基于BIM建模和人工神经网络的快/慢速火灾识别方法
An Identification Approach for the Fast/Slow-speed Fire Based on BIM Modeling and Neural Network
刘畅 1王张舒君 2张恩茂 3张涵清 2韩宇 4王艳敏2
作者信息
- 1. 中设数字技术有限公司,北京 100043
- 2. 哈尔滨工业大学电气工程及自动化学院,哈尔滨 150001
- 3. 中国建筑设计研究院有限公司,北京 100044
- 4. 哈尔滨哈飞建筑安装工程有限责任公司,哈尔滨 150060
- 折叠
摘要
通过BIM建模技术和神经网络方法,本文提出了一种适用于快/慢火灾识别的新算法.建立该建筑的BIM模型,针对模型进行FDS火灾数值模拟,获取建筑中各种火情的仿真数据;继而基于BP神经网络方法,对FDS数据进行训练分析,实现对火灾信息的提取.最后,分别选取休闲区为例的小空间慢速火燃烧和化妆室为例的小空间快速火燃烧两种情况进行火灾模拟,以验证本文所提方法的正确性和有效性.
Abstract
In this paper,a new approach of fire identification for the fast/slow-speed fire is proposed based on BIM and neural network technology.Create a BIM model of the building and perform FDS numerical simulations on the model to obtain simulated data for various fire scenarios within the building.And then by utilizing the neural network technology,the FDS data is collected.Finally,taking a recreational region and a dressing room as two examples to simulate the slow-speed fire and fast-speed fire,respectively,so that the proposed approach is validated.
关键词
神经网络/BIM/FDS/火灾模拟/火情判断Key words
neural network/BIM/FDS/fire simulation/fire identification引用本文复制引用
出版年
2024
NETL
NSTL
万方数据