Systematic Analysis of Chemical Dangerous Product Leakage based on Large Data Technology

Abstract

Due to facility upsizing of the chemical system itself, the high process continuity and the complicated inter-parameter mechanism, the chemical system has its own characteristics in addition to the above 4V (Volume, Variety, Velocity, High Value) features: high dimensionality, strong non-linearity, unevenly distributed sample data, low signal to noise ratio. It is due to these unique features of chemical system that there have some difficulties in the analysis and mining of its big data from the traditional methods. In this paper, Chemical characteristics of system is constructed based on big data platform architecture, and uses the hybrid diagnostic identification algorithm, preidentification of abnormal state of chemical system that may arise, to prevent and avoid the effect of practical application. By analyzing the error based on exception identification method of single point parameter, we can see that the method of abnormal identification of chemical systems based on large data technology has high stability and reliability.