文中制备三种沥青、五个粉胶比的沥青胶浆,开展135、145、155、165、175、185 ℃的旋转黏度试验.通过试验研究、对比分析,以及数学拟合等方法研究沥青类型、粉胶比和温度对试验结果的影响规律.将试验结果中的沥青类型用针入度进行量化,得到用于预测沥青胶浆黏度的数据库,并通过沥青胶浆的预测值和实际值的拟合效果及数据集的均方根误差(root mean square error,RMSE)和相关系数(correlation coefficient,R)验证天牛须搜索算法(beetle antennae search,BAS)和随机森林(random forest,RF)混合机器学习模型预测沥青胶浆黏度的准确性.结果表明:沥青类型显著影响胶浆的黏度;增大粉胶比可提高胶浆的黏度;粉胶比不影响胶浆黏度的温度敏感性,胶浆的黏度与温度的关系均可用半对数线性函数拟合;BAS和RF混合机器学习模型对胶浆黏度的预测准确性较高.
Study on Influencing Factors of Asphalt Mortar Viscosity and RF-BAS Prediction Model
Three kinds of asphalts and asphalt pastes with five powder/binder ratios were prepared,and the rotational viscosity tests were carried out at 135℃,145℃,155℃,165℃,175℃ and 185℃.Through experimental research,comparative analysis and mathematical fitting,the influence of as-phalt type,powder-binder ratio and temperature on the test results was studied.The asphalt types in the test results were quantified by penetration,and a database for predicting the viscosity of asphalt mortar was obtained.And through the fitting effect between the predicted value and the actual value of asphalt mortar,root mean square error(RMSE)and correlation coefficient(R)of the data set,the accuracy of predicting the viscosity of asphalt mortar was verified by the hybrid machine learning mod-el of beetle antennae search(BAS)and random forest(RF).The results show that the asphalt type significantly affects the viscosity of mucilage.Increasing the ratio of powder to glue can improve the viscosity of glue.The ratio of powder to binder does not affect the temperature sensitivity of mucilage viscosity,and the relationship between mucilage viscosity and temperature can be fitted by semi-loga-rithmic linear function.The mixed machine learning model of BAS and RF has high accuracy in predic-ting the viscosity of mucilage.
asphalt mortarviscosityhigh temperature performancehybrid machine learning model
万方数据
维普
