Research On Multi-sensor Data Fusion For On-line Measurement Of Suspended Sediment Concentration

In order to understand the regulation of river sediment flow,the change of river bed and channel scour and siltation,the situation of soil and water loss and the effect of control,the real-time detection of suspended sediment content has become an important part of hydrological test.However,there are some problems in the existing methods,such as the effective acquisition of sediment content information,the elimination of interference of environmental variables and the reasonable construction of multi-sensor data fusion model.At present,there is not a complete multi-sensor data fusion system for real-time detection of suspended sediment content.In this paper,aiming at the problems caused by environmental variables in sediment concentration detection and how to detect sediment concentration accurately,a real-time on-line detection method and a multi-sensor data fusion model are studied,which will be used for soil and water conservation in river basin.Water quality detection and drainage and sand discharge provide the corresponding reasonable basis.In the automatic monitoring system of sediment content,in addition to selecting the appropriate sensor,it must be able to adapt to the complex and changeable detection environment,as well as to ensure the accuracy of sediment content detection.With the development of related technology,the research of multi-sensor data fusion has gradually become one of the main directions of sediment content detection and data processing.Karamian filter theory,artificial neural network,genetic algorithm and so on can make a certain contribution to accurate detection of sediment content.Therefore,the study of suspended sediment content based on multi-sensor data fusion is of great significance.The main contents and innovations of this paper include:1.This paper discusses the methods and principles of suspended sediment content detection,summarizes the advantages and disadvantages of the existing methods,proposes a method based on audio resonance,and constructs a model based on Kalman temperaturefusion.Firstly,the different resonance frequencies of the audio resonance sensor are used to perceive the different suspended sediment content,and then the temperature signal is used as the control signal to reduce the measurement error by using the temperature signal as the control signal.2.A fusion model based on Kalman-GRBF neural network is proposed.The information collected by the sensor is normalized after Kalman filtering,and then the information value is processed by Kalman-GRBF fusion.In order to compare the processing effect of Kalman-GRBF multi-source data fusion model,under the same environment,Kalman-RBF GRBF,multivariate and univariate linear regression processing are also carried out,and the error analysis of suspended sediment content measurement is carried out.3.The hardware platform is designed,and the visualization platform based on LabVIEW is devised.Then,the Kalman-T fusion and the Kalman-GRBF coupling model processing are realized.Finally,the above research results are applied to the real life.Thus,the practicability and value of the designed system are verified,which provides great help for future research on soil and water conservation.