The Study Of Detection Model For Sediment Concentration In The Yellow River Based On Multi-source And Multi-scale Data Fusion

Real-time information detection of sediment concentration and sediment distribution can provide decision support for soil conservation, water quality assessment, dredging and sediment regulation in the high suspended sediment concentration (HSSC) river, such as the Yellow River. However, in the HSSC river, the existing sediment concentration detection methods still have many problems in the respects of information acquisition, environmental factors, the model organic presentation and so on. Currently, there is no effective method of online detection for sediment concentration in the Yellow River and other HSSC rivers.With the development of water conservancy information technology, multi-scale and multi-source data fusion research has become one of the main directions for online detection and data processing in the Yellow River. Except for relying on the sensors to be selected, the automatic detection system for sediment concentration built in the Yellow River not only requires to meet the needs of sophisticated environment detection and intelligent measurement, but also to ensure the accuracy and stability of the detection system.These techniques and theories, such as modern sensor technology, multi-source data fusion theory, multi-scale analysis theory and water information visualization technology, can greatly enhance the level of sediment concentration detection. Therefore, the study of multi-source and multi-scale data fusion for sediment concentration detection will have important theoretical and practical value. In this paper, the depth researches on the sediment concentration detection based on multi-source and multi-scale fusion theory are carried out, and the main results and conclusions have been harvested and presented as follows.1. The existing detection methods and techniques of the sediment concentration in the Yellow River have been analyzed and summarized, and the environmental factors as well as physical nature of the HSSC under different environmental factors have also been studied, and then the theoretical foundation of multi-source and multi-scale data fusion for HSSC detection has further been discussed.2. According to the physical characteristics of HSSC, which has different resonant frequency with different HSSC, a fuse model of multi-scale sequential block Kalman-temperature based on capacitive differential pressure is established. In this model, the Kalman filter is used to organically fuse the information of sediment concentration and temperature, and then dynamically adjusted the wavelet decomposition level according to the measurement error. This method realizes multi-scale motion detection for the HSSC based on acoustic resonance method.3. The optimized design for the RBF neural networks by an improved genetic algorithm (IGA) is proposed. And a mathematical model with multi-sensor and multi-scale fusion based on IGA-RBF is designed, which effectively reduces the impact of environmental factors on the Yellow River through fuse temperature, depth and flow rate with the RBF network. In addition, the genetic algorithm is utilized to optimize the RBF neural network, which effectively reduces the influence of environmental factors on the sediment detection for the Yellow River.4. An online detection system is designed based on HSSC detection methods for the Yellow River, such as capacitive differential pressure method, audio resonance, ultrasound law and optoelectronic method. A multi-source and multi-scale fusion model based on Wavelet-Curvelet is proposed, which utilizes the multi-scale data fusion section weighted by scalars to fuse the trend information of each item and uses the Curvelet transform to extract sediment concentration varying with temperature direction. Finally, a multi-source and multi-scale inversion model based on the Wavelet-Curvelet is established, which reconstructs the containing information of sediment term trend and directions and achieves optimal multi-source and multi-scale fusion.5. Through the integrated application of ZigBee, GPRS and other wireless transmission technologies, the detection system are built up for HSSC in the Yellow River, based on technology of the Internet of things (IOT) and LabVIEW. This platform solves the key technology of multi-source heterogeneous data organization, management, efficient transmission and visualization, and realizes the HSSC detection functions, such as information display, data analysis, fusion, data storage and so on.6. The above research results have been applied in the Yellow River garden slobber hydrologic station for the HSSC online detection to test the accuracy and reliability of the model. These research results will provide theoretical basis and technical support for HSSC dynamic measurement.