Effectiveness Evaluation of Anti-bird Devices based on Random Forest Algorithm
Qingdong Zhou, Weisong Lan, Yuyao Zhou, Guhuai Mo
Abstract
In the past several decades, transmission line fault caused by bird activities has gain more and more attentions from electric manufacturers, which makes effectiveness evaluation of anti-bird devices becomes an urgent problem. Although the study of this problem will give guidance for the research and the effective use of anti-bird devices, research results concerning this problem is few due to the diversity of transmission line tower and anti-bird device, as well as the adaptability of bird. Up to now, engineers still use anti-bird devices according to personal experience without theoretical guidance, which results in high cost and low efficiency in anti-bird. Consider what mentioned above, in this paper, a method based on random forest algorithm is proposed to evaluate the effectiveness of anti-bird devices, which makes full use of the historical anti-bird device utilization data collected. First, the random forest model is trained using data deriving from the numbers of bird nest in the tower before and after the installation of anti-bird devices. Then, this model can be applied to predict the effectiveness of an anti-bird device before installation. The experimental results show that the method proposed in this paper can provide analysis conclusion about the influencing level of each factor to effectiveness of anti-bird device as a percentage. Furthermore, by adjusting the parameters of the random forest model, evaluation results with high accuracy can be obtained, and compared with the single decision tree algorithm, the effectiveness evaluation results giving by the proposed method not only has higher accuracy, but also has greater stableness.