Fuzzy hydrocyclone modelling for particle separation using fuzzy rule interpolation

Abstract

This paper reports on the use of a fuzzy rule interpolation technique for the modelling of hydrocyclones. Hydrocyclones are important equipment used for particle separation in mineral processing industry. Fuzzy rule based systems are useful in this application domains where direct control of the hydrocyclone parameters is desired. It has been reported that a rule extracting technique has been used to extract fuzzy rules from the inputoutput data. However, it is not uncommon that the available input-output data set does not cover the universe of discourse. This results in the generation of sparse fuzzy rule bases. This paper examines the use of an improved multidimensional fuzzy rule interpolation technique to enhance the prediction ability of the sparse fuzzy hydrocyclone model. Fuzzy rule interpolation is normally used to provide interpretations from observations for which there are no overlaps with the supports of existing rules in the rule base.