| Water is the indispensable resource for the industrial production, agricultural development and human progress, while the human society is facing the water resources shortage. Growing water shortages and severe water pollution are plaguing the people's livelihood, and also the socio-economic sustainable development has become a major constraining factor. Chanzhi reservoir, the largest reservoir in Shandong peninsula, is an important water source in Qingdao. The reservoir water pollution, water quality monitoring and the water quality are widely concerned. At present, water quality warning based on geographic information system is a hotspot of the water quality warning research having a very broad application prospects. This topic to produce Chanzhi Reservoir as a research standpoint, investigating automatic monitoring of surface water systems, water quality assessment methods, water quality and water quality prediction method of early warning systems. Based on the above, the major research works as follows:(1) A situ water quality automatic detection system is designed in this paper, in which environmental monitoring data collector can link two kinds of multi-parameter water quality sensors is self-design. The collector with GM8125 Serial expansion, GPRS through mobile GPRS Wireless Communication Module transmission network will collect data to monitor the database server automatically; the same time, the server can control each point and each of the monitoring of instrument parameters. Monitoring parameters are temperature, salinity, depth, pH, turbidity, DO, ammonia nitrogen, nitrate nitrogen, chlorophyll-a. Automatic monitoring PC software could measure water quality parameters in real time to grasp the changes, with the exception of data and transmission error alarm function.(2) In this paper, support vector machine is applied in water quality evaluation, using the model to evaluate the water quality of Chanzhi reservoir, achieved some results, and the evaluation results can generally match the reservoir production capabilities. And compared with fuzzy comprehensive evaluation, the results are basically consistent. Thus see that support vector machine method in water quality evaluation has some prospect. But in the SVM training process, the training sample data contains uncertainties should be paid attentions, such as the results of 1-5 and 5-1 binary tree method should be consistent in theory, but in practice found Only 5-1 method results in line with the actual water feature. Thus SVM method as a new method of water quality evaluation needs further study.(3) This article uses the improved fuzzy comprehensive evaluation model, research findings were obviously reasonable. Comparative analysis of the evaluation model through extensive examination, evaluation results are more accurate and rigorous.(4) The paper uses BP artificial neural network to predict water quality, and the forecast from the forecast trend can reflect changes in water quality trends over time. Software automatically eliminates abnormal data, corrects the parameters, and the best prediction results in a mean relative error is 3.6%. Therefore, BP artificial neural network modeling of water quality prediction method has strong practicability.(5) Establishing a water quality evaluation and early warning systems based on Arc Engine. According to MS SQL Server database technology and combined with. Geodatabase, the system achieves the fundamental geographic information management, hydrological data management, quality assessment, water quality prediction and many other features. The system uses GIS technology component build efficient and stable platform to apply to the majority of surface water and basic water quality evaluation and the prediction of hydrological information management. In addition, the software has a good performance and easy visualization of information extraction and query capabilities to help the relevant departments for decision-makers to better manage and maintain water quality. |
PDF Full Text Request