Font Size: a A A

Research On UAV Search And Rescue Algorithm For Atmospheric Nuclear Proliferation Accident

Posted on:2021-05-26Degree:MasterType:Thesis
Country:ChinaCandidate:Y S BianFull Text:PDF
GTID:2370330605480542Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of the nuclear power industry,nuclear accident emergency response methods have become an important content of emergency response research.Based on the atmospheric prediction model of nuclear accident emergencies,this paper studies the trajectory planning of drone search and rescue,and realizes the real-time situation simulation of radionuclide pollutant leakage and diffusion.This article aims at the Hongyanhe Nuclear Power Station.Based on the site topography,meteorology and other basic conditions,building a pathfinding map in the plant and using the heuristic search A* algorithm to provide a trajectory planning solution for emergency search and rescue UAVs.Based on Tensor Flow artificial intelligence recognition algorithm,carring out the target recognition of the UAV emergency search and rescue process.This article uses the WRF model to calculate the integral of the meteorological field in the study area,and uses the visualization tool NCL to generate simulation results,such as the surface weather analysis map of meteorological elements including wind speed,wind direction,temperature,and pressure.Finally,the HYSPLIT model is used to set parameters to simulate the diffusion of nuclides after an accident at the target nuclear power plant,and to analyze the diffusion situation of nuclides with the help of climate simulation.According to the buildings in the plant,a pathfinding map is constructed,and the A* algorithm is run to obtain the optimal search path.The average search time is 6.911 seconds,and the average search steps are 54 steps.The artificial intelligence recognition algorithm is run in the Tensor Flow system,and the recognition rate can exceed 76% under the condition of high picture resolution,clear character characteristics and no interference factors.Combining the results of nuclide diffusion and the weather map on the ground,it is found that the direction of nuclide diffusion is roughly the same as the wind direction on the ground.The simulated regional meteorological elements have a significant influence on the diffusion of nuclide.The A* search algorithm's pathfinding process is significantly affected by nuclide diffusion.The average search time increases by 18.33%,and the average search steps increase by 18.52%.
Keywords/Search Tags:A* algorithm, artificial intelligence recognition, WRF-HYSPLIT, weather analysis
PDF Full Text Request
Related items