PDF全文阅读

王逸飞1 ,4 ,  瓦依提 · 力提甫2 ,  姬豪杰1 ,4 ,  刘尊青1 , 3 ,4

 

( 1 . 新疆农业大学 交通与物流工程学院 ，乌鲁木齐 830052 ;  2.  新疆交投建设管理有限责任公司，

乌鲁木齐 830000 ;  3 .  新疆农业大学 水利与土木工程学院 ，乌鲁木齐 830052 ;

4. 新疆道路工程试验检测研究中心（新疆农业大学道路工程研究中心），乌鲁木齐 830052)

 

摘   要 ：为系统研究钢渣粉－矿渣粉复合胶凝材料的力学性能演变规律 ，采用不同配比的矿渣粉 、钢渣粉与水泥制备胶砂试件 ，测试其在 1 、3 、7 、28 d 龄期抗折和抗压强度 ；建立 BP 神经网络和随机森林模型预测该复合胶凝材料的后期力学性能。结果表明： 当掺入 10% 钢渣粉与 30% 矿渣粉时 ，力学性能优异 ，28 d 抗折强度达 12. 22 MPa ,  抗压强度为 50. 10 MPa;  BP 神经网络预测精度显著优于随机森林模型 ，其抗压强度预测的 MAE、 RMSE 和 R2  分别为 1 . 823 、2. 136 和 0. 956 。研究为工业固废基胶凝材料的性能预测提供可靠方法 ，对推动绿色低碳建材的应用提供理论依据及参考。

关键词：钢渣粉 ；矿渣粉 ；BP 神经网络 ；随机森林 ；强度预测

中图分类号：TU528           

文献标志码：A           

文章编号： 1005- 8249   (2025)  04- 0001- 06

DOI:10. 19860/j.cnki.issn1005 - 8249.2025 .04.001

 

WANG Yifei 1 , 4  ,  WAIITI Litipu2 ,  JI Haojie1 , 4  ,  LIU Zunqing1 , 3 , 4

( 1 . School of Traffic and Logistics Engineering, Xinjiang Agricultural University, Urumqi 830052, China;

2. Xinjiang Transportation Investment and Construction Management Co . Ltd., Urumqi 830000, China;

3 . College of Hydraulic and Civil Engineering, Xinjiang Agricultural University, Urumqi 830052, China;

4. Xinjiang Road Engineering Test and Inspection Research Center (Xinjiang Agricultural University

Road Engineering Research Center), Urumqi 830052, China)

Abstract : To systematically investigate the evolution law of mechanical properties of steel slag powder - slag powder composite cementitious materials, mortar specimens were prepared using different proportions of slag powder, steel slag powder, and cement, and their flexural and compressive strengths were tested at curing ages of 1, 3, 7 and 28 days; BP neural network and random forest models were further established to predict the later mechanical properties of the composite cementitious material.The results showed that when 10% steel slag powder and 30% slag powder were added, the mechanical properties were excellent, with flexural strength reaching 12. 22 MPa and compressive strength reaching 50. 10 MPa at 28 days; The prediction accuracy of the BP neural network was significantly superior to that of the random forest model, with mean absolute error (MAE), root mean square error ( RMSE), and coefficient of determination ( R2 ) for compressive strength prediction being 1 . 823, 2. 136 and 0. 956, respectively. This study provides a reliable method for predicting the performance of industrial solid waste - based cementitious materials and offers theoretical basis and reference for promoting the application of green and low - carbon building materials .

Key words : steel slag powder; slag powder; BP neural network; random forest; strength prediction

 

 

基金项目： 国 家 自 然 科 学 基 金 资 助项 目  ( 52268072 ) ； 新 疆 农 业 大 学 交 通 运 输 工 程 校 级 重 点 学 科 开 放 课 题 资 助项 目(XJAUTE2022K04) 。

作者简介： 王逸飞  (2002—) ,  男 ，硕士研究生 ，研究方向：道路建筑材料。

通信作者：姬豪杰  (1995—) ,  男 ，硕士 ，讲师 ，研究方向：路基路面工程。

收稿日期：2024 - 10 - 28