基于逐步回归分析的峡江水库对吉安水位顶托作用量化分析  

Quantitative Analysis of the Uplift Effect of Xiajiang Reservoir onJi′an Water Level Based on Stepwise Regression Analysis

作　　者：何中政 钱伟[1,2] 傅春 蔡思宇 陈甜[1] HE Zhong-zheng;QIAN Wei;FU Chun;CAI Si-yu;CHEN Tian(School of Architectural Engineering,Nanchang University,Nanchang 330031,Jiangxi Province,China;Key Laboratory of Poyang Lake Environment and Resource Utilization,Ministry of Education,Nanchang University,Nanchang 330031,Jiangxi Province,China;School of Public Policy and Administration,Nanchang University,Nanchang 330031,Jiangxi Province,China;China Institute of Water Resources and Hydropower Research,Beijing 100048,China)

机构地区：[1]南昌大学工程建设学院,江西南昌330031 [2]南昌大学鄱阳湖环境与资源利用教育部重点实验室,江西南昌330031 [3]南昌大学公共政策与管理学院,江西南昌330031 [4]中国水利水电科学研究院,北京100048

出　　处：《中国农村水利水电》2024年第7期53-59,共7页China Rural Water and Hydropower

基　　金：国家自然科学基金项目(52209024,52009005);江西省水科院开放基金项目(2022SKSH01);江西省自然科学基金项目(20224BAB204075,20212BAB214065)。

摘　　要：峡江水库建设对赣江干流水位的顶托作用主要影响着上游吉安河段的水文情势变化,分析回水顶托作用下引起的水位变化以及解析河段水位抬升成因机制对研究水利工程修建改变河道水力特点的防洪减灾工作具有重要意义。基于历史实测水位和流量数据,构建计算水位顶托效果的逐步回归预测模型,利用皮尔逊相关分析定量评估水位顶托效果与各因素之间的相关性和影响程度,随后援用逐步回归方法量化分析了峡江水库运行对吉安水位顶托作用,即建立水位顶托效果与各因素之间的多元回归方程并验证方程的可靠性。结果表明:①基于逐步回归分析的预测模型模拟精度高,水位和流量要素模拟效果均良好,模拟水位过程与实测过程基本保持一致,在模拟流量方面也能够较好地再现实测过程且水位预测模型的各项评价指标优于流量预测模型;②水位顶托效果与峡江水库实测水位存在极强正相关性,与吉安实测水位和吉安水位落水过程存在中等正相关性,与吉安实测流量存在中等负相关性,与吉安水位涨水过程存在微弱负相关性;③多元逐步回归方程的修正后决定系数R2为0.892,表明方程拟合度较高且通过了后续的准确性检验,可表征吉安实测水位、吉安实测流量、峡江水库实测水位、吉安水位涨落过程的多重驱动作用。The jacking effect of Xiajiang Reservoir construction on the water level of the main stream of the Ganjiang River mainly affects the hydrological regime change of the upstream Jiʹan section. It is of great significance to analyze the water level changes caused by the jacking ef⁃fect of backwater and analyze the cause mechanism of water level rise in the river section to study the flood control and disaster reductionwork of water conservancy project construction to change the hydraulic characteristics of the river channel. Based on the historical measuredwater level and flow data, this paper constructs a stepwise regression prediction model to calculate the effect of water level jacking. Pearsoncorrelation analysis is used to quantitatively evaluate the correlation and influence degree between the effect of water level jacking and vari⁃ous factors. Then, the stepwise regression method is used to quantitatively analyze the effect of Xiajiang Reservoir operation on Jiʹan waterlevel jacking, that is, the multiple regression equation between the effect of water level jacking and various factors is established and the reli⁃ability of the equation is verified. The results show that: ① The prediction model based on stepwise regression analysis has high simulationaccuracy, and the simulation results of water level and flow factors are good. The simulated water level process is basically consistent with themeasured process. In terms of simulated flow, the measured process can also be well reproduced, and the evaluation indexes of the water lev⁃el prediction model are better than those of the flow prediction model. ② The effect of water level jacking has a strong positive correlationwith the measured water level of Xiajiang Reservoir, a moderate positive correlation with the measured water level of Ji ′an and the fallingprocess of Jiʹan water level, a moderate negative correlation with the measured flow of Jiʹan, and a weak negative correlation with the risingprocess of Jiʹan water level. ③ The modified

关 键 词：水位变化 顶托作用 逐步回归分析 皮尔逊相关分析 峡江水库 

分 类 号：TV697.1[水利工程—水利水电工程]