|本期目录/Table of Contents|

[1]娄镇涛,吴大志,刘程旭,等.基于改进 BP神经网络的粉煤灰-矿渣基地质聚合物固化软土强度预测模型[J].浙江理工大学学报,2026,55-56(自科三):249-257.
　LOUZhentao,WUDazhi,LIUChengxu,et al.A strength prediction model of fly ash-slag-based geopolymer-stabilized soft soil using an improved BP neural network[J].Journal of Zhejiang Sci-Tech University,2026,55-56(自科三):249-257.
点击复制

基于改进 BP神经网络的粉煤灰-矿渣基地质聚合物固化软土强度预测模型()

浙江理工大学学报[ISSN:1673-3851/CN:33-1338/TS]

卷:
55-56
期数:
2026年自科第三期
页码:
249-257
栏目:
出版日期:
2026-05-10

文章信息/Info

Title:
A strength prediction model of fly ash-slag-based geopolymer-stabilized soft soil using an improved BP neural network
文章编号:
1673-3851(2026) 05-0249-09
作者:
娄镇涛 ; 吴大志 ; 刘程旭 ; 陈柯宇
1. 浙江理工大学建筑与工程学院 ,杭州 310018;2. 浙江大学建筑与工程学院 ,杭州 310058
Author(s):
LOUZhentao ; WUDazhi ; LIUChengxu ; CHEN Keyu
1. School of Civil Engineering and Architecture, Zhejiang Sci-Tech University, Hangzhou310018, China; 2. College of Civil Engineering and Architecture, Zhejiang University, Hangzhou310058, China
关键词:
粉煤灰-矿渣基地质聚合物 ;  固化软土强度; BP神经网络; 预测模型 ; 敏感性分析
分类号:
TU433
文献标志码:
A
摘要:
传统水泥固化软土存在能耗高、碳排放量大且早期性能不足等问题 ,且粉煤灰-矿渣基地质聚合物固化软土强度受到众多因素影响 ,使得最佳配比难以准确预估 。针对上述情况 ,构建基于改进 BP神经网络的预测模型 ,用于快速准确预测粉煤灰-矿渣基地质聚合物固化软土强度并辨识关键影响因素 。首先 ,依据已有文献资料 ,选取前驱体掺量、粉煤灰氧化物成分、碱激发剂参数及土体物理性质等 14个关键变量 ,建立 178组样本数据集 。然后 ,分别采用粒子群优化算法(Particleswarm optimization, PSO)和自适应提升算法(Adaptiveboosting, AdaBoost) ,对反向传播神经网络(Back propagationneuralnetwork, BPNN)进行优化 ,建立 PSO-BPNN与 AdaBoost-BPNN预测模型 ,并与传统 BPNN、径向基函数神经网络(Radialbasisfunction neuralnetwork, RBFNN)及随机森林(Random forest, RF)模型进行对比 。最后 ,使用 Garson算法对最优模型进行敏感性分析 。结果表明 :构建的两种模型在预测精度与稳定性方面均显著优于传统模型 ,其中 PSO-BPNN模型在拟合精度与泛化能力方面表现最优 , 能够更准确地预测多因素下固化软土强度;NaOH 浓度、矿渣掺量及粉煤灰中 CaO含量是影响粉煤灰-矿渣基地质聚合物固化软土强度的主要因素 。该研究考虑了地质聚合物作用要素与土体物理性质 ,构建的 PSO-BPNN模型可以有效改善传统模型精度不足的缺陷 , 实现了固化软土强度的高效精确预测 ,可为工程实践中固化软土配比优化提供参考。

参考文献/References:

[1] Lin P Y, Chen X Y, Jiang M J, et al. Mapping shear strength and compressibility of soft soils with artificial neural networks [J]. Engineering Geology, 2022, 300: 106585.
[2] Wu Y G, Lu B W, Bai T, et al. Geopolymer, green alkali activated cementitious material: Synthesis, applications and challenges [J]. Construction and Building Materials, 2019, 224: 930-949.
[3] 王辉，高尚，郭美丽，等。粉煤灰 - 矿渣基泡沫地聚合物微观孔隙结构与宏观性能 [J]. 建筑材料学报，2024, 27 (5): 461-470.
[4] 张津津，李博，余闯，等。矿渣 - 粉煤灰地聚合物固化砂土力学特性研究 [J]. 岩土力学，2022, 43 (9): 2421-2430.
[5] 胡建林，赵天怡，冷发光，等。碱激发矿渣 - 粉煤灰基地质聚合物固化土力学特性及强度预测 [J/OL]. 复合材料学报，2025: 1-12. (2025-05-26)[2025-11-18]. https://doi.org/10.13801/j.cnki.fhclxb.20250526.003.
[6] Min Y F, Gao M J, Yao C Q, et al. On the use of one-part geopolymer activated by solid sodium silicate in soft clay stabilization [J]. Construction and Building Materials, 2023, 402: 132957.
[7] Yu J R, Chen Y H, Chen G, et al. Experimental study of the feasibility of using anhydrous sodium metasilicate as a geopolymer activator for soil stabilization [J]. Engineering Geology, 2020, 264: 105316.
[8] 骆展鹏，熊春林，韩泽军，等。矿渣 - 粉煤灰 - 玻璃粉复合固化盾构土力学性能及固化机制 [J]. 硅酸盐通报，2025, 44 (5): 1803-1812.
[9] 赵东三，徐正中，高旭，等。基于动态机器学习的水泥胶砂性能预测研究 [J]. 混凝土，2024 (3): 147-152.
[10] 蔡锦程，许子彦，董振勇，等。基于机器学习的砂浆流变特性预测 [J]. 公路交通科技，2024, 41 (6): 138-147.
[11] 胡艳香，陈昊，薛凯喜，等。基于 PSO-BP 和 GA-BP 算法神经网络的锂渣 - 橡胶砂浆抗压强度预测研究 [J]. 混凝土，2023 (10): 59-63.
[12] 马士宾，李盈霞，闫伟阳。基于两种方法优化磷酸镁水泥砂浆配合比研究 [J]. 混凝土，2023 (1): 140-145.
[13] Cai B, Lin X Q, Fu F, et al. Postfire residual capacity of steel fiber reinforced volcanic scoria concrete using PSO-BPNN machine learning [J]. Structures, 2022, 44: 236-247.
[14] Abdolrasol M G M, Suhail Hussain S M, Ustun T S, et al. Artificial neural networks based optimization techniques: A review [J]. Electronics, 2021, 10 (21): 2689.
[15] Turan C N, Javadi A A, Vinai R. Effects of class C and class F fly ash on mechanical and microstructural behavior of clay soil: A comparative study [J]. Materials, 2022, 15 (5): 1845.
[16] Li Y, Shen J, Lin H, et al. Optimization design for alkali-activated slag-fly ash geopolymer concrete based on artificial intelligence considering compressive strength, cost, and carbon emission [J]. Journal of Building Engineering, 2023, 75: 106929.
[17] Zhang J, Choi C E, Liang Z, et al. A generic framework for mix design of geopolymer for soil stabilization: Composition-informed machine learning model [J]. Computers and Geotechnics, 2024, 170: 106322.
[18] Odeh N A, Al-Rkaby A H J. Strength, Durability, and Microstructures characterization of sustainable geopolymer improved clayey soil [J]. Case Studies in Construction Materials, 2022, 16: e00988.
[19] Chen Z, Gao Y, Wei W, et al. Stabilization of soft clay by a low-calcium fly ash geopolymer [J]. Journal of Materials in Civil Engineering, 2023, 35 (11): 04023398.
[20] Bhavita Chowdary V, Ramanamurty V, Pillai R J. Experimental evaluation of strength and durability characteristics of geopolymer stabilised soft soil for deep mixing applications [J]. Innovative Infrastructure Solutions, 2021, 6 (1): 40.
[21] Phetchuay C, Horpibulsuk S, Arulrajah A, et al. Strength development in soft marine clay stabilized by fly ash and calcium carbide residue based geopolymer [J]. Applied Clay Science, 2016, 127/128: 134-142.
[22] Yaghoubi M, Arulrajah A, Disfani M M, et al. Impact of field conditions on the strength development of a geopolymer stabilized marine clay [J]. Applied Clay Science, 2019, 167: 33-42.
[23] Yaghoubi M, Arulrajah A, Disfani M M, et al. Effects of industrial by-product based geopolymers on the strength development of a soft soil [J]. Soils and Foundations, 2018, 58 (3): 716-728.
[24] Abdullah H H, Shahin M A, Sarker P. Use of fly-ash geopolymer incorporating ground granulated slag for stabilisation of Kaolin clay cured at ambient temperature [J]. Geotechnical and Geological Engineering, 2019, 37 (2): 721-740.
[25] 杨明辉，宋牧原，姚高占，等。机器学习预测盾构掘进地表沉降的研究进展及展望 [J]. 隧道建设 (中英文), 2024, 44 (11): 2119-2132.
[26] 渠建伟。基于 BP 神经网络的振冲碎石桩复合地基承载力预测 [J]. 铁道建筑，2017 (4): 87-90.
[27] 魏博文，柳波，徐富刚，等。融合 PSO-SVM 的混凝土拱坝多测点变形监控混合模型 [J]. 武汉大学学报 (信息科学版), 2023, 48 (3): 396-407.
[28] 胡军，董建华，王凯凯。隧道洞室地基稳定性双阶段多策略粒子群 BP 网络模型研究 [J]. 现代隧道技术，2017, 54 (3): 50-57.
[29] 赵佳亮，达列雄，郭鸿，等。基于 MLP-AdaBoost 模型的混凝土抗压强度预测研究 [J]. 混凝土，2025 (6): 17-22.
[30] 代硕，苏怀智，谷宇，等。基于 HO-BP-AdaBoost 的大坝变形监控模型 [J]. 水电能源科学，2025, 43 (7): 153-156.
[31] Zhang M, Du Q, Yang J, et al. Modeling the pile settlement using the Integrated Radial Basis Function (RBF) neural network by Novel Optimization algorithms: HRBF-AOA and HRBF-BBO [J]. Journal of Intelligent & Fuzzy Systems, 2022, 43 (6): 7009-7022.
[32] Liu Y D, Yang Z Y, Li X Y. Adaptive ensemble learning of radial basis functions for efficient geotechnical reliability analysis [J]. Computers and Geotechnics, 2022, 146: 104753.
[33] Chen W H, Ding J W, Wang T F, et al. An adaptive sequential sampling method based on ANN_MCD and RF: Application in geotechnical problems [J]. Applied Soft Computing, 2023, 144: 110462.
[34] Zhou J, Huang S, Zhou T, et al. Employing a genetic algorithm and grey wolf optimizer for optimizing RF models to evaluate soil liquefaction potential [J]. Artificial Intelligence Review, 2022, 55 (7): 5673-5705.
[35] 王泉。基于 BP-Garson 组合模型的堤岸稳定影响因素敏感性分析 [J]. 水电能源科学，2018, 36 (6): 122-124.
[36] 朱琎琦，牛晓凡，肖显斌。基于改良 BP 神经网络的生物质锅炉飞灰含碳量预测模型研究 [J]. 可再生能源，2020, 38 (2): 150-157.

备注/Memo

备注/Memo:
基金项目 : 国家自然科学基金项目(52578443)收稿日期 : 2025-10-20 网络出版日期 : 2026-01-27
作者简介 : 娄镇涛(2000 ) ,男 ,杭州人 ,硕士研究生 ,主要从事地基处理技术方面的研究。通信作者: 吴大志, E-mail: wudz@zstu. edu. cn
更新日期/Last Update: 2026-05-07