Available online at www.sciencedirect.

com

Procedia Food Science

www.elsevier.com/locate/procedia

Procedia Food Science 00 (2011) 000000

Procedia Food Science 1 (2011) 742 746

11th International Congress of Engineering and Food (ICEF11)

Numerical model of heat and mass transfer during roasting coffee using 3D digitised geometry
Angelo Fabbria, Chiara Cevolia, Santina Romanib, *a, Marco Dalla Rosab
a

Agricultural Economics and Engineering Department, University of Bologna, Piazza G. Goidanich, 60 - 47521 Cesena (FC), Italy (angelo.fabbri@unibo.it; chiara.cevoli3@unibo.it) b Food Science Department University of Bologna, Piazza G. Goidanich, 60 - 47521 Cesena (FC), Italy (santina.romani2@unibo.it;marco.dallarosa@unibo.it)

Abstract Nowadays roasters are equipped with electric control board, thermo-regulator device, display for temperatures and on-line sensors to detect defects and/or the roasting end, depending on coffee types and blends and the desired roasting degree. It is important to have a good control of the roasting process in order to obtain a high quality final beverage. In this work a numerical model, based on a 3D digitized geometry, able to describe the heat and moisture transfer inside a coffee bean during the roasting process, was developed. The model makes reference to a rotating cylinder roaster in natural convection conditions. For the multiphysics model development heat and mass transfer equations inside the coffee bean were numerically solved using a finite element technique. To validate the numerical model, green coffee bean was singularly roasted using a rotating drum roasters prototype. The calculated moisture concentration and time-temperature curves were then compared with the observed data. The calculated temperature values, in the centre of the bean, appear to be in good agreement with those measured inserting thermocouples into the coffee bean (RMSE 6C). Similarly the calculated volume averaged moisture was in good agreement with the experimental data (RMSE 265 mol/m3) over the entire time span. This model can be useful to optimise the roasting process control.
2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of 11th International Congress 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of ICEF11 on Engineering and Food (ICEF 11) Executive Committee.

Executive Committee Members.

Keywords: Finite element; coffee; roasting; 3D geometry

* Corresponding author. Tel.:+30-0547-636120; fax: +39-0547-382348. E-mail address: santina.romani2@unibo.it.

2211601X 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of 11th International Congress on Engineering and Food (ICEF 11) Executive Committee. doi:10.1016/j.profoo.2011.09.112

Angelo Fabbri al. //Procedia Science 1 00 (2011) 742 746 Authoret name Procedia Food Food Science (2011) 000000

743

1. Introduction Roasting is a crucial stage in the coffee processing, that is aimed to change markedly the chemical, physical, structural and sensorial properties of the green beans by heat induced reactions. In this way roasting process makes coffee beans suitable for brewing. The green beans are, in fact, characterized by only a weak and greenly aroma and a hard texture hinders their use as food. Roasting is an intense thermal treatment during which coffee beans are heated at high temperatures (160-240C) for times ranging between 8 and 20 min depending on the desired characteristics of the final product. In the case of an Italian-style coffee, the roasting temperature ranges between 200 and 240C [1]. Roasting is a complex process involving both energy (from the roaster to the bean) and mass (water vapor and volatile compounds from the bean to the environment) transfer implied in the main changes of the coffee beans in terms of weight, density, moisture, color and flavor [2; 3]. Endothermic process and reactions (water evaporation) occur in the first stages of the roasting, whilst the undesirable exothermic pyrolysis of saccharides may occur at the latter roasting stages. Process efficiency and quality of the roasted coffee depend on several factors including: gas composition and temperature, pressure, time, relative velocity of beans and gas flow rate. Each roasting equipment needs a proper choice of the process parameters as they affect the rate of heat transfer to the coffee bean and thus the development of the heat induced reactions during roasting [4]. It is important to have a good on-line estimation and prediction of the roasting process in order to obtain a high quality final beverage. For the process optimization the moisture and temperature evolution in coffee beans during the roasting process can be studied by modeling. Limited data have been found to describe simultaneous heat and mass transfer during coffee roasting. Schwartzberg [5] developed a semi-physical model to evaluate coffee bean temperature and moisture content during the roasting in batch system. This model gives predictions for the average temperature of coffee, neglecting the real internal distribution. Hernandez et al. [6], using the equations proposed by [5] carried out a theoretical analysis of heat and mass transfer to evaluate the coffee beans temperature and moisture content during the roasting in bed batch system including the internal distribution. Heyd et al. [7] improved the model of Hernandez et al. [6], particularly as it regards a more detailed physical description of the moisture transfer. Basile and Kikic, [8] developed a model found on a lumped specific heat capacity approach for predicting the non-stationary thermal profile of coffee roasting, validated by the experimental data found by the previous authors. In this work the process conditions of both fluidized bed and drum roasting were considered. All these models had dealt with simplified geometry, particularly with spheres or semi-ellipsoids. The aim of this work was to develop a numerical model, based on a 3D digitized geometry, able to describe the heat and moisture transfer inside the bean, during the coffee roasting process. The model makes reference to a rotating cylinder roaster in natural convection conditions, and was then validated with experimental data obtained by a lab scale roaster. This model could be useful to optimize the roasting process control. 2. Materials and Methods Commercial and well tested computational multiphysics codes are available nowadays, allowing an exceedingly flexible simultaneous numerical solution of the energy, mass and moment equation and are extensively adopted in the food engineering field The equations regarding heat and mass transfer inside the coffee bean during roasting were solved using Comsol Multiphysics 3.5 (COMSOL Inc., Burlington, MA, USA), a commercial partial differential equations solver based on finite element technique. During the roasting of coffee beans inside a small scale plant, heat is transferred mainly by natural convection from air to the product surface, and by conduction from surface toward the bean centre. Meanwhile moisture diffuses from the inside, towards the bean surface, and then vaporises. To simulate simultaneous heat and moisture transfer in bean during roasting, the following assumptions were made: a) moisture diffuses towards the product surface, and evaporates only at the surface;

744

Angelo et al. / Procedia Food Science 1 (2011) 742 746 Author Fabbri name / Procedia Food Science 00 (2011) 000000

b) the bean material is homogeneous and isotropic; c) the initial moisture and temperature of the bean are uniform; d) the heat produced by exothermic reactions during roasting is neglected; e) the coffee volume is constant. The geometry of the model was built acquiring the real geometry of an Arabica coffee bean using a 3D scanner (Roland Picza Pix-30, Roland DGA Corporation,USA). An unstructured mesh with tetrahedral elements of appropriate size was generated in Comsol Multiphysics. To validate the numerical model, green coffee bean was singularly roasted at 200C for a time of 10 min using a rotating drum roasters prototype (EXPO 500/E, STA impianti, Bologna, Italy). The internal temperature of the bean and the air temperature were measured by thermocouples type K (Chromel/Alumel; Tersid Came, Italia) with a precision of 0.5C. One bean was drilled in order to insert a 3 mm diameter thermocouple and was placed in the drum centre. The experiments were repeated five times. The calculated time-temperature curves were then compared with the observed data. Subsequently samples of 3 g of Arabica green coffee bens were roasted at 200C in the drum centre for 2, 4, 6, 8, 10 and 14 min to measure the moisture content [9] at different roasting stage. The moisture measurement was repeated three times. The calculated moisture concentration was then compared to the observed data. 3. Results and Discussion Figure 1 shows the temperature field of the coffee beans external surface, obtained by simulation carried out for 180 s at 200C. The temperature value is represented by the gradation of colours spanning from 176 to 180C. In particular in the figure the bean zones near the corner, are the first to equilibrate their temperature with that of roasting gas.

Fig. 1. Temperature field calculated on the bean coffee surface after 180 s of heat treatment at 200C.

In Figure 2 the time-temperature curves calculated in two bean points, respectively at the surface and in the bean centre, are shown. It can be observed that the lines are very close, especially after 300 seconds of treatment, when the temperature values are stabilised around at 200C. Before this time the surface of the bean, as expected, reached more quickly the highest temperatures; particularly, the calculated mean difference of temperature between the surface and the centre until 300 s was of about 14C. The internal bean temperature has an exponential behaviour similar to that described by Schwartzberg [5] when bean temperature is below 250C.

AngeloAuthor Fabbri et al./ / Procedia 1 (2011) 746 name Procedia Food Food Science Science 00 (2011) 742 000000

745

Fig. 2. The time-temperature curve calculated in two bean points during 600 seconds of roasting at 200C.

In Figure 3a the calculated and observed time-temperature curves for a 600 seconds treatment at 200C are reported. The simulated and experimental values appear to be in good agreement (RMSE 6C). The maximum observed temperature is reached after about 300 seconds of treatment as found by Hernandez et al. [6] and Heyd et al. [7] in similar studies. Similarly (Figure 3b) the calculated volume averaged moisture was in good agreement (RMSE 265 mol/m3) with the experimental data over all the time span. The maximum standard deviation in the observed data occurs at time 8 min, even if it is very small. Even if in same points there is not a perfect agreement between experimental and calculated data, probably due to the formation of steam inside the bean when the temperature increases.

Fig. 3. (a) calculated (black) and observed (gray) time-temperature curves for a point located at 3 mm from the surface; (b) simulated and experimental moisture concentration at different treatment time.

746

Angelo Fabbri et al. / Procedia Food00 Science (2011) 742 746 Author name / Procedia Food Science (2011)1000000

4. Conclusion The present study may be considered a first step toward the development of a more complete model of industrial interest. Particularly, the model could be useful to optimize the roasting process and to understand the relationships between process parameters and product quality better. The model may be used to study the roasting process of other food products (beans and/or seed), after appropriate implementations of the thermal-physical properties of interested materials. References [1]
Lerici CR, Nicoli MC. Il caff: aspetti chimici, tecnologici e qualit della bevanda. La Rivista della Societ Italiana di Scienza dellAlimentazione 1990. 19:116. [2] Eggers R, Pietsch A. Roasting. In Clarke RJ, Vitzthum OG, editors. Coffee: Recent Developments Blackwell Science, London, UK; 2001, p. 90107. [3] Severini C, Nicoli MC, Mastrocola D, Lerici CR. Influence of heating rate on some physico-chemical properties of coffee beans. In Proceedings of 14th Int. Sci. Colloquium on Coffee, ASIC, San Francisco, CA; 1992, p.641648. [4] Pittia P, Romani S. Coffee Processing. In Kristbergsson K, Otles S editors. Food Processing - Part IV Food Processing Operations, ISEKI Food Series , Springer; 2010. In press. [5] Schwartzberg HG. Modelling bean heating during batch roasting of coffee beans. In Welti-Chanes J, Barbosa-Canovas G, Aguilera MJ, editors. Engineering and food for the 21st century, Boca Raton: CRC Press LLC. 2002 [6] Hernandez JA, Heyd B, Irles C, Valdovinos B, Trystram G. Analysis of the heat and mass transfer during coffee batch roasting. J of Food Eng 2007; 78: 11418. [7] Heyd B, Broyart B, Hernandez JA, Valdovinos-Tijerino, Trystram G. Physical model of heat and mass transfer in a spouted bed coffee roaster. Drying technology 2007; 25:12438. [8] Basile M, Kikic I. Lumped specific heat capacity approach for predicting the non-stationary thermal profile of coffee during roasting. Chemical and Biochemical Engineering Quarterly, 2009; 23: 16777. [9] A.O.A.C. Official Methods of Analysis, 13th ed. Method n 968.11, Association of Official Analytical Chemists, The William Byrd Press, Inc., Richmond, Virginia; 2001.

Presented at ICEF11 (May 22-26, 2011 Athens, Greece) as paper MCF513.