Search

The important transmission lines have the characteristics of large transmission capacity, long line, high tower, complex corridor terrain, which make the risk of lightning trip high. According to the demand of lightning early warning for transmission lines, the construction scheme of lightning early warning network for transmission lines is proposed. Firstly, the lightning risk level of transmission lines is evaluated, and the topological relationship between lightning movement trend and transmission line trend is considered. Combined with the operation and site selection characteristics of the early warning device, the distribution method of lightning early warning devices is proposed. A construction scheme of lightning monitoring and early warning network based on lightning strike risk assessment level of transmission lines is proposed, which provides technical support for the application of lightning early warning in power grid.

In order to promote natural disaster protection level of compact transmission corridor, a comprehensive assessment method for evaluating the overall risk of lightning strike, windage, contamination and icing is proposed in this paper. On the basis of existing risk maps, the data is standardized as the risk value for each natural disaster accordingly. Combining with historical tripping data of compact transmission corridor and expert scores for the four natural disasters, the risk weights of lightning strike, windage, contamination and icing are determined by integrating analytic hierarchy process (AHP) with entropy weight. Through multiplying the risk values and risk weights, the overall risk of the four natural disasters is calculated. Using the tripping data of one compact transmission corridor from 2013 to 2016 for risk assessment analysis, the result shows that the accuracy of this comprehensive assessment method is up to 94.4%, which can be used to improve the stability and security of compact transmission corridor.

Transmission line monitoring plays a crucial role in the safe operation of the power grid. The application of image recognition technology based on artificial intelligence algorithms can improve the efficiency of fault identification and reduce the labor cost. Object detection is to detect the object of interest in the given picture and is an effective method to recognize the fault location. This paper proposed an insulator fault identification method based on the Atlas intelligent computing platform aiming at the characteristics of small insulator fault data set. This method adopts the SSD300 model for training and inference and analysis is carried out based on the Atlas intelligent computing platform. The experiment results show that the SSD300 model can be ported well to the Atlas intelligent computing platform without reducing the recognition accuracy. At the same time, the model size is decreased.

Aiming at the problem of poor image fusion in existing image fusion and enhancement methods, a dual light image fusion model of transformer based on DCGAN is proposed. It changes the original structure which constructs a cyclic GAN to allow the generator to generate fused image. The image feature of visible and infrared light is extracts from transformer in substation through the deep discriminant convolution network. The two image features are coded in the same feature dictionary for sparse coding which is fused and used as input of generating network. The fused image of transformer is obtained by deconvolution operation. Finally, using the error of training, the model can generate a useful fusion feature. The experiments show that it reduces the hardware pressure for existing of dual light images and has a good fusion effect.

Atmospheric electric field instrument is often used for lightning warning. The lightning alert is issued by setting the threshold of electric field. Because the threshold value directly affects the accuracy of the warning, the optimal threshold can be determined by evaluating the effectiveness of lightning early warning. But the conventional evaluation method is to match the alarm information with the lightning activity information. The method does not consider the operation state and network communication of the electric field instrument itself. There is a certain deviation in the evaluation accuracy, which affecting the optimal threshold. In this paper, the effectiveness evaluation method of the conventional lightning early warning is optimized. The evaluation method based on the abnormal operation state and data reception rate of the instrument is proposed. The original evaluation results are modified. The electric field data of Jiangsu in 2017 are analyzed. The results show that the evaluation method proposed in this paper can not only evaluate the early-warning effect of the electric field instrument more accurately, but also the efficiency of the lightning early-warning algorithm based on the atmospheric electric field instrument is improved.

In order to improve emergency protection level of ultra-high voltage (UHV) transmission line when facing large scale thunderstorm, back propagation neural network (BPNN) is adopted to establish a 72-hour thunderstorm prediction model. Firstly, standard deviation of the input meteorological parameters is standardized, and statistics of the output actual lightning activities are carried out according to unified grid division. On this basis, the three-layer back BPNN is used to complete the model training. Finally, the prediction results are evaluated by using the actual lightning activities across the prediction area and the lightning tripping events of UHV transmission lines. The results show that the 72-hour thunderstorm prediction method proposed in this paper is practical in lightning protection engineering application of power grid.

Cloud-to-Ground (CG) lightning in Guangdong Province, South China has been observed over decades by the Lightning Location System (LLS) operated by power sector in China. In this paper, we present the characteristics of CG lightning based on the data collected during 2004–2017 when the LLS performance was relatively steady from year to year. To the best of our knowledge, it is the first time to investigate CG lightning from such a long period in this area. Over 30 (67) million CG lightning flashes (strokes) were recorded during this 14-year period. Various characteristics, including flash density, peak current, multiplicity, ratio of positive CG lightning, lightning day, etc, are investigated. Interesting spatiotemporal patterns of the CG lightning parameters are obtained and the potential causes are discussed.

Cloud-to-ground (CG) lightning characteristic parameters deduced from the lightning location system (LLS) for five differing microenvironmental areas relative to megacity, city, municipal town, hilly suburban area, and mountainous rural area conditions were examined in our 2009–2011 Pearl River Delta (PRD) study. Our LLS data analysis showed that there were high variation of lightning characteristics and phenomenal changes among these areas. As a supplement to the usual study of physical characteristics, an observation-based methodology had been developed to study the lightning behavior, while the respective thunderstorms were traversing through these observation areas. Special features and phenomenal changes related to the lightning characteristic parameters, such as observable lightning stroke days (OLSDs) and observable lightning stroke frequency and density for an OLSD, were also addressed. Microenvironmental variation due to change in topography, degree of urbanization, urban effect, and thunderstorm strength was found to affect the spatial distribution of lightning stroke and the severity of lightning activities over the observation areas. This approach increases our understanding of lightning in subtropical China. It also tells us more about the behavior of lightning while the thunderstorm traverses through an observation area. This information is lacking in previous studies.

In view of the shortcomings of traditional lightning protection measures in the initiative defense, it is necessary to research lightning warning technology, which can provide decision support for avoiding the lightning strike risks of infrastructures. This paper adopted atmospheric electric field and time-differencial calculation, combined with prediction of lightning moving trend based on Lightning Location System, achieved the technology of small-scale and short-time lightning warning for targeted areas. Finally, its effectiveness and practicality are verified by applying this method to transmission line in power grid.

Aiming at the shortcomings of traditional lightning protection measures in active defense for lightning accident, it is necessary to carry out studies on lightning warning method. On the basis of lightning warning information, measures can be taken to avoid lightning risk in advance. This paper makes comparative analysis between fair-weather electric field and thunderstorm-weather electric field by using Hilbert-Huang Transform, and analyses their intrinsic oscillation characteristics for lightning warning. Combining the amplitude of atmospheric electric field with its proportion of energy on high frequency band, a lightning warning method is proposed. Then typical thunderstorm-weather is utilized to verify this lightning warning method workable and accurate.

The relationship between the cloud-to-ground lightning characteristics and the air quality data over Guangzhou megacity, South China was investigated in this report. A simple method was proposed to cluster lightning location data into two types of thunderstorms: the traverse thunderstorms and the local ones. It was found that the AQI (air quality index), PM10 concentration, PM2.5 concentration, NO 2 concentration and O 3 concentration were positively correlated to the number of cloud-to-ground lightning flashes for the local thunderstorms. It was also found that the ratio of the positive cloud-to-ground flashes showed decreasing trend and the duration of the local thunderstorm showed increasing trend as the PM10 concentration increased. No obvious correlation was observed for the cloud-to-ground lightning count and the SO2 concentration.