Estimation of suspended sediment concentration based on ADCP

Abstract

ADCP (Acoustic Doppler Current Profiler) is a kind of acoustic instruments which works according to the Principle of Doppler effect. Now, ADCP is widely used in the measurement of current velocity and flow of rivers and oceans. In fact, ADCP is not only limited to measure the current velocity and flow, as the data acquired through ADCP includes the acoustic backscattering signal which is related to the suspended sediment in the water. Based on this, a relationship between the backscattering intensity and suspended sediment concentration can be established, and then the concentration of suspended sediment can be estimated using the information of backscattering intensity. Compared with the traditional method, this approach has many advantages, and is worthy of further research. In this paper, the principle of estimating suspended sediment concentration from acoustic backscattering measured by ADCP has been studied systematically. Firstly, the basic principle of estimating the concentration of suspended sediment is detailed, and some main influence factors are analyzed. Secondly, on the basis of the above study, since the backscatter intensity is affected by many factors, there is a strong complex nonlinear relationship between the backscattering intensity and suspended sediment concentration. The Support Vector Machine (SVM) method is studied and applied in the estimation of suspended sediment concentration. The establishment of the network, training algorithm, and so on are researched in detail. Finally, an adaptable SVM model is established, its accuracy and generalization ability are also verified through simulation.