基于机器学习对三角堰流量系数的预测研究

为了准确高效得出三角堰泄流系数的大小,本研究基于广义神经网络、最小二乘支持向量机和遗传规划对三角堰流量系数进行智能建模．将无量纲参数堰顶角、弗劳德数、堰顶长与堰高之比、堰顶长与堰顶水深之比作为模型输入参数,流量系数作为模型输出参数．研究表明,最小二乘支持向量机性能优于广义神经网络和遗传规划．在测试阶段,均方根误差为0.000 73,散射指数为0.001 02,决定系数为0.999 59,最大误差为1.7 %．该模型精度较高,预测值较准确,具有较强的工程适用性．本研究通过对三角堰流量系数进行智能建模,采用机器学习对非线性物理关系进行高精度拟合,准确高效得出三角堰流量系数的大小,可为水利水电工程设计人员和我国灌区精确量水用水提供一种新的方法和思路．

Research on Prediction of Discharge Coefficient of Triangle Weir Based on Machine Learning

In order to accurately and efficiently obtain the flow rate of the triangular weir, in this study an intelligently model of the discharge coefficient of the triangular weir is established based on the generalized neural network, the least square support vector machine and the genetic programming. The dimensionless parameters including the weir crest angle, Froude number, the ratio of the length of the weir to the height of the weir, and the ratio of the length of the weir to the water head of the weir are taken as the input parameters of the model, and the discharge coefficient is taken as the model output parameter. The results show that the performance of the least square support vector machine is better than the generalized neural network and the genetic programming. In the test phase, the root mean square error is 0.000 73, the scattering index is 0.001 02, the coefficient of determination is 0.999 59, and the maximum error is 1.7 %. The model has high accuracy, high prediction precision, and strong engineering applicability. This study is devoted to the intelligent modeling of triangular weir discharge coefficients, in which the high-precision fitting of nonlinear physical relationships is carried out through machine learning, and the discharge coefficient of the triangular weir is accurately and efficiently obtained. It provides a new method and idea for the designers of water conservancy and hydropower engineering and the irrigation regions in China to accurately measure water use.