基于Stacking模型融合的高性能混凝土强度预测方法

摘要/Abstract

摘要： 针对传统经验公式对高性能混凝土强度预测时存在偏差大、效率低等问题,本文提出一种基于Stacking模型融合的高性能混凝土强度预测方法。首先,通过数据清洗和归一化对1 030组高性能混凝土抗压强度试验数据进行预处理,剔除异常数据及消除数据间量纲影响;其次,基于极端梯度提算法(XGBoost)、类别优先梯度提升算法、多层感知器和随机森林(RF)4种算法开展超参数优化、模型训练和评估,采用决定系数、均方根误差和平均绝对误差对比分析4种基学习器对强度预测的整体效果,在此基础上构建基于Stacking集成学习融合多种机器学习算法的高性能混凝土强度预测模型;最后,采用103组新的高性能数据集对模型进行验证,并开展可解释分析。结果表明:与其他基学习器的组合相比,XGBoost和RF融合模型的预测精度和性能均明显提高,泛化性能较好,且可解释分析显示最重要的输入特征变量是龄期和水泥,说明模型内在的预测逻辑与工程实践的经验较吻合,具有较高的合理性与可靠度。研究结果为进一步提高高性能混凝土强度的预测精度提供参考。

关键词: 混凝土, 强度预测模型, 集成学习, stacking算法, XGBoost算法, RF算法

Abstract: Strength prediction method of high performance concrete based on stacking model fusion was proposed to address the issues of large deviations and low efficiency of traditional empirical formulas for high-performance concrete strength prediction. Firstly, 1 030 sets of high-performance concrete compressive strength test data were preprocessed through data cleaning and normalization to eliminate abnormal data and the dimensional influence among data. Secondly, based on extreme gradient boosting (XGBoost), category boosting, multi-layer perceptron, and random forest (RF) algorithms, hyperparameter optimization, model training and evaluation were conducted, and the overall effect of the four base learners on strength prediction were compared and analyzed using coefficient of determination R², root mean square error and mean absolute error. Based on this, a Stacking ensemble learning model was constructed, which fuses multiple machine learning algorithms for concrete strength prediction. Finally, the model was validated using 103 sets of new data, and interpretable analysis was performed. The results show that compared to other combinations of base learners, the fusion model using XGBoost and RF significantly improves prediction accuracy and performance, and has good generalization performance. The interpretable analysis shows that the most important input feature variables are age and cement, indicating that the internal prediction logic of the model is more in line with engineering practice experience, having high rationality and reliability. The research results provide reference for further improving the accuracy of high-performance concrete strength prediction.

Key words: concrete, strength prediction model, ensemble learning, stacking algorithm, XGBoost algorithm, RF algorithm

中图分类号:

胡以婵, 梁铭, 谢灿荣, 解威威, 翁贻令, 池浩, 彭浩, 罗雪霜. 基于Stacking模型融合的高性能混凝土强度预测方法[J]. 硅酸盐通报, 2023, 42(11): 3914-3926.

HU Yichan, LIANG Ming, XIE Canrong, XIE Weiwei, WENG Yiling, CHI Hao, PENG Hao, LUO Xueshuang. Strength Prediction Method of High Performance Concrete Based on Stacking Model Fusion[J]. BULLETIN OF THE CHINESE CERAMIC SOCIETY, 2023, 42(11): 3914-3926.

参考文献

[1] 吴中伟. 混凝土科学技术近期发展方向的探讨[J]. 硅酸盐学报, 1979, 7(3): 262-270.
WU Z W. An approach to the recent trends of concrete science and technology[J]. Journal of the Chinese Ceramic Society, 1979, 7(3): 262-270 (in Chinese).
[2] 李宗津, 孙伟, 潘金龙. 现代混凝土的研究进展[J]. 中国材料进展, 2009, 28(11): 1-7+53.
LI Z J, SUN W, PAN J L. New development of contemporary concrete[J]. Materials China, 2009, 28(11): 1-7+53 (in Chinese).
[3] 俞桂良. 高强混凝土强度预测人工智能方法及应用[J]. 混凝土, 2010(10): 41-43.
YU G L. Strength prediction of high strength concrete using artificial intelligence method and its application[J]. Concrete, 2010(10): 41-43 (in Chinese).
[4] 季韬, 林挺伟, 林旭健. 基于人工神经网络的混凝土抗压强度预测方法[J]. 建筑材料学报, 2005, 8(6): 677-681.
JI T, LIN T W, LIN X J. Prediction method of concrete compressive strength based on artificial neural network[J]. Journal of Building Materials, 2005, 8(6): 677-681 (in Chinese).
[5] 唐明述. 水泥混凝土与可持续发展[J]. 中国有色金属学报, 2004, 14(增刊1): 164-172.
TANG M S. Cement, concrete and sustainable development[J]. The Chinese Journal of Nonferrous Metals, 2004, 14(supplement 1): 164-172 (in Chinese).
[6] LI Z Z, YOON J, ZHANG R, et al. Machine learning in concrete science: applications, challenges, and best practices[J]. NPJ Computational Materials, 2022, 8: 127.
[7] CHOU J S, TSAI C F, PHAM A D, et al. Machine learning in concrete strength simulations: multi-nation data analytics[J]. Construction and Building Materials, 2014, 73: 771-780.
[8] NUNEZ I, MARANI A, FLAH M, et al. Estimating compressive strength of modern concrete mixtures using computational intelligence: a systematic review[J]. Construction and Building Materials, 2021, 310: 125279.
[9] MOHTASHAM M M, SARADAR A, RAHMATI K, et al. Predictive models for concrete properties using machine learning and deep learning approaches: a review[J]. Journal of Building Engineering, 2023, 63: 105444.
[10] KHAMBRA G, SHUKLA P. Novel machine learning applications on fly ash based concrete: an overview[J]. Materials Today: Proceedings, 2023, 80: 3411-3417.
[11] BEN C W, FLAH M, NEHDI M L. Machine learning prediction of mechanical properties of concrete: critical review[J]. Construction and Building Materials, 2020, 260: 119889.
[12] YUAN X Z, TIAN Y Z, AHMAD W, et al. Machine learning prediction models to evaluate the strength of recycled aggregate concrete[J]. Materials, 2022, 15(8): 2823.
[13] SHANG M J, LI H J, AHMAD A, et al. Predicting the mechanical properties of RCA-based concrete using supervised machine learning algorithms[J]. Materials, 2022, 15(2): 647.
[14] IFTIKHAR B, ALIH S C, VAFAEI M, et al. Predictive modeling of compressive strength of sustainable rice husk ash concrete: ensemble learner optimization and comparison[J]. Journal of Cleaner Production, 2022, 348: 131285.
[15] FENG D C, LIU Z T, WANG X D, et al. Machine learning-based compressive strength prediction for concrete: an adaptive boosting approach[J]. Construction and Building Materials, 2020, 230: 117000.
[16] ASTERIS P G, SKENTOU A D, BARDHAN A, et al. Predicting concrete compressive strength using hybrid ensembling of surrogate machine learning models[J]. Cement and Concrete Research, 2021, 145: 106449.
[17] FAROOQ F, AHMED W, AKBAR A, et al. Predictive modeling for sustainable high-performance concrete from industrial wastes: a comparison and optimization of models using ensemble learners[J]. Journal of Cleaner Production, 2021, 292: 126032.
[18] KANG M C, YOO D Y, GUPTA R. Machine learning-based prediction for compressive and flexural strengths of steel fiber-reinforced concrete[J]. Construction and Building Materials, 2021, 266: 121117.
[19] 马高, 刘康. 基于BP神经网络CFRP约束混凝土抗压强度预测[J]. 湖南大学学报(自然科学版), 2021, 48(9): 88-97.
MA G, LIU K. Prediction of compressive strength of CFRP-confined concrete columns based on BP neural network[J]. Journal of Hunan University (Natural Sciences), 2021, 48(9): 88-97 (in Chinese).
[20] NGUYEN K T, NGUYEN Q D, LE T A, et al. Analyzing the compressive strength of green fly ash based geopolymer concrete using experiment and machine learning approaches[J]. Construction and Building Materials, 2020, 247: 118581.
[21] 韩建军, 赵道松, 李建平. 基于BP神经网络的垃圾飞灰混凝土抗压强度预测模型[J]. 混凝土, 2022(9): 78-81.
HAN J J, ZHAO D S, LI J P. Prediction model of compressive strength of garbage fly ash concrete based on BP neural network[J]. Concrete, 2022(9): 78-81 (in Chinese).
[22] 徐潇航, 胡张莉, 刘加平, 等. 基于机器学习回归模型的三峡大坝混凝土强度预测[J]. 材料导报, 2023, 37(2): 45-53.
XU X H, HU Z L, LIU J P, et al. Concrete strength prediction of the Three Gorges Dam based on machine learning regression model[J]. Materials Reports, 2023, 37(2): 45-53 (in Chinese).
[23] 梁宁慧, 游秀菲, 曹郭俊, 等. 基于机器学习的高温后聚丙烯纤维混凝土强度预测[J]. 硅酸盐通报, 2021, 40(2): 455-464.
LIANG N H, YOU X F, CAO G J, et al. Strength prediction of mechanical properties of polypropylene fiber reinforced concrete after high temperature based on machine learning[J]. Bulletin of the Chinese Ceramic Society, 2021, 40(2): 455-464 (in Chinese).
[24] 韩斌, 吉坤, 胡亚飞, 等. ANN-PSO-GA模型在湿喷混凝土强度预测及配合比优化中的应用[J]. 采矿与安全工程学报, 2021, 38(3): 584-591.
HAN B, JI K, HU Y F, et al. Application of ANN-PSO-GA model in UCS prediction and mix proportion optimization of wet shotcrete[J]. Journal of Mining & Safety Engineering, 2021, 38(3): 584-591 (in Chinese).
[25] 汪声瑞, 胡畔, 陈思宝, 等. 基于耦合BAS-MLP混凝土抗压强度的预测[J/OL]. 建筑材料学报: 1-15 [2023-04-03]. http://kns.cnki.net/kcms/detail/31.1764.TU.20230308.1421.004.html.
WANG S R, HU P, CHEN S B, et al. Prediction of concrete compressive strength based on coupled BAS-MLP[J/OL]. Journal of Building Materials: 1-15 [2023-04-03]. http://kns.cnki.net/kcms/detail/31.1764.TU.20230308.1421.004.html (in Chinese).
[26] 李杨, 刘庆华, 郭天添. 基于CART-SVR模型的混凝土抗压强度预测研究[J]. 混凝土, 2022(8): 40-44.
LI Y, LIU Q H, GUO T T. Prediction method of concrete compressive strength based on CART-SVR[J]. Concrete, 2022(8): 40-44 (in Chinese).
[27] 吴中伟. 绿色高性能混凝土与科技创新[J]. 建筑材料学报, 1998, 1(1): 1-7.
WU Z W. Green high performance concrete and innovation[J]. Journal of Building Materials, 1998, 1(1): 1-7 (in Chinese).
[28] 马骏. 高速公路预制T梁用高性能混凝土性能研究[J]. 中国测试, 2021, 47(6): 124-130.
MA J. Research on properties of high performance concrete for precast T-beam of expressway[J]. China Measurement & Test, 2021, 47(6): 124-130 (in Chinese).
[29] 秦涛, 韩方玉, 光鉴淼, 等. 铁路桥梁用高性能混凝土的力学性能试验研究[J]. 混凝土, 2021(3): 134-136.
QIN T, HAN F Y, GUANG J M, et al. Experimental study on mechanical properties of high performance concrete for railway bridges[J]. Concrete, 2021(3): 134-136 (in Chinese).
[30] 武春丽. 两部门联合推广应用高性能混凝土鼓励绿色建筑、保障房、政府投资工程率先应用[J]. 混凝土, 2014(4): 151.
WU C L.The two departments jointly promote the application of high-performance concrete and encourage green buildings, affordable housing and government investment projects to take the lead[J]. Concrete, 2014(4): 151 (in Chinese).
[31] 胥悦. 高性能混凝土技术在工业建筑工程中的实际应用[J]. 工业建筑, 2021, 51(8): 256.
XU Y. Practical application of high performance concrete technology in industrial construction engineering[J]. Industrial Construction, 2021, 51(8): 256 (in Chinese).
[32] 谷坤鹏, 王成启. 海工高性能混凝土常用胶凝材料抗硫酸盐侵蚀性能研究[J]. 水运工程, 2010(12): 8-13.
GU K P, WANG C Q. Resistance to sulfate corrosion of commonly used cement pastes of high-performance concrete for marine engineering[J]. Port & Waterway Engineering, 2010(12): 8-13 (in Chinese).
[33] 王跃全, 薛文强. 洋山港口工程中高性能混凝土施工质量控制[J]. 水运工程, 2005(6): 91-95.
WANG Y Q, XUE W Q. Quality control of high-performance concrete coustruction in Yangshan Port engineering[J]. Port & Waterway Engineering, 2005(6): 91-95 (in Chinese).
[34] WOLPERT D H. Stacked generalization[J]. Neural Networks, 1992, 5(2): 241-259.
[35] CHEN T Q, GUESTRIN C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22 nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco, California, USA. New York: ACM, 2016: 785-794.
[36] PROKHORENKOVA L, GUSEV G, VOROBEV A, et al. CatBoost: unbiased boosting with categorical features[EB/OL]. 2017: arXiv: 1706.09516. https://arxiv.org/abs/1706.09516.
[37] BREIMAN L. Random forests[J]. Machine Language, 2001, 45(1): 5-32.
[38] WIDIASARI I R, NUGROHO L E, WIDYAWAN. Deep learning multilayer perceptron (MLP) for flood prediction model using wireless sensor network based hydrology time series data mining[C]//2017 International Conference on Innovative and Creative Information Technology (ICITech). Salatiga, Indonesia. IEEE, 2017: 1-5.
[39] LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California, USA. New York: ACM, 2017: 4768-4777.
[40] YEH I C. Modeling of strength of high-performance concrete using artificial neural networks[J]. Cement and Concrete Research, 1998, 28(12): 1797-1808.
[41] YEH I C. Modeling slump flow of concrete using second-order regressions and artificial neural networks[J]. Cement and Concrete Composites, 2007, 29(6): 474-480.