Study On Source Inversion Methods For River Chemical Spill And Application Software Development

At present, early warning and reponse to water pollutions has become a urgent issues and hot topics in academic in china. Establishing Early Warning/Emergency Response System can provide effectively assistant support for pollution response and disposal. However, the detailed contamination source information is the precondition of the system. In fact, it is difficult to know the information of pollution source when pollution occured. In order to meeting the need of real-time assessment and early warning of pollutions, reseaches on pollution source inversion problem were carried out. Meanwhile, investigation on pollution source inverse problem of surface-water are insufficient so far. This thesis proposed solution framework of surface water pollution source inverse problems, and focuses on three inversion algorithms. First, a novel inversion algorithm based on an optimization approach for river point pollution sources is developed. Mass transport and kinetics processes of the contaminants in surface waters are combined along with the discharge history. And other relative parameters are deduced under the scenario that singular source instantly discharges degradable and soluble chemicals into one-dimensional rivers. A series of numerical experiments on the hypothetic cases, analyzing inversion effects associated with ambient river flow rates, contaminant decay rates, monitoring sites setting, sampling data errors and time intervals between two groups of sampling. The synthetical relative error of result is under 10%. Results show that parameters calculated fit well with the real values. In addition, the algorithms have the advantages such as efficient sampling process, minimum data requirement as well as easy programming. It is worthwhile to utilize this method for emergency environmental management practices.Next, two inversion methods were established based on genetic algorithm coupled with water quality model and BP artifical network, for instantaneous discharge of multiple pollutions sources and continuous discharge of single pollution source, respectively. The numerical cases test the effectiveness and accuracy of each algorithm, the synthetical relative errors are all below 20%. And the dispersion coefficient is the main facor of the algorithm, but the relative is under 50%, which responder can accepted.In order to control the detriment of the pollutions, river pollution source identification software was developed by C# and C++ programming language with water quality models and database technologies.The software integrated the control section information, pollution information and geographic information to realized pollution source identification on computer, which will play an auxiliary function for effective control of river point source pollutions. Moreover, a series of tracer experiments were conducted in labotory, providing data for concrete applied analysis, the results were satisfied. Therefore, this software can provide strong technical support for river basin water management. We recommned that this decision tool can be used by other relative administrative agency in China.