基于数字孪生的火电工程BIM数据映射与分析技术  

BIM data mapping and analysis technology of thermal power engineering based on digital twin

作　　者：项岱军 潘夏清 陈喜龙 张祖辉 左超 XIANG Daijun;PAN Xiaqing;CHEN Xilong;ZHANG Zuhui;ZUO Chao(CHN Energy Zhejiang Power Co.,Ltd.,Hangzhou 310000,China;Guoneng Zhejiang Zhoushan Power Generation Co.,Ltd.,Zhoushan 316012,China;China Energy Engineering Group Zhejiang Electric Power Design Instiute Co.,Ltd.,Hangzhou 310012,China;Beijing Glory PKPM Technology Co.,Ltd.,Beijing 100013,China)

机构地区：[1]国家能源集团浙江电力有限公司,浙江杭州310000 [2]国能浙江舟山发电有限责任公司,浙江舟山316012 [3]中国能源建设集团浙江省电力设计院有限公司,浙江杭州310012 [4]北京构力科技有限公司,北京100013

出　　处：《电子设计工程》2025年第8期102-106,共5页Electronic Design Engineering

基　　金：国家能源投资集团有限责任公司科技创新项目(GJNY-22-9-1)。

摘　　要：针对当前国产火电工程BIM数据提取分析能力不足、处理效率低等问题,文中以数字孪生理论为基础,提出了一种基于自动映射与改进粒子群优化反向传播神经网络(PSO-BP)的火电工程BIM数据映射与分析技术。在构建数字孪生平台架构的基础上,构造了工业基础类数据(IFC)到关系型数据库的自动映射方法,并设计基于改进PSO算法优化的BP神经网络结构,用于实现映射后数据的智能化分析与趋势预测。经实验可知,所提方法在大数据量场景下的数据修改效率比达到传统方法的374倍,改进PSO-BP算法所得到测试结果的均方根误差为0.163,精度为0.986,均优于现有智能算法,具有良好的实际工程应用价值。In response to the current problems of insufficient data extraction and analysis capabilities and low processing efficiency in domestic thermal power engineering BIM,this paper proposes a thermal power engineering BIM data mapping and analysis technology based on automatic mapping and improved Particle Swarm Optimization Back Propagation(PSO-BP)neural network based on digital twin theory.On the basis of constructing a digital twin platform architecture,an automatic mapping method from Industrial Foundation Classes(IFC)to relational databases was constructed,and a BP neural network structure optimized based on an improved PSO algorithm was designed to achieve intelligent analysis and trend prediction of mapped data.Through experimental analysis,the data modification efficiency of the proposed method in large data scenarios is 374 times higher than that of traditional methods.The root mean square error of the improved PSO-BP algorithm in the test results is 0.163,with an accuracy of 0.986,which is better than existing intelligent algorithms and has good practical engineering application value.

关 键 词：建筑信息模型 火电工程 自动映射 反向传播神经网络 粒子群优化算法 

分 类 号：TM621[电气工程—电力系统及自动化]