Spatial And Temporal Distribution Of Lightning And Prediction Of Thunderstorm Based On LLS

Lightning is a strong phenomenon in the natural world.It is extremely destructive.The lightning disaster will not only cause damage to electronic equipment,power distribution network and household appliances,but also cause forest fires,plant explosions,large-scale power outages and other accidents,and even threatens the lives of citizens and brings huge economic losses to humanity as a whole and serious social impact.And it has brought enormous economic losses and serious social impacts to the entire human race.Therefore,for many industries and departments,it is of great necessity and significance to learn in advance important information such as the occurrence,development,and disappearance of lightning.Under the premise of the current research background and project requirements,this paper has conducted in-depth research on lightning location system data,combined with clustering algorithm and geographic information system,aiming at proposing a practical and feasible approach to forecasting thunderstorms.And the prediction results presented in visualized regional maps provide important help for lightning protection and disaster reduction.First of all,this paper makes an in-depth analysis of the lightning data provided by LLS(Lightning Location System),and studies the monthly distribution of the total number of lightning and the comparison of monthly distribution of positive and negative lightning.Studies have shown that more than half of the total number of lightning strikes in a decade is concentrated on very few days.Digging deeper into relevant data can provide new possibilities for lightning protection in power systems.Then using the exponential distribution model in statistics to analyze the lightning activity,and using the lognormal distribution model provided by the IEEE working group to analyze the polarity distribution characteristics of lightning current amplitude and the probability distribution of lightning current amplitude in Hong Kong..Finally,we compare the distribution of lightning and lightning in Hong Kong and can accurately reflect the laws of lightning activity in Hong Kong.In order to improve the passiveness of lightning forecasting and warning,this paper proposes a prediction method for the thunderstorm movement trend with the help of active forecasting and improved density clustering.First,the Hong Kong region is divided into grids of 0.01°x0.01o(about 1 km2),and grids with lightning strike points are defined as lightning strike areas.Then the kernel density estimation and the weighted Euclidean distance improved clustering algorithm are introduced to cluster the real-time lightning data provided by the lightning location system.Finally,the speed and direction of the thunderstorm cloud are calculated according to the changes in the temporal position of the thunderstorm centroid.The least squares quadratic curve fits the thunderstorm motion trajectory,and the Kriging interpolation method calculates the density of lightning and realizes the prediction of thunderstorms.The verification results show that the predicted thunderstorm center of mass deviation from the actual position within 3 km,the degree of deviation is less than 5%,and the prediction range of the thunderstorm cloud coverage is within 20%.The thunderstorm process can be reproduced completely and intuitively,providing important technical support for improving the accuracy of lightning warning and reducing the false alarm rate.