Bulanık Mantık ile Yoğurt Üretiminin Modellenmesi

Abstract

Yoghurt production was modeled in this study based on different incubation temperatures, inoculum ratio of starter culture and incubation times. Experimental yoghurts were produced in two replicates and incubation final pH values of 343 yoghurt samples were measured. Resultant pH values were used in fuzzy logic modeling system. Fuzzy logic modelling was conducted in two sections: fuzzy rules were set and membership function was generated in the first section and defuzzification was conducted in the second section. Three different fuzzy sets (triangle membership function) were used for fuzzification of incubation temperature, inoculum ratio of culture and incubation time values. Since there were 7 membership functions of input parameters, 343 (7 x 7 x 7) rows of rule were generated. Mamdani method was used to tabulate fuzzy rules. Three trapezoidal sections of membership functions generated for defuzzification were used and membership function values were determined with the use of weighted average method. Incubation final pH values of 343 samples were assessed in modeling study and model outputs were compared with the expert decisions. Matlab (R2016b) software was used to assess model performance and model general performance was calculated as 90.27%. For present model to be used in industrial scale, automated yoghurt production lines should be designed and put into service of food industry.