基于K近邻算法的高粘结性能混凝土抗压强度预测  

Prediction of compressive strength of concrete with high cohesive properties based on K-nearest neighbor algorithm

作　　者：伍晓圆[1] 刘艳[1] WU Xiaoyuan;LIU Yan(Ningxia Vocational and Technical College,Yinchuan 750001,China)

机构地区：[1]宁夏职业技术学院,宁夏银川750001

出　　处：《粘接》2025年第3期24-27,共4页Adhesion

基　　金：宁夏回族自治区重点研发计划项目(项目编号:2022BEG02018)。

摘　　要：针对掺合料种类繁多,无法适应粘结界面的粗糙度,降低了抗压强度的预测精度问题,从不同硅灰掺量、钢纤维掺量、粉煤灰掺量角度,制备不同配合比条件的高粘结性能混凝土试件,将不同配合比掺量数据作为K近邻算法的输入,以适应粘结界面的粗糙度,计算新配比样本与参考配比样本配比特征的欧几里得距离,将距离最小的参考配比样本中混凝土抗压强度作为新配比样本中混凝土抗压强度预测值,提高抗压强度的预测精度。试验结果表明,硅灰掺量、钢纤维掺量、粉煤灰掺量分别是25%、4%、10%时,高粘结性能混凝土抗压强度较优。In view of the wide variety of admixtures,which cannot adapt to the roughness of the bonding interface and reduce the prediction accuracy of compressive strength,concrete specimens with high bonding performance under different mix conditions were prepared from the perspectives of different silica fume content,steel fiber content and fly ash content,and the data of different mix ratios were used as the input of the K-nearest neighbor algorithm to adapt to the roughness of the bonding interface,and the Euclidean distance between the new ratio sample and the reference ratio sample was calculated.The compressive strength of concrete in the reference ratio sample with the smallest distance was used as the predicted value of concrete compressive strength in the new ratio sample to improve the prediction accuracy of compressive strength.The test results showed that when the silica fume,steel fiber and fly ash content were 25%,4%and 10%,respectively,the compressive strength of concrete with high bonding performance is better.

关 键 词：K近邻算法 高粘结性能 抗压强度 超高性能混凝土 配合比 

分 类 号：TQ177.68[化学工程—硅酸盐工业] TU528[建筑科学—建筑技术科学]