PDF全文阅读

宋立立 ， 孙    妍

（ 北京市建筑设计研究院有限公司 ，北京 100045)

摘   要：针对民用建筑供电不稳定，评估结果的适应度和准确性较低的问题，基于决策树 ，设计民用建筑供电可靠性智能评估算法 。提取民用建筑供电故障节点信息熵，结合信息源输出平均信息熵与系统内的总信息熵，得到成功与失败节点的信息熵；基于决策树修正可靠性评估数据，对训练样本进行梯度排序与预测，引入随机变量，获取不同时段的数据组统计量；建立民用建筑供电可靠性评估模型，完成民用建筑供电可靠性智能 评估算法设计。试验结果表明，在获取网络模型中的最佳节点数量后，最小均方差和平均适应度较优，保证评估算法的精度。

关键词 ：决策树算法 ； 民用建筑 ；供电可靠性 ；智能评估模型 ；可靠性评价

中图分类号 ：TN06           

文献标志码 ：A            

文章编号 ：1005- 8249  (2023)   06- 0128- 06

DOI : 10. 19860/j . cnki . issn1005- 8249. 2023 . 06. 021

SONG Lili ,  SUN Yan

( Beijing Institute of Architectural Design ,   Beijing 100045 ,   China)

Abstract: Aiming at the problem of unstable power supply of civil buildings and low adaptability and accuracy of evaluation results, an intelligent evaluation algorithm for power supply reliability of civil buildings is designed based on decision tree. Extract the information entropy of power supply fault nodes of civil buildings, and obtain the information entropy of successful and failed nodes by combining the average information entropy of information source output and the total information entropy in the system; The reliability evaluation data is modified based on the decision tree, the training samples are graded and predicted, and the random variables are introduced to obtain the data group statistics in different periods; The power supply reliability evaluation model of civil buildings is established, and the intelligent evaluation algorithm design of power supply reliability of civil buildings is completed. The experimental results show that after obtaining the optimal number of nodes in the network model, the minimum mean square deviation and average fitness are better to ensure the accuracy of the evaluation algorithm.

Keywords: decision tree algorithm; civil buildings; power supply reliability; intelligent evaluation model; reliability evaluation;

作者简介 ：宋立立  ( 1983—) ,   男 ，硕士 ， 高级工程师 ， 主 要研究方向 ：工业与民用建筑电气系统设计。

收稿日期 ：2023- 05- 07