Multi-dimensional Data Mining Technology And Its Application In Reservoir Water Quality

With the popularization and development of computer technology,all sorts of new theories,techniques and methods of hydrology has come out.Data mining technology is a hot pot in the database area.It is a significant work that how to effectively analyze and deal with and find useful information from these data and support the water quality management and decision.This paper takes the Biliuhe reservoir as an example coordinate the water sampling points,then using GIS geostatisties and the function of image display from the surface,lengthways and crosswise of the reservoir to directly show the level changes under the dynamic water level and its quality character,analyze the temporal and spatial variation characteristics during the process of water level decreasing and analyze its influence factor.The research results are as follows:(1)Water quality in reservoir has multi attribute with multi index and temporal variation.The change of water quality is not only the pollutant content,but also influenced by change of reservoir water level and space.According to the characteristics of multi index temporal and spatial variation,establish the method of water quality multidimensional data mining,put forward the dynamic monitoring scheme of sampling points,study sampling points and space coordinate transformation method,to provide support for the analysis and management of reservoir water quality.(2)The pollution in the reservoir migrate and transform with the movement of water,causing water factors differ in the distribution of time.Meanwhile,the different water quality data gathering from different time causes the difference of water factors in the distribution of time.The water quality difference not only is the temporal and spatial evolvement of water itself but also indicate the water level variation during the reservoir operation.So its boundary environment also in the movement.At the same time,the position of the sampling points on each vertical line is changed with the change of water level.The layout monitoring section and poins which combines the reservoir’s shape and hydrological characteristics,the continuous layered monitoring and the sample of water quality which obtain from the crosswise,lengthways and vertical three-dimension reservoir provide the reliable data to the follow-up research work.(3)With GIS geostatisties and the function of image display,from the surface,lengthways and crosswise of the reservoir analyze the spatiotemporal variation of the reservoir’s temperature,DO,total nitrogen,total phosphorus,pH,iron,manganese and ammonia nitrogen and its influence factors.As the result,from the surface point,the temperature reaches its highest point in summer,and lowest in winter.DO and total nitrogen,pH,iron and manganese uniformly distribute on the surface.Total phosphorus and ammonia nitrogen gradually decrease from entrance to dam.From the lengthways point,the temperature and DO in the vertical level will lay and differ with time while total nitrogen and total phosphorus won’t.Iron and manganese well distributed in time and space but on some months they will increase sharply.Ammonia nitrogen increase in the entrance.From the crosswise point,the distribution and variation of Liudian,library center and the cross section in the front of the dam are similar to those on the lengthways point.Some are not well distributed on some months.Because water on the cross section in the front of the dam is deeper than the other parts,it has a more obvious stratification while the library center with its wide surface and shallower water don’t has an obvious stratification among the three.This paper apply data mining technology to reservoir water quality multidimensional characteristic analysis,and use GIS to achieve the multidimensional data visualization of water quality,analyze and know the water quality condition,playing an a certain effect on water quality management and decision making.