摘要
寒区沥青路面施工面临骤然降温、大温差、大风等不利天气,造成沥青路面摊铺、碾压温度起伏大,影响沥青路面压实质量,进而诱发诸多病害。以沥青路面施工温度为核心,通过物联网系统和智能元件采集沥青路面施工工艺参数并进行分析,采用机器学习算法构建沥青路面施工温度预估模型,准确掌控沥青路面摊铺和碾压施工温度,确保沥青路面施工质量。结果表明:沥青混合料拌和参数控制精度高,运输、摊铺和碾压阶段的施工工艺参数变异性大,需对参数进行异常值剔除处理,表明当前施工工艺控制技术仍有提升空间;采用随机森林算法对施工工艺参数进行重要性评估,出料温度和施工速度对施工温度影响最显著;基于4种机器学习算法建立了沥青路面施工温度预测模型,其中多层感知机模型最优,对多层感知机模型的隐藏层、神经元个数和学习率进行优化,优化后模型的周期数、均方误差和平均绝对误差降低,整体性能显著提升;考虑气象参数后,施工温度预测模型的训练效率降低,但预测精度提高。工程应用表明:提出的基于多层感知机沥青路面施工温度预测模型与实际工况相符,通过调节出料温度、摊铺速度、碾压速度可以有效减少混合料温度损失。
Asphalt pavement construction in cold regions is faced with unfavorable weather such as sudden cooling,large temperature difference,and strong wind,resulting in large fluctuations in the temperature of asphalt pavement paving and rolling,affecting the compaction quality of asphalt pavement and causing many diseases.Taking the asphalt pavement construction temperature core,the asphalt pavement construction process parameters are collected and analyzed through the Internet of Things system and intelligent components,and the asphalt pavement construction temperature prediction model is constructed by machine learning algorithm,so as to accurately control the asphalt pavement paving and rolling construction temperature and ensure the quality of asphalt pavement construction.The results show that the control accuracy of asphalt mixture mixing process parameters is high,and the variability of construction process parameters in the stages of transportation,paving and rolling is large,and outliers need to be eliminated through data cleaning,indicating that there is still room for improvement in the current construction process control technology.The importance of construction process parameters was evaluated by the random forest algorithm,and the discharge temperature and construction speed had the most significant influence on the construction temperature.The construction temperature prediction model of asphalt mixture is established based on four machine learning algorithms,and the effect of multi-layer perceptron model is the best.After optimizing the hidden layer,the number of neurons and the learning rate of the multi-layer perceptron model,the cycle number,mean squared error and mean absolute error of the optimized model are reduced,and the overall performance of the model is significantly improved.After considering the meteorological parameters,the training efficiency of model is reduced,but the prediction accuracy of model is improved.The engineering application shows that the asphalt pavement construction temper
作者
司伟
茆纬杰
李宁
石岩
次旦多杰
陈星
马骉
SI Wei;MAO Wei-jie;LI Ning;SHI Yan;Cidanduo jie;CHEN Xing;MA Biao(Key Laboratory of Special Area Highway Engineering of Ministry of Education,Chang'an University,Xi'an 710064,Shaanxi,China;Tibet Tianlu Co.Ltd.,Lhasa 850000,Tibet,China;School of Civil Engineering,Xi'an University of Architecture and Technology,Xi'an 710055,Shaanxi,China;Zhongke Huayan(Xi'an)Technology Co.Ltd.,Xi'an 710199,Shaanxi,China)
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2023年第3期81-97,共17页
China Journal of Highway and Transport
基金
国家自然科学基金项目(52278430)
西藏自治区自然科学基金重点项目(XZ202101ZR0046G)
西藏自治区科技厅重大科技专项西藏天路科创基金项目(XZ 2019TL-G-05)。
关键词
道路工程
温度预估
机器学习
沥青混合料
施工过程
road engineering
temperature prediction
machine learning
asphalt mixture
construction process
