基于朴素贝叶斯网络的输电走廊山火风险评估模型

Wildfire Risk Assessment Model of Power Transmission Line Corridors Based on Naive Bayes Network

针对山火风险评估中物理模型构建复杂的问题,提出了一种基于因子选择的朴素贝叶斯网络输电走廊山火风险评估模型。首先,收集和分析了对山火发生产生影响的人为、地表环境和气象3大类共14个影响因子的数据。然后采用Relief算法按照对山火风险影响的重要性对因子进行排序。逐一删减最不重要因子后,基于朴素贝叶斯网络构建最优输电走廊山火风险评估模型。最后,以广东省为例进行山火风险评估,并将评估结果可视化绘制成输电走廊山火风险分布图,然后构建了代价函数验证山火风险分布图的适用性。筛选2020年第1季度卫星监测火点进行验证表明,80.9%的火点落在较高和高风险区域,可指导运维人员开展差异化的山火防治工作。

In wildfire risk assessment,numerous factors are contributed to wildfire occurrence.And the physical model construction is complex.This paper develops a method to evaluate the wildfire risk of power transmission line corridors based on Naive Bayes Network.First,the data of 14 wildfire-related factors including anthropogenic,physiographic and meteorologic factors,are collected and analyzed.Then,the relief algorithm is used to rank the importance of factors according to their impacts on wildfire occurrence.After eliminating the least important factors in turn,an optimal wildfire risk assessment model for power transmission line corridors is constructed based on the Naive Bayes Network.Finally,this model is carried out and visualized in Guangdong Province in southern China.Then a cost function is proposed to further verify the applicability of the wildfire risk distribution map.Then a cost function is proposed to further verify the applicability of the wildfire risk distribution map.The fire-spots monitored by satellites during the first season in 2020 shows that 80.9%of fire-spots fall in higher and high risk area.This method can guide operation and maintenance personnel to carry out differentiated wildfire management work.