基于SVR的智能建筑火灾预警模型设计

The Intelligent Building Fire Pre-Warning Model Design Based on SVR

火灾信息处理算法的有效性直接决定着建筑火灾自动预警系统的可靠性,开发新型智能火灾预警算法是目前建筑火灾探测预警领域研究的热点之一.针对现有火灾预警算法的不足,研究设计提出一种基于支持向量回归机(SVR)的智能建筑火灾预警算法.为了验证该算法在多传感器复合式建筑火灾预警系统信息处理中的可靠性与优越性,以普通火灾和欧洲试验火历史数据为例,通过Matlab模拟仿真,进行实证分析,并将预警结果与BP神经网络预警结果进行对比分析.研究成果可为新型建筑火灾自动预警系统的设计提供科学的依据.

The effectiveness of the fire information processing algorithm directly determines the reliability of the building auto-fire warning system, so developing new fire pre-warning algorithm is one of the hotspot in the field of building fire detection and pre-warning research. Aiming at the deficiencies of the existing fire warning algorithm, an intelligent building fire warning algorithm based on the support vector regression machine （SVR） is designed and proposed in the study. In order to verify the information processing reliability and superiority of the algorithm in the composite building fire pre-warning system, with the ordinary fire history data and the European standard test fire history data as example, to make empirical analysis through Matlab simulation, and the pre-warning results were compared with the BP neural network pre-warning results. The research results can provide a scientific basis for the design of new building auto-fire pre-warning system.