Chemical Engineering and Processing 46 (2007) 472–476 Heat transfer during thermal processing of a temperature dependent non-Newtonian fluid in a tubular heat exchanger C. Ditchfield a , C.C. Tadini a,∗ , R.K. Singh b , R.T. Toledo b a S˜ ao Paulo University, Escola Polit´ ecnica, P.O. Box 61548, 05424-970 S˜ ao Paulo, SP, Brazil b Department of Food Science and Technology, University of Georgia, 30602-7610 Athens, GA, USA Received 10 February 2006; received in revised form 20 April 2006; accepted 30 May 2006 Available online 16 September 2006 Abstract Heat transfer is highly dependent upon fluid properties and operating conditions for a particular heat exchanger. Determination of heat transfer coefficients in aseptic processing of a temperature dependent non-Newtonian fluid (banana puree) as a function of steam temperature, flow rate and heat exchanger’s length/diameter ratio is essential for modeling the thermal process. A commercial acidified aseptic banana puree with an average pH of 4.49 and average soluble solids concentration of 22.1 ◦ Brix was processed in a tubular heat exchanger with two heating sections. Three flow rates (2.5 × 10 -5 , 3.7 × 10 -5 and 4.7 × 10 -5 m 3 s -1 ), three steam temperatures (110.0, 121.1 and 132.2 ◦ C) and two length (L)/diameter (D) ratios (250 and 500) were studied. Density, thermal conductivity and specific heat capacity were considered constant and heat transfer coefficients were calculated from the temperature data obtained. For an L/D ratio of 500, heat transfer coefficients varied from 654.8 to 842.2 W m -2 K -1 , while for an L/D ratio of 250 heat transfer coefficients varied from 735.5 to 1070.4 W m -2 K -1 . An empirical correlation was proposed and verified, which explained the experimental data within 10% error. © 2006 Elsevier B.V. All rights reserved. Keywords: Banana puree; Heat transfer; Heat transfer coefficients; Empirical correlation 1. Introduction Knowledge of heat transfer and the parameters that govern it is essential for understanding the processing of foods. During many heat transfer operations in food processing, particularly continuous thermal processing of fluid foods, the governing heat transfer mode is convection. To model heat transfer by convec- tion heat transfer coefficients (h) are required. They depend on thermo-physical properties of product, heat exchanger geometry and surface roughness, and fluid flow regime [1]. Banana puree is a non-Newtonian temperature dependent Herschel–Bulkley fluid whose rheological behavior changes significantly with tem- perature [2]. There are numerous expressions in the literature to determine the heat transfer coefficient, but experimental determinations that include process parameters are important because only a few such expressions are found in the literature [3]. Heat trans- ∗ Corresponding author. Tel.: +55 1130912258; fax: +55 1130912255. E-mail address: catadini@usp.br (C.C. Tadini). fer coefficients are a function of the Reynolds number (Re), the Prandtl number (Pr), the length/diameter (L/D) ratio, the ratio between average viscosity and viscosity at the wall tempera- ture (μ/μ W ) and the flow behavior index (n) for non-Newtonian fluids [4]. Liu et al. [5] concluded that temperature variations induce viscosity variations that cause distortion in the veloc- ity profiles of non-Newtonian fluids, thus existing correlations many times fail to predict heat transfer coefficients correctly. Wichterle [6] demonstrated that Sieder and Tate’s empirical for- mula that considers viscosity variation as a ration of bulk and average viscosities elevated to the power 0.14 is only valid for certain operating conditions and fluid types. For a fluid that has such a complex rheological temperature dependent behavior as banana puree, the characteristics of the fluid flow will certainly influence heat transfer causing a deviation from the expected behavior. Quast et al. [7] studied heat transfer to banana puree in an agitated vessel for different types of agitation. The authors report that the agitation speed had a great influence on the heat transfer coefficient up to 30 rpm, and a further increase in agi- tation speed did not result in an increase in the heat transfer 0255-2701/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.cep.2006.05.018