Published in: Proceedings Institut Teknologi Bandung, Lembaga Penelitian ITB, Volume 32, No. 1, pp. 411-417 NON-LINEAR PARAMETER ESTIMATION OF EMC/ERH CORRELATIONS FOR GRAIN-TYPE PRODUCTS J.P. Sitompul, Istadi (*) , and S. Sumardiono Dept. of Chemical Engineering, Institute of Technology Bandung, Jl. Ganesha 10, Bandung, 40132, Indonesia E-mail: johnner@termo.pauir.itb.ac.id ABSTRACT A mathematical prediction of corn sorption isotherm was developed for three-parameter EMC correlation such as Modified-Henderson, Modified-Chung-Pfost, Modified- Oswin, and Modified-Halsey. Moisture sorption experimental data were determined using a dynamic method at temperature 30 o C, 40 o C and water activity range from 0.07 to 0.75. The vapor pressure regulation of air surrounding the grains was adjusted by varying sulfuric acid solution. Non-linear regression algorithms such as Simplex and/or Rosenbrock algorithm were used to fit the measured data to the EMC relations. The generalized correlation coefficient (R 2 ) and the mean relative deviation were in the range of 0.90-0.99 and 2 % - 8 % respectively, in which comparison between predicted and experimental sorption isotherms. From the parameters of correlation, it is possible to get the best equations among them to represent equilibrium moisture content of corn grains. The Modified-Henderson and Modified-Oswin equations give the best model for predicting the adsorption and desorption EMC of corn grain for given range of temperature and water activity. NOTATION A,B,C parameters used in EMC relations A w water activity or relative humidity decimal R 2 correlation coefficient T absolute temperature K X * equilibrium moisture content % i X ˆ calculated data i X mean value of experimental data X i experimental data n number of experimental data (*) Permanent address: Dept. of Chemical Engineering, Diponegoro University, Jl. Prof. Sudarto, Semarang, Indonesia, 50239