Study on seepage leakage evaluation model of pumped storage power station in karst area
-
摘要: 岩溶地区水库的渗漏问题受多重复杂因素影响，渗漏量定量预测评价一直是工程地质领域的难点问题。传统评价方法难以有效处理多源异质数据间的非线性关系，而机器学习算法凭借其强大的数据挖掘和模式识别能力，在水库渗漏评价中展现出显著优势。本研究构建了基于随机森林（RF）、人工神经网络（ANN）和支持向量机（SVM）的岩溶水库渗漏预测模型，通过对比分析发现三种模型均能取得较好的预测效果。其中，随机森林模型表现出最优的预测性能，在训练和验证阶段，其模拟值与实测值具有高度一致性，能够精确捕捉渗漏量的动态变化规律。该模型不仅预测精度高，且稳定性好，可作为岩溶地区水库渗漏评价的首选预测模型。优选出岩体渗漏性、断层发育程度、渗漏通道形态、水动力条件4个指标为水库渗漏的主要的影响因子。本研究为岩溶地区水库渗漏风险评估提供了新的技术手段和支撑。Abstract:
So far, there is still no good method to quantitatively predict and evaluate the seepage volume of karst reservoirs.The reason is that the factors influencing the leakage of karst reservoirs are numerous and very complex, such as being affected by the lithology of the strata, the degree of structural development, the type of groundwater, the development of karst, hydrodynamic conditions, and anti-seepage treatment measures, etc. Most of these factors are not completely quantitative, and even random and ambiguous, making it difficult to describe them with deterministic models .On the other hand, the various factors influencing leakage have complex cross-effects and dynamic effects of mutual influence and mutual restraint. There exist complex nonlinear relationships among these factors, which are far beyond the description of a single (group of) simple algebraic equation. Machine learning algorithms have the characteristics of high accuracy and stability in simulating complex groundwater runoff and assessing reservoir seepage, which have attracted the interest of many researchers in recent years.Machine learning algorithms are a type of algorithm that can automatically analyze and obtain patterns from data and use these patterns to predict unknown data. They can automatically select the features with the highest correlation to evaluation events and solve problems such as noise, missing values, and outliers in the data, thereby improving the quality and integrity of the data.It has obvious advantages in reducing the time required for data processing. Compared with other classic statistical methods, its ability to identify nonlinear patterns of input and output is more reliable.Since most of the factors influencing reservoir leakage are qualitative and interact with each other, there is a certain correlation. Therefore, machine learning algorithms can be adopted to study the intrinsic connection and regularity between these qualitative influencing factors and the seepage of karst reservoirs, establish mathematical models, achieve the transformation from qualitative variables to quantitative evaluation, and conduct quantitative evaluation of the seepage volume of karst reservoirs. In this study, a model for predicting the leakage of karst reservoirs based on random forest (RF), artificial neural network (ANN), and vector machine (SVM) was innovatively constructed. By comparing and analyzing the results, it was found that all three models could achieve good prediction results. In this regard, the random forest model exhibited the best prediction performance: its simulated values were highly consistent with the measured values during the training and validation phases, accurately the dynamic change law of the leakage volume. The model has high prediction accuracy and good stability, and can be used as the preferred prediction model for the evaluation of reservoir leakage in karst areas. Four indicators, namely rock mass permeability, fault development degree, seepage channel morphology and hydrodynamic conditions, were selected as the main influencing factors of reservoir seepage. This research result provides new technical means and support for the risk assessment of leakage in pumped storage power stations. -
图 1 随机森林模型训练样本实测与预测渗漏量拟合曲线
Figure 1. Random forest model training sample measured and predicted leakage amount fitting curve
图 2 随机森林模型各影响因子重要性排序
Figure 2. Ranking of the importance of each influencing factor of the random forest model
图 3 神经网络模型训练样本实测与预测渗漏量拟合曲线
Figure 3. The fitted curve of the measured and predicted leakage amount of the training sample of the neural network model
图 4 神经网络模型各影响因子重要性排序
Figure 4. Importance ranking of each influencing factor of the neural network model
图 5 支持向量机模型训练样本实测与预测渗漏量拟合曲线
Figure 5. The fitting curve of the measured and predicted leakage amount of the support vector machine model training sample
图 6 支持向量机模型各影响因子重要性排序
Figure 6. Importance ranking of each influencing factor of the support vector machine model
表 1 各因子皮尔逊相关系数
Table 1. Pearson correlation coefficients of each factor
x1 x2 x3 x4 x5 x6 x7 x8 x9 DF x1 1 0.18 0.124 0.209 0.116 0.076 0.191 0.182 −0.034 0.297* x2 0.180 1 0.363** 0.656** 0.574** 0.547** 0.651** 0.684** 0.413** 0.677** x3 0.124 0.363** 1 0.440** 0.288* 0.426** 0.421** 0.491** 0.276* 0.451** x4 0.209 0.656** 0.440** 1 0.649** 0.497** 0.741** 0.804** 0.406** 0.589** x5 0.116 0.574** 0.288* 0.649** 1 0.577** 0.638** 0.584** 0.531** 0.637** x6 0.076 0.547** 0.426** 0.497** 0.577** 1 0.570** 0.555** 0.562** 0.561** x7 0.191 0.651** 0.421** 0.741** 0.638** 0.570** 1 0.736** 0.719** 0.667** x8 0.182 0.684** 0.491** 0.804** 0.584** 0.555** 0.736** 1 0.414** 0.595** x9 −0.034 0.413** 0.276* 0.406** 0.531** 0.562** 0.719** 0.414** 1 0.458** DF 0.297* 0.677** 0.451** 0.589** 0.637** 0.561** 0.667** 0.595** 0.458** 1 注：**在0.01水平上显著相关；*在0.05水平上显著相关 
下载: 导出CSV
表 2 评价因子共线性分析
Table 2. Collinearity analysis of evaluation factors
因子 容差 方差膨胀因子VIF x1 0.898 1.114 x2 0.442 2.260 x3 0.708 1.411 x4 0.263 3.808 x5 0.446 2.242 x6 0.483 2.071 x7 0.200 4.989 x8 0.264 3.783 x9 0.356 2.809
下载: 导出CSV
表 3 随机森林模型验证样本实测与预测渗漏量对比
Table 3. Comparison of measured and predicted leakage amount of validation sample in random forest model
样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% 样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% Y1 0.20 0.35 5.43 Y22 0.15 0.35 7.25 Y2 1.00 0.35 23.55 Y23 0.01 0.35 12.32 Y7 0.00 6.33 229.35 Y28 0.00 0.35 12.68 Y8 3.00 3.15 5.43 Y32 5.00 6.33 48.19 Y13 1.00 0.35 23.55 Y38 26.00 25.40 21.74 Y16 1.80 0.35 52.54 Y49 0.50 0.35 5.43 Y19 0.21 0.35 5.07 Y56 0.20 0.35 5.43 Y20 0.70 0.35 12.68 Y58 4.40 2.43 71.38 平均值 33.87
下载: 导出CSV
表 4 神经网络模型验证样本实测与预测渗漏量对比
Table 4. Comparison of measured and predicted leakage amount of verification sample in neural network model
样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% 样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% Y1 0.20 0.28 2.90 Y22 0.15 0.38 8.33 Y2 1.00 0.52 17.39 Y23 0.01 0.38 13.40 Y7 0.00 0.55 19.92 Y28 0.00 0.30 10.87 Y8 3.00 0.21 101.06 Y32 5.00 6.11 40.21 Y13 1.00 0.38 22.46 Y38 26.00 20.83 187.28 Y16 1.80 0.43 49.63 Y49 0.50 0.28 7.97 Y19 0.21 0.43 7.97 Y56 0.20 0.46 9.42 Y20 0.70 0.21 17.75 Y58 4.40 0.87 127.87 平均值 40.27
下载: 导出CSV
表 5 支持向量机模型验证样本实测与预测渗漏量对比
Table 5. Comparison of measured and predicted leakage amount of support vector machine model verification sample
样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% 样本号 实测渗漏量/m3∙L−1 预测渗漏量/m3∙L−1 相对误差/% Y1 0.20 0.00 7.24 Y22 0.15 0.19 1.45 Y2 1.00 0.70 10.87 Y23 0.01 0.19 6.52 Y7 0.00 2.68 97.08 Y28 0.00 0.00 0.00 Y8 3.00 3.00 0.00 Y32 5.00 5.15 5.43 Y13 1.00 0.52 17.39 Y38 26.00 18.16 284.00 Y16 1.80 1.01 28.62 Y49 0.50 0.06 15.94 Y19 0.21 0.10 3.98 Y56 0.20 0.12 2.90 Y20 0.70 0.15 19.92 Y58 4.40 2.93 53.25 平均值 34.66
下载: 导出CSV
表 6 3种机器学习方法模型误差统计
Table 6. Model error statistics of 3 machine learning methods
方法 训练样本 验证样本 主要影响
因子决定系数
（R2）平均绝对误差
（MAE）均方根误差
（RMSE）决定系数
（R2）平均绝对误差
（MAE）均方根误差
（RMSE）随机森林 0.98 0.58 0.31 0.91 0.94 1.76 岩体渗漏性x7
断层发育程度x2
渗漏通道形态x8神经网络 0.87 1.18 1.81 0.79 1.40 2.16 断层发育程度x2
岩体渗漏性x7
水动力条件x5支持向量机 0.97 0.30 0.87 0.82 0.87 1.93 岩体渗漏性x7
断层发育程度x2
水动力条件x5
下载: 导出CSV
-
[1] 韩冬, 郑静, 周力, 崔正辉, 何国建. 中国抽水蓄能站点资源评价与分布特点[J]. 人民长江, 2025, 56(3): 1-9.HAN Dong, ZHENG Jing, ZHOU Li, CUI Zhenghui, HE Guojian. Comparison and selection of design schemes for water transmission and powergeneration system of pumped storage power stations[J]. Yangtze Rirev, 2025, 56(3): 1-9. [2] 曹建文, 杨海洋, 王喆, 赵良杰, 林玉山, 栾松, 李录娟. 湖北某抽水蓄能电站建设中地下河系统多元信息综合分析评价[J]. 中国岩溶, 2023, 42(4): 699-710. doi: 10.11932/karst20230406CAO Jianwen, YANG Haiyang, WANG Zhe, ZHAO Liangjie, LIN Yushan, LUAN Song, LI Lujuan. Comprehensive analysis and evaluation of the multi-information on the underground river system in construction of a pumped-storage hydropower station in Hubei[J]. Carsologica Sinica, 2023, 42(4): 699-710. doi: 10.11932/karst20230406 [3] 宋晨, 潘晓东, 刘天云, 罗飞, 曾洁, 彭聪, 程瑞瑞, 张文平. 云南省一把伞水库主封堵区岩溶渗漏条件与管道特征分析.[J]. 中国岩溶, 2024, 43（5）: 1088-1098.SONG Chen, PAN Xiaodong, LIU Yuntian, LUO Fei, ZENG Jie, PENG Cong, CHENG Ruirui, ZHANG Wenping. Analysis of karst leakage conditions and pipeline characteristics in the main blocked area of Yibasan reservoir, Yunnan.[J]. Carsologica Sinica, 2024, 43(5): 1088-1098. [4] 丁 瑜, 甘伟琪, 冯 磊, 颜英军, 马飞鹏. 秦岭某抽蓄电站水库区岩溶发育特征及透水性研究[J]. 中国岩溶, 2025, 44(1): 70-78.DING Yu, GAN Weiqi, FENG Lei, YAN Yingjun, MA Feipeng. Study on karst development characteristics and water permeability of a pumped- storage upper reservoir in the Qinling Mountains, Southern Shaanxi[J]. Carsologica Sinica, 2025, 44(1): 70-78. [5] 刘克勤. 某抽水蓄能电站岩溶地层库盆防渗研究与评价[J]. 福建建设科技, 2016(3): 9-10+31. doi: 10.3969/j.issn.1006-3943.2016.03.004LIU Keqin. Research and Assessment of lower reservoir leakage in Jurong Pumped Storage Power Station[J]. Fujian Construction Science & Technology, 2016(3): 9-10+31. doi: 10.3969/j.issn.1006-3943.2016.03.004 [6] 闵征辉, 朱义刚, 郭建华. 新疆某抽水蓄能电站下水库渗控分析[J]. 水利规划与设计, 2024(12): 85-90+108. doi: 10.3969/j.issn.1672-2469.2024.12.016MIN Zhenghui, ZHU Yigang, GUO Jianhua. Seepage control analysis of lower reservoir of a pumped storage power station in Xinjiang[J]. Water Resources Planning and Design, 2024(12): 85-90+108. doi: 10.3969/j.issn.1672-2469.2024.12.016 [7] 邹成杰. 天生桥二级水电站坝基岩体工程地质分类与评价[J]. 红水河, 1994, 13（4）: 49-53.ZOU Chengjie. Engineering geological classification and evaluation of dam foundation rock mass of Tianshengqiao second-stage hydropower station.[J]. Hongshui River, 1994, 13(4): 49-53. [8] 韩行瑞 . 岩溶水文地质学[M]. 科学出版社, 2015.HAN Xingrui. Karst hydrogeology[M]. Science Press, 2015. [9] 沈春勇. 水利水电工程岩溶勘察与处理[M]. 中国水利水电出版社, 2015. SHEN Chunyong. Karst investigation and treatment of water conservancy and hydropower projects [M]. 中国水利水电出版社, 2015. [10] 林兴超, 汪小刚, 陈文强, 任爱武. 高泉水库三维渗流数值模拟及坝基渗漏原因分析[J]. 中国水利水电科学研究院学报, 2013, 11(4): 260-265.LIN Xingchao, WANG Xiaogang, CHEN Wenqiang, REN Aiwu. Three dimensional numerical simulation of seepage for Gaoquan Hydropower Station andanalysis of reasons of foundation seepage[J]. Journal of China Institute of Water Resources and Hydropower Research, 2013, 11(4): 260-265. [11] 赵瑞, 许模. 水库岩溶渗漏及防渗研究综述[J]. 地下水, 2011, 33(2): 20-22.ZHAO Rui, XU Mo. Summary on reservoir karst seepage and anti-seepage research[J]. Ground Water, 2011, 33(2): 20-22. [12] 许模, 王士天. 工程水文地质学及其在水电工程中的应用概述[J]. 工程地质学报, 2003, 11(4): 396-401. doi: 10.3969/j.issn.1004-9665.2003.04.011XU Mo, WANG Shitian. A brief of engineering-hydrogeology and its application to hydroelectric power project[J]. Journal of Engineering Geology, 2003, 11(4): 396-401. doi: 10.3969/j.issn.1004-9665.2003.04.011 [13] 罗明明, 周宏, 郭绪磊, 陈乾龙, 齐凌轩, 况野. 峡口隧道间歇性岩溶涌突水过程及来源解析[J]. 地质科技通报, 2021, 40(6): 246-254.LUO Mingming, ZHOU Hong, GUO Xulei, CHEN Qianlong, QILingxuan, KUANG Ye. Processes and sources identification ofintermittent karst water inrush in Xiakou Tunnel[J]. Bulletin ofGeological Science and Technology, 2021, 40(6): 246-254. [14] 颜慧明, 常威, 季怀松, 邓争荣, 郭绪磊, 陈林, 黄琨. 黄陵背斜东北翼岩溶水系统特征及其对引调水隧洞工程的影响[J]. 地质科技通报, 2022, 41(5): 315-323.YAN Huiming, CHANG Wei, JI Huaisong, DENG Zhengrong, GUO Xulei, CHEN Lin, HUANG Kun. Characteristics of the karst water system on the northeast wing of the Huangling anticline and its impact on water diversion tunnel engineering[J]. Bulletin of Geological Science and Technology, 2022, 41(5): 315-323. [15] 曹贤发, 刘玉康, 刘之葵, 张炳晖. 基于强溶蚀带特征的地基岩溶发育程度评价方法[J]. 中国岩溶, 2020, 39(4): 577-583.CAO Xianfa, LIU Yukang, LIU Zhikui, ZHANG Binghui. Evaluation method of development degree based on features of intense dissolution layer[J]. Carsologica Sinica, 2020, 39(4): 577-583. [16] 王宇, 王梓溦. 岩溶地下水富集的地貌组合形态[J]. 中国岩溶, 2015, 34(4): 314-324. doi: 10.11932/karst20150402WANG Yu, WANG Ziwei. Patterns of karst geomorphologic combinations in areas with rich groundwater[J]. Carsologica Sinica, 2015, 34(4): 314-324. doi: 10.11932/karst20150402 [17] 贾立维, 赵克欣. 窄巷口水电站深岩溶, 高流速, 大流量防渗堵漏处理实践[C]//2013水利水电地基与基础工程技术——中国水利学会地基与基础工程专业委员会第12次全国学术会议论文集: 273-280.JIA Liwei, ZHAO Kexin. Seepage prevention and plugging treatment practice of Zhaixiangkou Hydropower Station with deep karst, high flow rate and large flow [C]//2013 Foundation and foundation engineering technology of water conservancy and hydropower—Proceedings of the 12th National Academic Conference of Foundation and Foundation Engineering Committee of China Water Conservancy Society: 273-280. [18] 张伟, 许季军, 陈劲松. 江口水电站坝址区三维渗流计算分析[J]. 长江科学院院报, 2002, 19(2): 31-33. doi: 10.3969/j.issn.1001-5485.2002.02.008ZHANG Wei, XU Jijun, CHEN jinsong. Analysis on three-dimensional FEM seepage of Jiangkou Hydroelectric Station's dam area[J]. Journal of Yangtze River Scientific Research Institute, 2002, 19(2): 31-33. doi: 10.3969/j.issn.1001-5485.2002.02.008 [19] 丁士群, 傅瑞金. 琅琊山抽水蓄能电站上水库副坝地段岩溶发育特征及渗漏分析[J]. 中国岩溶, 1997, 16(1): 74-81.DING Shiqun, FU Ruijin. Analysis on karst development and seepage in the auxiliary dam area of the upper reseroir of langyashan pumped storage hydroplant[J]. Carsologica Sinica, 1997, 16(1): 74-81. [20] 王家明, 杨淇淇, 李金龙. 广西桂林抽水蓄能电站上水库防渗方案综合比选[J]. 红水河, 2023, 42(6): 52-55.WANG Jiaming, YANG Qiqi, LI Jinlong. Comprehensive selection on seepage prevention schemes for upper reservoirof Guilin pumped storage power station in Guangxi[J]. Hongshui River, 2023, 42(6): 52-55. [21] 李权, 党发宁, 毛钟毓. 某抽水蓄能电站上水库防渗方案的比选研究[J]. 甘肃水利水电技术, 2014, 50(8): 7-9+50.LI Quan, DANG Faning, MAO Zhongyu, Comparative study on seepage control schemes of upper reservoir of a pumped storage power station[J]. Gansu Water Resources and Hydropower Technology, 2014, 50(8): 7-9+50. [22] 赵国斌, 王汉斌. 可溶岩地区抽水蓄能电站主要工程地质问题探讨[J]. 水力发电, 2024, 50(7): 24-30. doi: 10.3969/j.issn.0559-9342.2024.07.005ZHAO Guobin. Discussion on Major Engineering Geological Problems of Pumped-storage Power Stations in Soluble Rock Areas[J]. Water Power, 2024, 50(7): 24-30. doi: 10.3969/j.issn.0559-9342.2024.07.005 [23] 肖万春. 水库岩溶渗漏勘察技术要点与方法研究[J]. 水力发电, 2008, 34(7): 52-55. doi: 10.3969/j.issn.1006-2890.2023.07.066XIAO Wanchun. Research on Technical Points and Methods of Reservoir KarstSeepage Investigation[J]. Water Power, 2008, 34(7): 52-55. doi: 10.3969/j.issn.1006-2890.2023.07.066 [24] 李芳涛, 李华明, 胡志平, 陈南南, 晏长根. 峨汉高速廖山隧道岩溶发育规律及其工程效应浅析[J]. 中国岩溶, 2020, 39(4): 592-603. doi: 10.11932/karst20200415LI Fangtao, LI Huaming, HU Zhiping, CHEN Nannan, YAN Changgen. Features of karst development and geotechnical effects in the Liaoshan tunnel on the E-Han expressway[J]. Carsologica Sinica, 2020, 39(4): 592-603. doi: 10.11932/karst20200415 [25] 覃绍媛. 黑滩河水库多层复杂岩溶水系统及水库防渗研究[D]. 成都理工大学, 2020.QIN Shaoyuan. Research on multi-layered complex karst water system andreservoir anti-seepage in heitanhe reservoir [26] 能昌信, 张慧敏, 孙晓晨, 徐亚, 刘景财, 刘玉强. 机器学习模型精细表征填埋场渗漏风险及其不确定性[J]. 中国环境科学, 2025, 45(9): 4986-4996. doi: 10.3969/j.issn.1000-6923.2025.09.024NAI Changxin, ZHANG Huimin, SUN Xiaochen, XU Ya, LIU Jingcai, LIU Yuqiang. Machine learning models for fine characterization of landfill leakage and uncertainty analysis[J]. China Environmental Science, 2025, 45(9): 4986-4996. doi: 10.3969/j.issn.1000-6923.2025.09.024 [27] 伍仪霖, 周子勇. 运用机器学习方法提取Sentinel-2影像烃微渗漏异常[J]. 地球物理学报, 2024, 67(4): 1330-1341. doi: 10.6038/cjg2023Q0773WUYilin, ZHOU, Ziyong. Extraction of hydrocarbon microseepage information from sentinel-2 remotesensing image using machine learning[J]. CHINESE JOURNAL OF GEOPHYSICS, 2024, 67(4): 1330-1341. doi: 10.6038/cjg2023Q0773 [28] 赵立冬, 李振, 刘罡. 人工神经网络模型在贾庄闸枢纽渗漏量计算中的应用[J]. 水文, 2009, 29(2): 64-65.ZHAO Lidong, LI Zhen, LIU Gang. Application of artificial neural network model in seepage calculation for jiazhuang gate[J]. JOURNAL OF CHINA HYDROLOGY, 2009, 29(2): 64-65. [29] 綦娅, 陈兴帅, 褚学伟. 基于MATLAB BP神经网络的岩溶渗漏污染预测一例[J]. 工程勘察, 2010, 38(10): 41-45+56.QI Ya, CHENG Xingshuai, CHU Xuewei. Prediction of pollution of karst leakage in Baiji ardealite site with BP neural network system[J]. Geotechnical Investigation & Surveying, 2010, 38(10): 41-45+56. [30] 程旋谟, 胡伟伟. 基于WOA-RF的隧道渗漏预测模型[C]//2024中国自动化大会论文集. 2024.CHEN Xuanmo, HU Weiwei. A Tunnel leakage prediction model based on WOA-RF algorithm[C]//2024 Proceedings of China Automation Conference. 2024. [31] 光耀华. 岩溶水库渗漏预测的地质-数学模型数量化理论在岩溶方面的应用[J]. 中国岩溶, 1992, 11(3): 181-197.GUANG Yaohua. The geologic-mathematical models for predicting leakage from karst reservoirs application of quantitative theory in karst[J]. Carsologica Sinica, 1992, 11(3): 181-197. [32] 毛健全, 李景阳, 顾悦, 任仲魁, 肖万春. 岩溶水库渗漏的数学-地质模型—逻辑信息法在岩溶水库渗漏评价中的应用[J]. 中国岩溶, 1998, 7(1): 45-55.Mao Jianquan, Li Jingyang, Gu Yue, Ren Zhongkui, Xiao Wanchun. Mathematical model of karst reservoir leakage-application of logic information methods to karst reservoir leakage[J]. Carsologica Sinica, 1998, 7(1): 45-55. [33] 刘坚, 李树林, 陈涛. 基于优化随机森林模型的滑坡易发性评价[J]. 武汉大学学报: 信息科学版, 2018, 43(7): 7.LIU Jian, LI Shulin, CHEN Tao. Evaluation of landslide susceptibility based on optimized random forest model[J]. Journal of Wuhan University: Information Science Edition, 2018, 43(7): 7. [34] 林荣福, 刘纪平, 徐胜华. 随机森林赋权信息量的滑坡易发性评价方法[J]. 测绘科学, 2020, 45(12): 131-138. doi: 10.16251/j.cnki.1009-2307.2020.12.020LIN Rongfu, LIU Jiping, XU Shenghua. Evaluation method of landslide susceptibility based on random forest weighted information[J]. Science of Surveying and Mapping, 2020, 45(12): 131-138 . doi: 10.16251/j.cnki.1009-2307.2020.12.020 [35] 杨桂芳, 姚长宏, 王增银, 万军伟. BP神经网络在岩溶水库渗漏评价中的应用[J]. 中国岩溶, 2000, 19(1): 73-80. doi: 10.3969/j.issn.1001-4810.2000.01.011YANG Guifang, YAO Changhong, WANG Zhengyin, WAN Junwei. Application of back propagation artificial neural network in leakage evaluation for karst reservoirs[J]. Carsologica Sinica, 2000, 19(1): 73-80. doi: 10.3969/j.issn.1001-4810.2000.01.011 [36] 张强, 黄生志, 陈晓宏. 基于支持向量机的土壤湿度模拟及预测研究[J]. 土壤学报, 2013, 50(1): 59-67.ZHANG Qiang, HUANG Shengzhi, CHEN Xiaohong. Simulation and prediction of soil moisture based on support vector machine technology[J]. Acta Pedologica Sinica, 2013, 50(1): 59-67. -
点击查看大图
图(6) / 表(6)
计量
- 文章访问数: 87
- HTML浏览量: 40
- PDF下载量: 57
- 被引次数: 0
