Premium
An Improved CBR Model Based on Mechanistic Model Similarity for Predicting End Phosphorus Content in Dephosphorization Converter
Author(s) -
Feng Kai,
Xu Anjun,
He Dongfeng,
Wang Hongbing
Publication year - 2018
Publication title -
steel research international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.603
H-Index - 49
eISSN - 1869-344X
pISSN - 1611-3683
DOI - 10.1002/srin.201800063
Subject(s) - similarity (geometry) , phosphorus , computer science , euclidean distance , content (measure theory) , biological system , data mining , artificial intelligence , mathematics , chemistry , mathematical analysis , image (mathematics) , biology , organic chemistry
There has been a lack of reliable online measuring equipment on end phosphorus content in dedicated dephosphorization converter, which leads to the low control accuracy of end phosphorus content. A Case‐based Reasoning (CBR) method based on mechanistic model similarity is proposed to predict the end phosphorus content in dephosphorization converter. The correlation between the influencing factors and target as well as the similarity between influencing factors of different cases are considered comprehensively in the proposed method. Firstly, the mechanistic model of end phosphorus content in dephosphorization converter is established by using the slag phosphorus distribution ratio and Matter Conservation. Then, the similarity calculation equation of mechanistic model is designed. Lastly, the prediction result is obtained by bringing the mechanistic model of end phosphorus content into the designed similarity calculation method. CBR method based on mechanistic model similarity (CBR‐MM), mechanistic model, Back Propagation Neural Network (BPNN), CBR based on Euclidean Distance similarity (CBR‐ED), and CBR based on Grey Distance similarity (CBR‐GD) are used to predict end phosphorus content by applying practical production data. The result shows CBR‐MM has higher prediction accuracy compared to other methods.