基于PCA-FPP-BP神经网络的高层建筑火灾安全评价

Fire Safety Assessment of High Rise Building Based on PCA-FPP-BP Neural Network

为解决高层建筑构造复杂、人员密度大、火灾触发因素繁多而造成高层建筑火灾安全评价困难的问题,本文提出基于PCA-FPP-BP神经网络的高层建筑火灾安全评价模型。首先运用主成分分析(PCA)对构建的高层建筑火灾安全评价指标降维处理,筛选主要信息;接着基于三角模糊数构建模糊评判矩阵,利用模糊优先规划(FPP)求解指标的权重值,减少主观的影响;最后考虑到指标间关系错综复杂彼此交叉和反馈的特性,选择BP神经网络对高层建筑火灾安全进行评价。通过工程案例证明该评价模型的实用性以及可靠性。

In order to solve the difficult problem of the high- rise building fire safety evaluation caused by the complex high- rise building structure,personnel density and various fire- caused factors,this paper proposes a high- rise building fire safety assessment model based on PCA- FPP- BP neural network. Firstly,the principal component analysis（ PCA） is used to reduce the dimension of the building fire safety evaluation index,and the main information is selected; then,fuzzy judgment matrix is constructed based on triangular fuzzy number,the weight value of the index is solved by fuzzy priority programming（ FPP） and the subjective influence is reduced; finally considering the complicated relations between indexes and the characteristics of cross each other and feedback,the BP neural network is selected to evaluate the fire safety of high- rise buildings. The practicability and reliability of the evaluation model are proved through the engineering case.