Font Size: a A A

Sequential Patterns Mining Application In Aluminium Electrolysis

Posted on:2014-02-12Degree:MasterType:Thesis
Country:ChinaCandidate:Z WangFull Text:PDF
GTID:2248330395998589Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
In aluminum electrolysis and the actual daily production process, a large amount of historical data is generated, including the measurement data, control data and laboratory data which are related to the status of the trough, daily data for decision making and process that are issued by personnel, in these data for decision making, the production instruction execution of outstanding technical personnel in the face of electrolysis condition different for a period of time in the future is reflected. How to dig out the way of decision-making process and technical personnel from these data, thus for the future intelligent management of electrolytic cell production, timely adjustment of cell condition, ensure the normal production of electrolytic tank, which has important application value.The electrolytic aluminum production lots of history data, especially by setting the voltage adjustment, the amount of aluminum, aluminum fluoride added production decision data structure are analyzed, studied various sequential pattern mining algorithm, designed and developed the decision data for production of electrolytic aluminium display, analysis, sequence pattern mining and application system.1、 The PrefixSpan and its improved algorithm are analysed, combining with the characteristics of decision data for production of electrolytic aluminium, using pseudo projection technology, namely the use of the index indicates the sequential pattern mining algorithm instead of the entity database, the decision data for production of electrolytic aluminium is mined, the speed of mining is greatly improved.2、 The frequent decision sequences mined from decision data of historical production in the (like a week) pattern are analysed, combining sequence pattern involves electrolysis cell (such as a month) changes of measurement data, the decision definition is established.3、 Sequence of decisions in line with the current status of the trough in the future time according to the electrolyzer recent practical measurement, testing and other historical data, sequence of decisions by choosing beforehand excavated, can be facilitatedly developed.4、 A suitable electrolytic cell production decision sequence analysis, mining and application system is developed.
Keywords/Search Tags:Sequential Pattern Mining, Aluminum Electrolysis, Decision Serialization
PDF Full Text Request
Related items