Application Of OLAP And Data-cube Compressing Technologies In The Process Control Of Aluminum Electrolysis

The size of the materialized views of a data-cube enlarges with the increase of the data amount, the dimensionality and the complicacy of the dimension hierarchy of the data-cube in data warehouse and, therefore, results in data explosion. To solve this problem, compressing technologies for data-cube emerge and become a significant branch of data-cube technologies.In the traditional aluminum electrolysis control systems, the ability of multidimensional data query and analysis is bad. Aiming to solve this problem, this dissertation introduced OLAP and data-cube technologies into the aluminum electrolysis control systems. Combined with the application of real data in aluminum electrolysis control systems, a thorough theoretical and experimental study of compressing technology for data-cube was made and the main contents and achievements were as follows:1. By introducing the technologies of OLAP and data-cube into the aluminum electrolysis control systems, a multidimensional query system (MQ system) was designed and developed. Main work in this dissertation was as follows: begining with the function design of the MQ system, the structure of the MQ system was brought up and each module of the MQ system was explained; a multidimensional data model of the instance database was designed with a star mode and the ETL process was realizedin a simple way; the multidimensional query of aluminum electrolytic process data and, further more, the visualization of the query results were realized with the index information in the data-cube structure. As a sub-system of the electrolytic process analyzing system in the control systems, the MQ system also supplied data for other sub-systems in the process analyzing system.2. Through the analysis and comparison of the strongpoints and shortcomings of the main data-cube compressing technologies used currently and the analysis of the characteristics of aluminum electrolytic process data, Dwarf algorithm was introduced into the MQ system, realizing the full materialization and compressing storage of the data-cube views.3. To solve the problem of poor query respose ability of the MQ system, the Dwarf data-cube index structure was expanded and, therefore, the query response time was shortened without increasing the storage burden and computing time.