基于电阻率和支持向量机算法的混凝土振捣范围预测  

Predicting radius of action of vibrator based on resistivity data and support vector machine algorithm

作　　者：徐德[1] 田正宏[2] 姜林峰 XU De;TIAN Zhenghong;JIANG Linfeng(Collaborative Innovation Research Center of Intelligent Construction and Blockchain,Jiangsu Open University,Nanjing 210036,China;College of Water Resources and Hydropower,Hehai University,Nanjing 210000,China)

机构地区：[1]江苏开放大学智能建造与区块链协同创新研究中心,江苏南京210036 [2]河海大学水利水电学院,江苏南京210000

出　　处：《混凝土》2021年第12期141-146,共6页Concrete

基　　金：国家自然基金资助(51879094);国家住房与城乡建设部科技项目的阶段性成果(2020-K-061)。

摘　　要：基于混凝土电阻率和支持向量机算法,提出了一种新拌混凝土振捣范围的原位预测方法。首先采集振捣过程中混凝土不同区域的电阻率数据,经小波去噪后提取特征值;将受振混凝土各区域作为样本群,以各区域的混凝土电阻率特征值、与振捣棒的相对位置以及混凝土流动性特征值作为样本信息,以振捣作用下铝球是否沉降至该区域作为分类标签,搭建支持向量机分类模型,对混凝土各个区域是否处于振捣范围内进行二分类判断,并将两类区域的分界位置作为振捣范围的表征;预测结果表明,以坍落度为60、90 mm的混凝土试验数据作为监督学习的训练组,训练得到的支持向量机分类模型能够准确预测相同振捣棒在坍落度170 mm的混凝土中的振捣范围,证明了使用电阻率数据表征新拌混凝土振捣范围的可行性。Based on fresh concrete resistivity and support vector machine(SVM)algorithm,an in situ prediction method for radius action of vibrator was proposed.Firstly the resistivity of different areas of fresh concrete during the vibration were recorded,and the high frequency noise points in the original data were removed by wavelet filtering.Then the support vector machine algorithm was introduced and concrete was divided into a few areas as sample cluster.The characteristic value of resistivity,position relative to the vibrator of a certain area,fluidity of concrete were adopted as information of simples,whether the aluminum ball settled into a certain area was used as category label.Based on these data,a support vector machine classification model was built to classify and judge whether a certain area was within the radius of action of vibrator.Prediction results show that using the data of experiment carried in concrete with slump of 60 mm,90 mm as training groups of supervised learning mode,the proposing SVM classification model can accurately predict the radius of action of vibrator in concrete with slump of 170 mm,which proves the feasibility of using resistivity data as a characteristic value to reveal the radius of action of vibrator.

关 键 词：混凝土电阻率 振捣范围预测 小波去噪 监督学习 支持向量机 

分 类 号：TU528.01[建筑科学—建筑技术科学]