火蔓延速度模拟原始数据

2018年实验室通过利用学习型山火传播预测模型，计算火锋面的正常移动测量实时扩散率(RoS)，利用机器学习将局部活性氧与环境参数关联起来预测未燃烧区域的活性氧，利用窄带水平集方法模拟火锋的演化过程，通过这些数值对生成的火灾锋面进行验证，应用于真实规模的灌木丛火灾场景。数据包括：含水率、热辐射、点燃时间、点燃温度、蔓延速度共168条/2030.88kb

In 2018, the laboratory employed a machine learning-based wildfire spread prediction model. Specifically, the model calculated the real-time spread rate (RoS) of the fire front via normal movement measurement. Then, machine learning was utilized to correlate local reactive oxygen species (ROS) with environmental parameters to predict ROS levels in unburned areas. The narrow-band level set method was adopted to simulate the evolution of the fire front, and the generated fire fronts were validated against these numerical results. This approach was applied to full-scale bushfire scenarios. The dataset contains 168 entries with a total size of 2030.88 kb, covering moisture content, thermal radiation, ignition time, ignition temperature, and spread rate.

搜集汇总

背景与挑战

背景概述

该数据集为2018年实验室生成的火灾蔓延模拟原始数据，通过机器学习模型和数值方法预测火锋面演化，应用于真实灌木丛火灾场景。数据包含含水率、热辐射等168条记录，隶属于人工林重大灾害研究项目，用于支持植物学领域的火灾行为分析。