基于PSO-BP神经网络高速公路建设期碳排放预测方法

打开文本图片集

中图分类号：U415 文献标识码：A DOI：10.7535/hbkd.2025yx03009

Carbon emission prediction method for expressway construction period based on PSO-BP neural network

ZHAO Quansheng¹，LI Fei¹，GUO Feng'ai1，YU Jianyou 2 ，XU Shizhao 3 ， HU Yunpeng4，CHU Xiaomeng5

ool of Civil Engineering，Hebei Universityof Science and Technology，Shijiazhuang，Hebei O5ool8，China; 2.Hebei Expressway Han Gang Port Company Limited，Cangzhou，Hebei O6l5oo，China;

3.SchoolofMechanicalEngineering，Hebei UniversityofScienceand Technology，Shijiazhuang，HebeiO5oo18，China; 4.Qinhuangdao Highway Construction and Development Center，Qinhuangdao，Hebei O66o99，China; 5.School of Chemical and Pharmaceutical Engineering，Hebei University of Science and Technology，Shijiazhuang， Hehei .China）

Abstract：To solvetheproblemof inaccuratecarbonemissons predictionduring the highwayconstruction period，amethodof Optimizing theback propagation（BP）neuralnetwork byparticleswarmoptimization（PSO）algorithmwas proposed topredictcarbon emisions.The12keyndicators，includingroutelength，subgradelength，pavementlngth，tunellength，ridgeandculvertlength， interchangelength，excavationvolume，filingvoue，dieselconsumption，cementconsumption，crushedstoneconsumptionadteel consumption，wererefinedfromthefourdimensionsof project length，construction，energyconsumptionandmaterialconsumption usingtheanalytichierarchyprocess（AHP）.Thedatafrom36hghwayprojectswereusedasempiricalsamplesformodelrainng，nd a comparative analysis was conducted based on error indicators. The results show that the R2 value of the obtained PSO-BP model is 0.974，while the R2 value of the BP model is0.89o，with the former being closer to1.Compared to the results of life cycle assessment，thePSO-BPmodelhasasmalerdeviationfromtheactualvaluethantheunoptimizedBPmodel.Thefourlayersof the hierarchyandtheselected12keyindicatorsenablethepredictionofcarbonemisionsduringthedesignandplaningstageofhighway construction，providing reference for low-carbon highway construction.

Keywords：otherdisciplines of road engineering;carbon emisson prediction；PSO-BP neural network；modeloptimization; factor analysis

开展碳排放量预测是研究高速公路建设期碳达峰、碳中和的基础工作，这一工作不仅是制定减排策略的前置步骤，更是评估环境影响的关键工具[1-2]。（剩余13746字）