Probability Prediction of Bird Damage Based on Multiple Linear Regression

In recent years, bird damage has become one of the major hidden dangers of power grid security. How to reduce the harm of bird activities to power equipment and promote the harmonious coexistence of human and birds has become a new social topic for power grid enterprises to carry out transmission line operation and maintenance work. Due to the wide distribution area of transmission lines, it is a waste of human and material resources to drive away birds blindly, and it needs a long time to adhere, so it is difficult to effectively prevent bird damage. Therefore, the prediction of the probability of bird damage is a powerful tool which can effectively reduce bird damage. This paper provides an effective method to predict bird damage. Firstly, the effects of topography, surrounding environment, voltage class, tower type, wire arrangement and season on bird damage are analyzed respectively. Then, multiple linear regression is used to establish the model between the occurrence probability and characteristic factors of bird damage, and correlation coefficient is used to evaluate the accuracy of the model. Experimental results show that the proposed model has high accuracy and reliability.