| Water quality monitoring not only guards the social sustainable development and the safety of human existence but also has become an important foundation for water resources management and water environment protection. At present, the common water quality on-line monitoring methods mainly include chemical analysis, chromatographic analysis and spectral analysis. Among them, the chemical or chromatographic analysis is difficult to be used for monitoring and early warning of sudden water pollution accidents because of its single location detection and long testing time. Comparatively, the UV-Vis spectroscopy as one of the spectral analysis methods not only can directly or indirectly measure most of the metal ions, organic pollutants, COD and other parameters in water, but also has the advantages of simple structure, easy operation, and maintenance cost, etc. Based on above-mentioned reasons, the UV-Vis spectroscopy has been widely concerned.In view of this, we combined the UV-Vis spectroscopy with the distributed sensor network technology to realize the real-time online monitoring of the important water quality parameters, such as TURB, NO3-N, COD and TOC. This work was jointly supported by the Key Technology Research and Development Program of Sichuan Province(2012SZ0111) and Graduate Research and Applied Research Innovation Project of Chongqing(CYS14039). The data processing methods of the monitoring system, which was set up by our laboratory, had been studied. My main research contents include:(1) Study on the compression and the recovery of spectral data. In view of the problem of the large amount of data transmission in the multi node and remote monitoring system, the compressed sensing algorithm based on wavelet transform had been researched to solve the data loss problem. The influence of the measurement matrix and the reconstruction algorithm on the exact reconstruction had been studied as well as the maximum packet loss rate by using the proposed algorithm.(2) Study on a global calibration algorithm based on extinction methods. When the UV-Vis spectroscopy was used to detect the parameters in the natural water, the turbidity interference caused by suspended particles in the water would result in the overall nonlinear uplift of the spectral line and degraded accuracy. To solve the problem, a global calibration algorithm was developed. It used the part of the full band spectrum to inverse particle size distribution, then the whole band spectrum of suspended particulate matter was fitted by the particle size distribution. Finally, the spectrum of the suspended particulate matter was deducted from the original spectrum to realize the global calibration.(3) Study on the optimal wavelengths for light extinction spectroscopy measurement based on the genetic algorithm-particle swarm optimization. In the process of particle size distribution reconstruction, if the number of the measuring wavelengths was too large, this would take lot of memory and need more computing time, which were not conducive to make the rapid, real-time and simple measurement. Aiming at this problem, the optimal wavelengths extraction algorithm based on GAPSO was developed, and this method had important significance to improve the accuracy of particles size distribution inversion, reduce the complexity of the full band method and exclude subjectivity and empiricism during the selection of measuring wavelengths.Finally, the experiment was designed to verify the validity and stability of the proposed algorithm. |
PDF Full Text Request