The prevalence of more lean ore,less rich ore and high smelting costs necessitates a substantial import of iron ore for China's iron and steel industry.However,occurrences such as using waste to replace ore and substituting substandard goods for high-quality ones during the importation process pose threats to both the ecological and economic security.Establishing an identification model for iron ore adulteration and promptly verifying the solid waste properties of iron ore are crucial for supporting risk regulation of imported iron ore,promoting trade facilitation,and protecting ecological environment security.This study focuses on samples of imported iron ores and iron-containing materials produced in the production process of steel mills in China.The elemental composition and content of the samples were determined using WDXRF without standard sample analysis,and the KNN classification algorithm is used to establish the identification model of iron ore and iron-containing materials.The tenfold cross-validation method is used to optimize the model parameters and evaluate the model's recognition ability.The model's precision,recall rate and F1 score all reached 1.0,respectively,and the model's recognition accuracy for validation samples is 100%.The WDXRF without standard sample analysis method has the advantage of simple pre-processing and good data stability.Combined with the KNN classification algorithm,this method can achieve fast and accurate identification of imported iron ore and iron-containing materials.
iron oreiron-containing materialX-ray fluorescence spectrometry(XRF)k-nearest neighbors(KNN)
NETL
NSTL
万方数据
