Video Fault Diagnosis Based On Crystallization Quality Control Of Molten Magnesia Furnace

Due to the rapid development of modern industry and science technology,the application of industrial control equipment and computer technology in the process industry has been accelerated.As a result,the complexity of the system will also increase,the degree of automation will continue to increase,and the entire system will gradually become larger,more continuous and more intelligent.For such a huge system in the event of failure,the resulting loss is immeasurable.Therefore,safety issues have drawn much attention.With the extensive application of sensors and industrial field monitoring equipment,the video information of industrial processes has been greatly accumulated,so the fault diagnosis of industrial processes based on video data has drawn great attention.At the same time,with the rapid development of magnesium industry,the purity requirements of magnesium oxide are getting higher and higher,and improving the quality of magnesium oxide crystal is an important task facing the magnesium industry.Firstly,crystallization process of fused magnesia was studied,and the factors influencing the crystallization quality were analyzed.Starting from the establishment of nucleation model and the construction of growth model,ultimate goal was to improve the crystallization quality.Analyze the conditions that the crystal growth rate and nucleation quantity should meet,and control the crystal quality by adjusting the parameters of the nucleation model.Aimed at the industrial process of fused magnesia furnace,a fault diagnosis method based on video data was proposed.The method used rapid kernel independent source analysis as a data modeling method to detect the blast furnace of the magnesia furnace in time.Data extraction method adopted here is the information entropy image feature extraction method.Through the data modeling of the normal video frequency of magnesia furnace,online monitoring of faulty video of blast furnace was carried out and simulation was conducted from different angles to verify the effectiveness of the method.At the same time,we compare the FastKICA method with KICA and find out the advantages of FastKIC A for video data with obvious advantages in missing rate.