摘要
水平式ADCP的误差主要有速脉动流速测量误差和横向流速实测区间的代表性误差。以北碚水文站为分析对象,为提高测验及数据拟合精度,采用静水装置的物理安装方法,选取紊动较小的水平式ADCP单元格段作为断面代表流速,利用Matlab小波工具箱对数据突变点进行剔除,并采用代表流速法和BP神经网络法拟合断面代表流速和流速仪法断面平均流速的相关关系。结果表明:静水装置固定安装能避免水流直接冲击水平式ADCP的换能器而减小振动,从而减小脉动误差。BP神经网络法的关系线拟合精度较代表流速法好。受水利工程蓄放水影响的低水水位过程线锯齿严重的情况,是由于流速仪法测次时间较长或测量时机掌握不佳,低水部分关系线精度较差,可用遥控船走航式ADCP代替流速仪法比测,重新单独分析低水部分的关系线精度。
The main errors of horizontal ADCP are velocity fluctuation error and representative error of lateral velocity layer.In this study,Beibei hydrological station is taken as the analysis object.In order to improve the measurement accuracy and data fitting accuracy,the static water device physical installation method is adopted,and the horizontal ADCP unit section with small turbulence is selected as the representative velocity of the section.The wavelet toolbox of Matlab is used to eliminate the data mutation points,the representative velocity method and BP neural network method are used to fit the representative velocity and average flow of the section,and the current meter method is used to fit the relationship between the section velocity and the flow velocity.The results show that the static water device can avoid the direct impact of water flow on the horizontal ADCP transducer,reduce the vibration and the pulsation error.The fitting accuracy of BP neural network method is better than that of representative velocity method.It is mainly affected by water storage and discharge of water conservancy projects,and the saw tooth phenomenon of low water level hydrograph is serious.The main reason is that the accuracy of low water level hydrograph is poor due to the length of time measured by current meter method or not in time.Therefore,remote control ship navigation ADCP can be used instead of current meter method for comparative measurement,and the accuracy of low water level hydrograph can be analyzed separately.
作者
黄幸
Huang Xing(School of Water Conservancy and Hydropower,Xi'an University of Technology,Xi'an 710048,Shaanxi)
出处
《陕西水利》
2021年第5期35-39,共5页
Shaanxi Water Resources