基于LS-SVM的侧堰泄流能力预测模型  

Prediction Model of Side Weir Discharge Capacity Based on LS-SVM

作　　者：李国栋[1] 沈桂莹 李珊珊[1] 陆庆楠 LI Guodong;SHEN Guiying;LI Shanshan;LU Qingnan(State Key Laboratory of Ecological Water Resources in Northwest Arid Area,Xi'an University of Technology,Xi'an 710048,China;Powerchina Huadong Engineering Corporation Limited,Hangzhou 311112,China)

机构地区：[1]西安理工大学,省部共建西北旱区生态水利国家重点实验室,陕西西安710048 [2]中国电建集团华东勘测设计研究院有限公司,浙江杭州311112

出　　处：《应用基础与工程科学学报》2023年第4期843-851,共9页Journal of Basic Science and Engineering

基　　金：国家自然科学基金项目(51579206,52079107)。

摘　　要：为了准确高效地得出矩形侧堰流量系数(C_(d)),首先设计矩形侧堰模型试验,得出6种不同流量工况下的流量系数试验值,利用MATLAB搭建不同核函数的最小二乘支持向量机(LS-SVM)模型,将影响C_(d)的各无量纲参数作为模型输入,C_(d)作为模型输出.研究表明:LS-SVM模型可用于矩形侧堰流量系数预测,且高斯核函数优于多项式核函数和线性核函数,在测试阶段最佳模型的性能指标平均绝对误差(MAE)为0.005,均方根误差(RMSE)为0.005,决定系数(R2)为0.966,表明该模型性能较好,精度较高,预测值较准确.本文提出了一种预测侧堰泄流能力的智能模型,讨论了不同无量纲参数对该模型的影响,验证了该模型的适用性,为类似水利工程提供参考依据,同时为解决复杂水力学问题提供新思路.In order to obtain the discharge coefficient(C_(d))of the rectangular side weir accurately and efficiently,this paper designed the rectangular side weir model experiment to obtain the experimental values of C_(d) for six different discharge conditions.The least squares support vector machine(LS-SVM)model with different kernel functions was developed using MATLAB,and the dimensionless parameters affecting C_(d) were used as model inputs and C_(d) as model outputs.It is shown that the LS-SVM model can be used to predict the C_(d) of rectangular side weirs.The mean absolute error(MAE),root mean square error(RMSE)and coefficient of determination(R2)are 0.005,0.005 and 0.966 respectively,which indicates that the model has better performance,higher accuracy and more accurate prediction.An intelligent model for predicting the discharge capacity of side weirs is proposed in this paper,and the influence of different dimensionless parameters on the model is discussed to verify the applicability of the model,which provides a reference basis for similar hydraulic engineering and also provides new ideas for solving complex hydraulics problems.

关 键 词：矩形侧堰 流量系数 核函数 LS-SVM 人工智能 机器学习 

分 类 号：TV135.2[水利工程—水力学及河流动力学]