Optimization of short-path distillation of ginger extraction pilot plant (GEPP) process using multivariate statistical process control

• Main Authors: Yunus, M. Y. M., Abdul Samad, Noor Asma Fazli, Harun, N., Siti Aishah, Mustafa
• Format: Research Report
• Language: English
• Published: 2007
• Subjects: TP Chemical technology
• Online Access: http://umpir.ump.edu.my/id/eprint/36516/1/Optimization%20of%20short-path%20distillation%20of%20ginger%20extraction%20pilot%20plant%20%28GEPP%29%20process%20using%20multivariate%20statistical%20process%20control.wm.pdf; http://umpir.ump.edu.my/id/eprint/36516/
• Institution: Universiti Malaysia Pahang

The research objectives are to extract vanilla oleoresins from vanilla beans using ultrasonic solvent extraction technique to define oleoresins yield produced and to develop Quality Control Charts for the process in the UMEPPS. The scopes covered studies on the influence of the variation of vanilla surface area, type of solvent and sonication or exposure time to the yield of vanilla oleoresins, the solvent to material ratio in vanilla solvent extraction, vanillin detection in the vanilla oleoresins using HPLC and development of the Polynomial Regression Models and Shewhart Control Charts. The lab data provided by the UMEPPS are then used for the development of Polynomial Regression Models and Shewhart Control Charts using MATLAB software tool as yield been the key process variable that is base for Statistical Process Control (SPC) technique. Ethanol appeared to be the best solvent in maximizing extraction yield of 52.8 wt% after 24 hours of ultrasonic solvent extraction when the grinded vanilla beans used. The employment of ultrasound to the solvent extraction gives a significant reduction in extraction time of 120 minutes and increasing maximum extraction yield of 39.7 wt% when using UMEPPS. The 6th degree and 5th degree Polynomial Regression Models is the best fitted models because the values of goodness of fit statistics are closest to the reference values. The Shewhart Control Charts for Experimental Data 1 indicated that the process in the UMEPPS is normally distributed however for Experimental Data 2; the X Bar Control Chart revealed that the process in the UMEPPS is out of control. However, the validation of the models cannot be done with other sets of experimental data from the UMEPPS.