were implemented on some users. Features were extracted using a sliding window method， and the LightGBM model was employed for multi-class detection. Experimental results showed that this model excelled in detection accuracy and real-time performance， accurately identifying different types of FDIA with quick and efficient detection， meeting the real-time requirements of practical applications. This model could help ensure the safe operation of power systems.

Photovoltaic power forecasting model based on probabilistic TCN-Transformer

SHENG Ruixiang, ZHANG Xiaoyu

2024, 46(11): 10-18. doi:10.3969/j.issn.2097-0706.2024.11.002

Asbtract ( 174 )

HTML ( 5)

PDF (1450KB) ( 346 )

Figures and Tables | References | Related Articles | Metrics

A short-term PV power prediction method based on a temporal convolutional network （TCN） and a Transformer structure is proposed. Firstly， the main factors affecting PV power generation，such as wind speed， rainfall， light intensity and cloudiness， are analysed. Then， TCN is used to extract the global spatial features of the sequence， and Transformer is used to extract the temporal features of long-term dependencies in the sequence， so that a TCN-Transformer composite model with a high prediction precision is applied to PV power deterministic and probabilistic prediction. Simulation analyses are performed on the dataset from DKASC（Australia）， and the results show that the improved TCN-Transformer model exhibits excellent prediction performance under different weather conditions， improving the short-term prediction accuracy on PV power.

An information extraction method for electric power accidents based on BERT-BiLSTM-CRF model

ZHAO Guizhong, HUANG Miaohua

2024, 46(11): 19-28. doi:10.3969/j.issn.2097-0706.2024.11.003

Asbtract ( 106 )

HTML ( 5)

PDF (1019KB) ( 156 )

Figures and Tables | References | Related Articles | Metrics

Investigating patterns in electric power accidents and establishing a safety warning model require accurate， automated information extraction from large-scale accident samples for multidimensional analysis. However， traditional methods for extracting Chinese information entity features have shown low accuracy. Therefore， based on a novel named entity recognition technique for Chinese processing and leveraging multiple machine learning and deep learning models， a BERT-BiLSTM-CRF model tailored to the power grid accident domain was proposed. High-quality word vectors were generated by a pre-trained model of bidirectional encoder representations from transformers（BERT） within a transformer framework. A semantic enhancement masking strategy was employed to improve the model's understanding of the overall text structure. Then， a bidirection long short-term memory（BiLSTM） model was applied to capture contextual information， completing feature extraction. The conditional random field（CRF） model produced the optimal prediction sequence. Experimental results demonstrated the superiority of this customized model， as its accuracy， recall， and F₁ score exceeded those of three existing entity recognition models， including a general large model pre-trained using Generative pre-trained transformer（GPT） technology. These experiments validate that the proposed method achieves high accuracy and displays significant advantages in Chinese electric power accident information extraction.

Defect detection method of PV panels based on multi-scale fusion and improved YOLOv8n

ZHANG Wenqiang, LI Jiashu, XUAN Yang, LI Chen, QIAN Hang, ZHANG Xiaoyu

2024, 46(11): 29-37. doi:10.3969/j.issn.2097-0706.2024.11.004

Asbtract ( 113 )

HTML ( 3)

PDF (5277KB) ( 60 )

Figures and Tables | References | Related Articles | Metrics

Photovoltaic power generation plays a critical role in modern energy systems， and its stable operation is essential for ensuring energy supply. However， photovoltaic panels are exposed to complex environmental factors， such as ultraviolet radiation， corrosion， moisture， which can lead to defects like cracks and broken grids， thereby potentially compromising their functionality. To address the problems of high model complexity and low detection accuracy for small defects in current photovoltaic panel defect detection algorithms， a defect detection algorithm based on improved YOLOv8n was proposed. A new backbone network was designed to reduce the model's parameters and computation； The C2F-Efficient Local Attention （C2F-ELA） module was introduced to enhance the model's capability to precisely localize subtle features. Weight-Bidirectional Feature Pyramid Network （W-BiFPN） was then proposed to replace the original network structure and integrate the P2 small target detection layer. This effectively enhanced the model's ability to capture multi-scale features while leveraging shallow network information to strengthen the local feature perception， boosting the accuracy of small target recognition. Experimental results showed that， compared to the baseline YOLOv8 algorithm， this method reduced the number of parameters by 16.7% and increased the mean average precision by 3.1 percentage points， demonstrating an improvement in detection performance.

Research on transformer fault diagnosis method based on improved TCN model

XU Bo, WEI Yijun, DENG Fangming

2024, 46(11): 38-45. doi:10.3969/j.issn.2097-0706.2024.11.005

Asbtract ( 114 )

HTML ( 1)

PDF (1419KB) ( 289 )

Figures and Tables | References | Related Articles | Metrics

Transformer diagnosis methods based on traditional machine learning have been limited by low accuracy， single-source data， and a scarcity of fault samples. This paper proposes a transformer fault diagnosis model utilizing multi-source heterogeneous data fusion and the Mud Ring Algorithm （MRA） to optimize the Temporal Convolutional Network （TCN）. Oil chromatography data， infrared high-voltage bushing detection images， ultrasonic discharge detection images， and ultra-high frequency partial discharge detection images were selected as input information for the transformer fault diagnosis model. The Informer network and ResNet （Residual Network） were applied to extract and learn features from different data types， followed by feature fusion of multi-type data. The MRA algorithm was used to optimize the parameters of the TCN network， and the integrated results were used for fault classification. Experimental results showed that the proposed method achieved a Nash efficiency coefficient of 0.82 and an accuracy of 94.83%， with faster convergence， demonstrating its effectiveness in enhancing transformer fault diagnosis performance.

Path planning strategy of UAV inspection of large-scale photovoltaic power stations

WU Zhangyu, WU Chili, YU Huiming, ZHENG Xingnan, ZHANG Xiaoyu

2024, 46(11): 46-53. doi:10.3969/j.issn.2097-0706.2024.11.006

Asbtract ( 139 )

HTML ( 6)

PDF (1328KB) ( 61 )

Figures and Tables | References | Related Articles | Metrics

With the development of the photovoltaic industry， daily operation and maintenance costs for large-scale photovoltaic power stations， which mainly rely on manual inspections， are increasing. The widespread application of unmanned aerial vehicle（UAV）inspection technology effectively reduces inspection costs and improves inspection efficiency. To address the inspection challenges of large-scale photovoltaic power stations， a UAV path planning method based on clustering algorithm and ant colony algorithm was proposed. Based on the regular layout of photovoltaic clusters in large-scale photovoltaic power plants， and considering the field of view and flying altitude of drones， photographic points were planned to cover all clusters， and a scatter plot of these points was constructed. Considering the endurance of UAV， a condensed K-means clustering algorithm was proposed to divide the large-scale photovoltaic power station into multiple inspection sub-regions. For multiple inspection sub-regions， a coverage-based elite ant colony algorithm was proposed for path planning， ensuring that all photovoltaic strings within the inspection area were covered， which effectively reduced the inspection path length and energy consumption of UAV. The effectiveness of the proposed algorithm was then validated in a real-world photovoltaic power station （30 MWp） scenario.

Application and prospect of federated learning in new power systems

LYU Yongsheng, ZHANG Xiaoyu, WANG Xirong, GUO Peiqian

2024, 46(11): 54-64. doi:10.3969/j.issn.2097-0706.2024.11.007

Asbtract ( 207 )

HTML ( 6)

PDF (1033KB) ( 363 )

Figures and Tables | References | Related Articles | Metrics

New power systems aiming to make clean， low-carbon， safe， flexible and efficient power supply is a key measure to achive the "dual carbon" target. However， with the widespread access of renewable energy， the fusion of artificial intelligence technologies and the rapid development of smart microgrids and distributed energy sources， such as electric vehicles， traditional centralized data processing methods fall short in protecting data privacy and enabling intelligent management. Federated learning （FL）， an innovative distributed machine learning technology， offers an effective efficiency optimization solution for new power systems due to its data privacy protection capability and intelligence. In reviews on the applications of FL in new power systems， basic principles and main algorithms of FL are expounded， practical cases of FL applied in load forecasting， anomaly detection， distributed power control and energy management under data privacy protection are analysed. Then the current technical challenges encounter by FL are also discussed. Finally， the prospects of FL in new power systems are made.

Energy storage capacity optimization of wind-PV-energy storage systems for buildings considering battery life loss

FAN Pengcheng, ZHANG Yifan, YIN Wenqian, SHI Jiahao, YE Jilei

2024, 46(11): 65-72. doi:10.3969/j.issn.2097-0706.2024.11.008

Asbtract ( 91 )

HTML ( 2)

PDF (1023KB) ( 63 )

Figures and Tables | References | Related Articles | Metrics

Driven by the "dual carbon" goal， the development of smart buildings incorporating wind power，PV power， and other new energy sources alongside energy storage has become an important approach to promoting the consumption of distributed renewable energy. Reasonable configuration of energy storage can be decided based on the complementarity of wind power，PV power and load， and net load of buildings can be regulated by"charing the battery during off-peak hours and discharging the battery during peak hours"， thereby enhancing the operational economy of buildings equipped with wind‒PV‒storage systems. A capacity optimization model for the energy storage device in smart buildings with wind‒PV‒energy storage systems was proposed， considering the battery life loss. Focusing on the impact of depth of discharge and cycle time on battery service life， a daily loss cost model for the energy storage device is established. To minimize total investment and operation and maintenance costs， an capacity optimization model for the battery of the building with wind‒PV‒energy storage systems was established， meeting various constraints for operation on typical days. To address the nonlinear terms in life loss model solving， piecewise linearization and binary expansion methods were successively employed， converting the model into a mixed-integer linear programming problem. A case study was conducted on a commercial building in Jiangsu to validate the effectiveness of this storage capacity optimization model considering life loss.

Optimal capacity configuration of off-grid wind-solar hybrid hydrogen production and green ammonia synthesis system

WANG Haiming, ZHANG Runzhi, ZHOU Jiahui, XU Gang, GENG Jiaxi

2024, 46(11): 73-82. doi:10.3969/j.issn.2097-0706.2024.11.009

Asbtract ( 144 )

HTML ( 4)

PDF (1748KB) ( 55 )

Figures and Tables | References | Related Articles | Metrics

To address the significant fluctuations and storage and transportation challenges associated with renewable energy， an off-grid wind-solar hybrid hydrogen production and green ammonia synthesis system was proposed. The levelized cost of ammonia（LCOA） between the wind-solar hybrid system and standalone wind and solar energy systems was compared， and sensitivity analysis on the green ammonia cost of the system was performed. Based on actual wind-solar output data， the system tightly coupled capacity design with operation scheduling. Hourly scheduling optimization was performed under constraints of equipment operation status and capacity configuration， aiming to maximize the system's economic returns through a mixed-integer linear programming （MILP） algorithm. The hybrid strategy effectively reduced renewable energy volatility. Energy storage devices smoothed out hydrogen and electricity fluctuations， ensuring continuity and stability of energy supply. The research results provide a feasible basis for implementing wind-solar hydrogen-ammonia integrated projects in remote and weak-grid areas or regions where grid connection is challenging， and serve as a reference for the construction of future demonstration projects.

Reactive power coordination control strategy for sending-end hybrid cascaded HVDC transmission system with high proportion of wind power integration

SHI Shengyao, JIANG Minglei, ZHANG Heyi, MA Kerui, WANG Ziqiang, MA Zhiqiang

2024, 46(11): 83-91. doi:10.3969/j.issn.2097-0706.2024.11.010

Asbtract ( 102 )

HTML ( 2)

PDF (1292KB) ( 22 )

Figures and Tables | References | Related Articles | Metrics

Compared with traditional line commutated converter-based high voltage direct current （LCC-HVDC） transmission systems， the sending-end hybrid cascaded transmission system offers higher flexibility and strong adaptability to weak grids. However， as the proportion of renewable energy increases and the output from renewable energy bases becomes more unpredictable， the sending-end system faces challenges with poor voltage quality. Using static voltage sensitivity coefficients， the impact of the sending-end hybrid cascaded transmission system on the static voltage stability of the sending-end AC bus was analyzed. To improve voltage quality at the sending-end AC bus under high wind power scenarios， a reactive power coordination control strategy was designed for the hybrid cascaded system. This strategy combined a reactive power supplementary control strategy based on the modular multi-level converter （MMC） with a DC voltage supplementary control strategy， further stabilizing the AC bus voltage even when the reactive support capacity of the MMC was limited. A simulation model of the sending-end hybrid cascaded transmission system was built on the PSCAD/EMTDC simulation platform to validate the proposed coordination control strategy， demonstrating its effectiveness in regulating the sending-end AC bus voltage and enhancing voltage quality in the sending-end AC system.

Table of Content