改进自适应指数平滑法及其在基坑变形预测中的应用

燕俊松; 石泉彬; 刘如兵

doi:10.20265/j.cnki.issn.1007-2993.2024-0168

改进自适应指数平滑法及其在基坑变形预测中的应用

doi: 10.20265/j.cnki.issn.1007-2993.2024-0168

泰州职业技术学院建筑工程学院，江苏泰州　225300

基金项目: 江苏省发改委省级工程研究中心建设项目（JPERC2021-168）；江苏省高等职业教育高水平专业群建设项目（2020-80）

详细信息

作者简介:
燕俊松，男，1992年生，博士，讲师，主要从事地下工程风险管理研究。E-mail：yanjunsong@tzpc.edu.cn

中图分类号: TU473; TU196
计量
- 文章访问数: 41
- HTML全文浏览量: 14
- PDF下载量: 9
- 被引次数: 0
出版历程
- 收稿日期: 2024-04-16
- 修回日期: 2024-07-29
- 录用日期: 2024-10-29
- 刊出日期: 2025-10-10

Improved adaptive exponential smoothing method and its application evaluation in excavation deformation prediction

Institute of Architectural Engineering, Taizhou Polytechnic College, Taizhou 225300, Jiangsu, China

摘要

摘要: 对基坑开挖变形作出可靠预测有助于提高施工安全性。对现有自适应指数平滑法进行了改进，摒弃采用遍历法求解最优平滑系数α的传统思路，以均方误差（MSE）作为损失函数，推导了一次、二次和三次指数平滑模型的MSE对α的梯度表达式，进而采用梯度下降法求解最优α值。与原方法相比，改进方法在最优平滑模型选择及平滑系数优化方面展现出与原方法相当的性能，且在适当的超参数（尤其是学习率）设置情况下，寻优效率显著提升。在此基础上，基于1900组基坑变形序列的应用，较为系统地评估了改进自适应指数平滑法在基坑变形预测中的适用性和可靠性，并探讨了预测步长和训练序列长度对预测性能的影响，据此提出了基坑变形预测优化方案。
- 深基坑 /
- 变形预测 /
- 改进自适应指数平滑法 /
- 梯度下降法 /
- 应用评价
Abstract: Reliable prediction of excavation deformation contributes to improved construction safety. The existing adaptive exponential smoothing method has been improved by abandoning the traditional approach of solving for the optimal smoothing coefficient α through exhaustive search. Instead, mean squared error (MSE) was used as the loss function, and the gradient expressions of MSE with respect to α for single, double, and triple exponential smoothing models were derived. Subsequently, the gradient descent method is applied to solve for the optimal α value. Compared to the original method, the improved method demonstrates comparable performance in optimal model selection and coefficient optimization, and significantly enhances optimization efficiency with appropriate hyper-parameter (especially learning rate) settings. Based on the application to 1900 sets of excavation deformation sequences, the applicability and reliability of the improved adaptive exponential smoothing method in deformation prediction have been systematically evaluated, and the effects of prediction step size and training sequence length on prediction performance have been explored. Consequently, an optimized scheme for excavation deformation prediction is proposed.
- deep foundation pit /
- deformation prediction /
- improved adaptive exponential smoothing method /
- gradient descent algorithm /
- application evaluation

HTML全文

图 1 改进自适应指数平滑法流程图

下载: 全尺寸图片幻灯片

图 2 改进方法与原方法的平滑系数优化结果对比

下载: 全尺寸图片幻灯片

图 3 指数平滑模型的MSE-$ \alpha $曲线以及改进方法与原方法的平滑系数优化过程对比

下载: 全尺寸图片幻灯片

图 4 改进方法与原方法拟合训练数据的结果对比

下载: 全尺寸图片幻灯片

图 5 改进方法与原方法训练时间对比

下载: 全尺寸图片幻灯片

图 6 不同预测步长下的预测表现

下载: 全尺寸图片幻灯片

图 7 不同训练序列长度下的预测表现

下载: 全尺寸图片幻灯片

表 1 改进自适应指数平滑法预测评价指标

训练序列长度	预测步长/期	$ \mathrm{RMSE} $/mm	$ {P}_{2} $
10	1	0.8	0.98
	2	1.13	0.94
	3	1.59	0.82
	4	2.1	0.75
	5	2.7	0.66
15	1	0.72	0.99
	2	1.04	0.96
	3	1.46	0.86
	4	1.91	0.78
	5	2.43	0.69
20	1	0.72	0.99
	2	1.03	0.95
	3	1.41	0.86
	4	1.83	0.78
	5	2.31	0.7
25	1	0.69	0.99
	2	0.98	0.96
	3	1.33	0.87
	4	1.73	0.8
	5	2.18	0.71
30	1	0.7	0.99
	2	0.99	0.96
	3	1.31	0.87
	4	1.7	0.79
	5	2.08	0.71

下载: 导出CSV

参考文献(24)

[1]	周红波, 高文杰, 蔡来炳, 等. 基于WBS-RBS的地铁基坑故障树风险识别与分析[J]. 岩土力学, 2009, 30(9): 2703-2707,2726. (ZHOU H B, GAO W J, CAI L B, et al. Risk identification and analysis of subway foundation pit by using fault tree analysis method based on WBS-RBS[J]. Rock and Soil Mechanics, 2009, 30(9): 2703-2707,2726. (in Chinese) ZHOU H B, GAO W J, CAI L B, et al. Risk identification and analysis of subway foundation pit by using fault tree analysis method based on WBS-RBS[J]. Rock and Soil Mechanics, 2009, 30（9）: 2703-2707,2726. (in Chinese)
[2]	胡冬, 张小平. 基于灰色系统理论的基坑变形预测研究[J]. 地下空间与工程学报, 2009, 5(1): 74-78,168. (HU D, ZHANG X P. Research on predicting deformation of foundation pit based on grey system theory[J]. Chinese Journal of Underground Space and Engineering, 2009, 5(1): 74-78,168. (in Chinese) HU D, ZHANG X P. Research on predicting deformation of foundation pit based on grey system theory[J]. Chinese Journal of Underground Space and Engineering, 2009, 5（1）: 74-78,168. (in Chinese)
[3]	CUI D G, ZHU C Q, LI Q F, et al. Research on deformation prediction of foundation pit based on PSO-GM-BP model[J]. Advances in Civil Engineering, 2021, 2021: 8822929. doi: 10.1155/2021/8822929
[4]	LIU Q, YANG C Y, LIN L. Deformation prediction of a deep foundation pit based on the combination model of wavelet transform and gray BP neural network[J]. Mathematical Problems in Engineering, 2021, 2021: 2161254.
[5]	刘俊城, 谭勇, 张生杰. 地铁车站深基坑开挖变形智能多步预测方法[J]. 上海交通大学学报, 2024, 58(7): 1108-1117. (LIU J C, TAN Y, ZHANG S J. Multi-step prediction of excavation deformation of subway station based on intelligent algorithm[J]. Journal of Shanghai Jiao Tong University, 2024, 58(7): 1108-1117. (in Chinese) LIU J C, TAN Y, ZHANG S J. Multi-step prediction of excavation deformation of subway station based on intelligent algorithm[J]. Journal of Shanghai Jiao Tong University, 2024, 58（7）: 1108-1117. (in Chinese)
[6]	ZHAO H J, LIU W, SHI P X, et al. Spatiotemporal deep learning approach on estimation of diaphragm wall deformation induced by excavation[J]. Acta Geotechnica, 2021, 16(11): 3631-3645. doi: 10.1007/s11440-021-01264-z
[7]	许莉, 林鑫, 徐杨, 等. 深基坑地下连续墙支护结构侧向变形的改进指数平滑法预测[J]. 应用基础与工程科学学报, 2024, 32(4): 1094-1107. (XU L, LIN X, XU Y, et al. Prediction of lateral deformation of diaphragm wall support structures in the deep foundation pit by using improved exponential smoothing method[J]. Journal of Basic Science and Engineering, 2024, 32(4): 1094-1107. (in Chinese) XU L, LIN X, XU Y, et al. Prediction of lateral deformation of diaphragm wall support structures in the deep foundation pit by using improved exponential smoothing method[J]. Journal of Basic Science and Engineering, 2024, 32（4）: 1094-1107. (in Chinese)
[8]	苏勇, 高政国, 廖翌棋. 基于监测数据的几种地铁施工变形预测方法[J]. 岩土工程技术, 2009, 23(4): 172-177. (SU Y, GAO Z G, LIAO Y Q. Several deformation forecasting methods for metro construction based on monitoring data[J]. Geotechnical Engineering Technique, 2009, 23(4): 172-177. (in Chinese) SU Y, GAO Z G, LIAO Y Q. Several deformation forecasting methods for metro construction based on monitoring data[J]. Geotechnical Engineering Technique, 2009, 23（4）: 172-177. (in Chinese)
[9]	程胜. 基于KF-ES融合模型的地铁基坑沉降变形预测[J]. 测绘与空间地理信息, 2023, 46(10): 161-164,167. (CHENG S. Prediction of subway foundation pit settlement and deformation based on KF-ES fusion model[J]. Geomatics & Spatial Information Technology, 2023, 46(10): 161-164,167. (in Chinese) CHENG S. Prediction of subway foundation pit settlement and deformation based on KF-ES fusion model[J]. Geomatics & Spatial Information Technology, 2023, 46（10）: 161-164,167. (in Chinese)
[10]	吴德会. 动态指数平滑预测方法及其应用[J]. 系统管理学报, 2008, 17(2): 151-155. (WU D H. Dynamic exponential smoothing prediction method and its applications[J]. Journal of Systems & Management, 2008, 17(2): 151-155. (in Chinese) WU D H. Dynamic exponential smoothing prediction method and its applications[J]. Journal of Systems & Management, 2008, 17（2）: 151-155. (in Chinese)
[11]	叶德谦, 王毅. 一种对二次指数平滑预测中参数和初值的研究[J]. 燕山大学学报, 2008, 32(2): 173-175. (YE D Q, WANG Y. Study on choosing coefficient of secondary exponential smoothing and initial value[J]. Journal of Yanshan University, 2008, 32(2): 173-175. (in Chinese) YE D Q, WANG Y. Study on choosing coefficient of secondary exponential smoothing and initial value[J]. Journal of Yanshan University, 2008, 32（2）: 173-175. (in Chinese)
[12]	董杰, 李欣, 方运海, 等. 基于改进模糊综合-指数平滑法的地下水水质评价和预测[J]. 中国海洋大学学报, 2020, 50(1): 126-135. (DONG J, LI X, FANG Y H, et al. Evaluation and prediction of groundwater quality based on improved fuzzy synthesis-exponential smoothing[J]. Periodical of Ocean University of China, 2020, 50(1): 126-135. (in Chinese) DONG J, LI X, FANG Y H, et al. Evaluation and prediction of groundwater quality based on improved fuzzy synthesis-exponential smoothing[J]. Periodical of Ocean University of China, 2020, 50（1）: 126-135. (in Chinese)
[13]	刘造保, 徐卫亚, 张开普, 等. 基于改进指数平滑法的岩体边坡变形预测[J]. 河海大学学报(自然科学版), 2009, 37(3): 313-316. (LIU Z B, XU W Y, ZHANG K P, et al. Prediction of rock slope deformation based on optimized exponential smoothing method[J]. Journal of Hohai University (Natural Sciences), 2009, 37(3): 313-316. (in Chinese) LIU Z B, XU W Y, ZHANG K P, et al. Prediction of rock slope deformation based on optimized exponential smoothing method[J]. Journal of Hohai University (Natural Sciences), 2009, 37（3）: 313-316. (in Chinese)
[14]	宋芳, 张亚平, 张宗郁, 等. 基于遗传退火三次指数平滑的期货预测研究[J]. 计算机工程与设计, 2012, 33(5): 1963-1967. (SONG F, ZHANG Y P, ZHANG Z Y, et al. Futures forecasting with genetic/simulated annealing algorithm and cubic exponential smoothing[J]. Computer Engineering and Design, 2012, 33(5): 1963-1967. (in Chinese) SONG F, ZHANG Y P, ZHANG Z Y, et al. Futures forecasting with genetic/simulated annealing algorithm and cubic exponential smoothing[J]. Computer Engineering and Design, 2012, 33（5）: 1963-1967. (in Chinese)
[15]	高春雷, 安泰, 向兵, 等. 基于粒子群算法的指数平滑系数优化方法[J]. 西北大学学报(自然科学版), 2014, 44(3): 383-386. (GAO C L, AN T, XIANG B, et al. The exponential smoothing coefficient optimization method based on particle swarm algorithm[J]. Journal of Northwest University (Natural Science Edition), 2014, 44(3): 383-386. (in Chinese) GAO C L, AN T, XIANG B, et al. The exponential smoothing coefficient optimization method based on particle swarm algorithm[J]. Journal of Northwest University (Natural Science Edition), 2014, 44（3）: 383-386. (in Chinese)
[16]	LIU S G, SUN C Z, ZHOU H, et al. Intelligent prediction model (IPM) of foundation pit displacement based on extreme learning machine (ELM) and its application[J]. Processes, 2022, 10(5): 896. doi: 10.3390/pr10050896
[17]	MA Q W, LIU S H, FAN X Y, et al. A time series prediction model of foundation pit deformation based on empirical wavelet transform and NARX network[J]. Mathematics, 2020, 8(9): 1535. doi: 10.3390/math8091535
[18]	刘清龙, 吕颖慧, 秦磊, 等. 基于改进最小二乘支持向量机组合模型的深基坑沉降变形预测[J]. 济南大学学报(自然科学版), 2024, 38(1): 8-14. (LIU Q L, LÜ Y H, QIN L, et al. Deep foundation pit settlement deformation prediction based on improved least square support vector machine combined model[J]. Journal of University of Jinan (Science and Technology), 2024, 38(1): 8-14. (in Chinese) LIU Q L, LÜ Y H, QIN L, et al. Deep foundation pit settlement deformation prediction based on improved least square support vector machine combined model[J]. Journal of University of Jinan (Science and Technology), 2024, 38（1）: 8-14. (in Chinese)
[19]	岳建伟, 仲豪磊, 赵丽敏, 等. 深基坑变形的滚动预测方法[J]. 中国科技论文, 2022, 17(6): 639-646. (YUE J W, ZHONG H L, ZHAO L M, et al. Rolling prediction method of deep foundation pit deformation[J]. China Sciencepaper, 2022, 17(6): 639-646. (in Chinese) YUE J W, ZHONG H L, ZHAO L M, et al. Rolling prediction method of deep foundation pit deformation[J]. China Sciencepaper, 2022, 17（6）: 639-646. (in Chinese)
[20]	曹净, 丁文云, 赵党书, 等. 基于LSSVM-ARMA模型的基坑变形时间序列预测[J]. 岩土力学, 2014, 35(S2): 579-586. (CAO J, DING W Y, ZHAO D S, et al. Time series forecast of foundation pit deformation based on LSSVM-ARMA model[J]. Rock and Soil Mechanics, 2014, 35(S2): 579-586. (in Chinese) CAO J, DING W Y, ZHAO D S, et al. Time series forecast of foundation pit deformation based on LSSVM-ARMA model[J]. Rock and Soil Mechanics, 2014, 35（S2）: 579-586. (in Chinese)
[21]	马琳. 基于ICEEMD-ICA准则进行数据处理的基坑变形组合预测研究[J]. 地质与勘探, 2023, 59(5): 1074-1082. (MA L. Combined prediction of foundation pit deformation based on ICEEMD-ICA criterion for data processing[J]. Geology and Exploration, 2023, 59(5): 1074-1082. (in Chinese) MA L. Combined prediction of foundation pit deformation based on ICEEMD-ICA criterion for data processing[J]. Geology and Exploration, 2023, 59（5）: 1074-1082. (in Chinese)
[22]	张生杰, 谭勇. 基于LSTM算法的基坑变形预测[J]. 隧道建设(中英文), 2022, 42(1): 113-120. (ZHANG S J, TAN Y. Deformation prediction of foundation pit based on long short-term memory algorithm[J]. Tunnel Construction, 2022, 42(1): 113-120. (in Chinese) ZHANG S J, TAN Y. Deformation prediction of foundation pit based on long short-term memory algorithm[J]. Tunnel Construction, 2022, 42（1）: 113-120. (in Chinese)
[23]	钟正雄, 杨林德, 杨金松. 基坑变形的实时建模预报时序分析方法[J]. 工业建筑, 2000, 30(3): 1-3,42. (ZHONG Z X, YANG L D, YANG J S. Method of time series analysis of real-time modeling prediction for deep excavation deformation[J]. Industrial Construction, 2000, 30(3): 1-3,42. (in Chinese) doi: 10.3321/j.issn:1000-8993.2000.03.001 ZHONG Z X, YANG L D, YANG J S. Method of time series analysis of real-time modeling prediction for deep excavation deformation[J]. Industrial Construction, 2000, 30（3）: 1-3,42. (in Chinese) doi: 10.3321/j.issn:1000-8993.2000.03.001
[24]	王飞. 基坑变形组合预测分析及安全性评价[J]. 隧道建设(中英文), 2019, 39(2): 204-210. (WANG F. Combined prediction analysis and safety evaluation of foundation pit deformation[J]. Tunnel Construction, 2019, 39(2): 204-210. (in Chinese) WANG F. Combined prediction analysis and safety evaluation of foundation pit deformation[J]. Tunnel Construction, 2019, 39（2）: 204-210. (in Chinese)