Net-based Study Of Industrial Process Visualization Data Warehouse

Now the production process automation system is widely employed in industrial enterprises. These automation systems are set up by varied manufacturers, and the historical data are deposited heterogeneous system, which can't be stored , used , and managed uniformly. They are wasted extremely. So, many problems haven't been sloved reasonably because being lack for plenty of data.This paper studies the relative theory and technical of the industrial process plant data warehouse. The three tier system structure is employed, C sharp program language is used to code, and the visual industrial process data warehouse is accomplished. To integrate and utilize the historical data resources, we mine the knowledge for production with technical background. Application environment is also discussed.First, the basic concept of data warehouse and application instance are introduced, the product capability similar partially with the project being researched in the paper are presented, the distributed application palntform .NET and three tier structure of software system are refered. On these basic, we put forward the system structure in the mass, the system consists of four big granularity components, the communications of components are based on .NET Remoting. Then, It is introduced that some important technique being used in the system, it includes how to extract data from other automatic system,to integrate data, to store data with deadband value , and to manage metadata. At last, it is presented that some visual analysis tools having been completed, the tools are used to analyze and mine multidimensional data. It is brought forward that the algorithm whose name is Mini-Check is used to mine the linearcorrelation of long time series data in the paper, which translate time series into the Haar wavelet coefficients and filter large amount time series utilizing the relation between the linear correlation coefficient and the Eucliden distance coefficient, at last, only Pearson correlation coefficient of the pair of time series need calculate. The efficiency of this method has enhanced being compared with the old method by experiment.If they have built the data warehouse in the paper, the engineers should get useful information, in virtue of the information , they can optimize technics flow and save production energy sources.