Soft-sensor Approach with Case-based Reasoning and Its Application in Grinding Process
Chai Tian-you · Kongzhi yu juece · 2006
A soft-sensor approach with case-based reasoning(CBR) is developed to deal with the problem that some key technical parameters cannot be measured directly online in complex industrial processes.The structure of case representation is composed of case time,case descriptors,case solution and case similarity.The improved k-nearest neighbor(k-NN) technique with multiple case similarity threshold computation is used in case retrieval.The methods based on static and dynamic similarity threshold are employed in case reuse.New strategies of case revision and case retaining are presented.A softsensor model for grinding particle size based on the proposed CBR is established for a grinding process in a mineral processing plant,and the successful industrial application shows the effectiveness and future prospect of the proposed method.