Intelligent Bird Repellent System Design

In recent years,bird-induced transmission line emergencies have become increasingly common in the coastal hilly areas of Guangdong,and bird repellent strategies for defending against potential bird damage have attracted much attention.How to solve the problem of bird adaptation to bird repellent strategies during the prevention and control of bird-involved faults on transmission lines is important for the quality of power supply to the power grid and the economic and stable development of the power-using units.The parameters of existing bird-repelling strategies are often selected using random selection methods,and bird-repelling devices often have some effectiveness at the initial stage of placement,but gradually decrease in effectiveness over time,making it difficult to cope with the adaptation of birds.Therefore,this thesis will develop an intelligent bird repellent system for bird adaptation.In this thesis,the intelligent bird repellent system with adaptive function is studied for the bird adaptation problem in bird-related fault prevention and control,including the software and hardware components of the intelligent bird repellent system,the algorithm for selecting parameters of individual bird repellent strategy and the algorithm for selecting bird repellent strategy.The specific research work of this thesis is as follows.Research was conducted on the dominant bird species and pest-involved birds in some coastal hilly areas of Guangdong,and the software and hardware of the intelligent bird repellent system were designed based on the research results.First,the pest-involved bird species,mainly the black-collared starling and the spotted wench,are identified; then,Doppler radar is selected as the main detection method among various bird detection methods; then,based on sufficient research,acoustic waves and laser are determined as the main repellent methods,considering that the bird repellent effects of different frequency acoustic waves and different wavelength laser are different,and the conclusion that there are differences in the sensitivity of birds to different parameters of bird repellent strategies is verified.The effective frequency range of acoustic wave and the effective wavelength range of laser are determined based on the findings of existing literature combined with experimental results.subsequently,the hardware modules of the intelligent bird repellent system are selected and the circuit schematic and PCB are designed.finally,the software framework of the intelligent bird repellent system is designed.The application of reinforcement learning algorithm in the bird-involved fault scenario of transmission lines is studied for the problem of selecting the parameters of individual bird repelling strategy for bird adaptation,and the self-learning of reinforcement learning is used to cope with the bird adaptation.Firstly,the characteristics of small amount of data and unknown model in the process of bird-involved fault prevention and control are clarified,and the reinforcement learning algorithm is proposed to be applied to the search problem of optimal parameters of bird-repelling strategy; then,the improved Q-Leanring algorithm is proposed to address the problems of long blind search time and slow convergence speed in the application of reinforcement learning algorithm.Finally,by comparing with Q-Learning,SARSA and SARSA（λ）algorithms,it is demonstrated that the improved Q-Learning algorithm has less search time,faster convergence speed and more accurate search results.The application of comprehensive evaluation algorithm to multiple bird repellent strategies is studied for the problem of frequency selection of bird adaptation.Based on the improved Q-Learning algorithm for individual bird repellent strategies,the AHP-EW-TOPSIS algorithm for multiple bird repellent strategies is proposed to determine the effective ranking of each bird repellent strategy by comparing the Q-values of parameters among different bird repellent strategies,and the frequency of use is determined based on the ranking.Compared with the traditional TOPSIS method,the AHP-EW-TOPSIS method has stronger anti-interference capability and can change the frequency of each bird repellent strategy by controlling the weight proportion factor.Finally,the effectiveness of the intelligent bird repelling system is verified on a small number of towers of transmission lines.