Skip to Main content Skip to Navigation
Theses

Etude de la torréfaction : modélisation et détermination du degré de torréfaction du café en temps réel

Abstract : In order to guarantee and optimize the quality of coffee roasting, it is important to control a large number of factors during the process. Today's, robust sensors and algorithmes are used to measure and on-line analize essential values such as color, surface, temperature, weight... In this work, roasting coffee has been investigated, a control strategy is applied to estimate coffee roasting quality, considering sensors-algorithmes based on real time air temperature. Coffee beans were roasted using hot air as heating medium. Bean temperature, weight, color and expansion of surface were measured on-line during the roasting. These experiences allow better understanding of the phenomena that appear during roasting. A dynamical model is proposed to predict coffee bean temperature and moisture during the roasting. Gray value and expansion kinetics of beans surface are estimed using artificial neural network (ANN), considering the time during the process. Recurrents ANN are used to estimate the bean gray value in the next time (t+1). The gray value estimed is the key factor to stop process. The roasting degree wished is obtained when the gray value reaches the optimal quality.
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00003699
Contributor : Ecole Agroparistech <>
Submitted on : Tuesday, May 13, 2008 - 8:00:00 AM
Last modification on : Saturday, June 6, 2020 - 11:15:09 PM
Long-term archiving on: : Friday, September 10, 2010 - 12:24:17 PM

Identifiers

  • HAL Id : pastel-00003699, version 1
  • PRODINRA : 246693

Collections

Citation

José Alfredo Hernández Pérez. Etude de la torréfaction : modélisation et détermination du degré de torréfaction du café en temps réel. Sciences du Vivant [q-bio]. ENSIA (AgroParisTech), 2002. Français. ⟨NNT : 2002EIAA0124⟩. ⟨pastel-00003699⟩

Share

Metrics

Record views

2114

Files downloads

1057