Data Cleaning And Restoration Of Large-Scale Iron Ore Sintering

The steel industry is an important basic industry of the country and a pil-lar industry of national economic development,and its development level is an important manifestation of a country’s comprehensive national strength.Iron ore agglomeration is the first process in the modern steel smelting process.At present,domestic and foreign steel companies generally use two types of agglom-eration methods,sintering and pelletizing.Among them,sintering has always been the main method of iron ore agglomeration in China.Therefore,sinter-ing production has an important position in China’s iron and steel enterprises.The research on equipment,technology and related basic theories in the sintering process has always been the focus of metallurgy..With the rapid development of the sintering industry,the production process,technology and equipment have also made great progress,the automation level of the sintering process has also been greatly improved,and the large-scale and automation of the sintering equipment has higher requirements for the modeling and control of the sintering process.Modeling and control of the sintering process relies on a large amount of real-time business data,one of which is widely used is the temperature of the bellows,which fully reflects the combustion of materials on the trolley.However,there are often numerical errors in the collected bellows temperature data.Therefore,when analyzing and controlling the state of the sintering process,it is necessary to clean and repair the temperature data of the bellows collected during the sintering process.Based on the thermal process state during the sintering process,this paper models the sintering process and combines the real production experience and requirements on site to study and design a system for identifying and repair-ing the temperature data of the air box during the sintering process.By using real production data for testing,the algorithm described in this article can well identify abnormal temperature data,detect and locate errors in the data,and repair the erroneous data.Moreover,the algorithm described in this article runs fast,meeting the requirements of the algorithm and model running speed in the production environment.