A new arrangement of blades in scraped surface heat exchangers for food pastes

The present study analyses the performance of a scraped surface heat exchanger (SSHE) equipped with a new arrangement of blades to achieve higher thermal efficiencies than the conventional SSHE. This new design of exchanger, named here as scraped surface heat exchanger with alternate blades, A-SSHE, may be particularly suited to treat high viscous fluids, like food pastes. Experimental and numerical tests were carried out on an industrial scale A-SSHE used to heat hazelnut paste, an intermediate product widely used in Italian confectionery industry. Experimental tests include physical, chemical and rheological characterization of the hazelnut paste and the evaluation of the overall heat transfer coefficient as a function of rotational speeds and mass flow rates. Three-dimensional axial-symmetric CFD simulations of the A-SSHE were performed by using the software Fluent 6.2. For comparison, the same numerical tests were carried out for an equivalent SSHE with a conventional blades design (C-SSHE). Our studies show that the A-SSHE gives heat transfer coefficient values almost twice that of an equivalent C-SSHE, and that the numerical results are consistent with the experimental observations. The analysis of the fluid dynamics and of the thermal profiles suggests that the higher heat transfer efficiency of A-SSHE may be attributed to the occurrence of back mixing phenomena.

Efficient heating or cooling of complex fluids is an important problem for several operations in food industry. High viscosities and non-Newtonian behaviors of the processed fluids are often encountered in this field and tend to complicate their handling and to reduce the heat transfer rates in process equipment. In addition, the presence of solid food pieces causes additional problems related to the formation of solid deposits on hot surfaces. During heating, these deposits can eventually heat up to roasting temperatures altering the aromatic properties of the treated fluid. Some of these issues can be avoided or minimized by adopting a scraped surface heat exchanger (SSHE) in place of classical shell and tubes or double-pipe exchangers.

An SSHE is a double-pipe heat exchanger with a rotor coaxially inserted in the internal tube, known as the stator. The rotor is equipped with scraping blades to clean the inner surface of the stator. The blades play a double role in the SSHE: on the one hand, they literarily scrape the inner surface of the stator, avoiding particle deposition and cooking; on the other hand, they enhance the heat transfer rates by increasing the fluid velocity near the exchange surface and by generating turbulence. The typical design of an SSHE presents a given number of blades arranged longitudinally on the rotor surface; the blades have the same length of the rotor (Fig. 1). In the following, we call this configuration C-SSHE, where C stands for ‘‘continuous blades’’.

The principal source of information on C-SSHE derives from numerical simulations (De Goede and De Jong, 1993; Baccar and Abid, 1997, 1999; Sun et al., 2004; Bongers, 2006; Yataghene et al., 2008). Experimental studies on C-SSHE (e.g. Harrod, 1986; De Goede and De Jong, 1993; Dumont and Fayolle, 2000; Dumont et al., 2000; Yataghene et al., 2009) suffer the lack of reliable techniques and protocols to measure the local values of temperature, pressure and velocity. Numerical and experimental tests were usually carried out by considering Newtonian or model shear thinning fluids in order to mimic the properties of alimentary fluids as ice creams, tomato sauces etc. For these systems, practical operating conditions refer to a turbulent – or vortex flow – regime (Dumont et al., 2000). When high viscous fluids or concentrated food pastes (i.e. products with zero shear viscosity above 104 Pa s) are considered, the C-SSHE operates at low rotational velocity so that the operative conditions are usually within the laminar regime. The transition between laminar and turbulent regime usually appears for a value of the dimensionless generalized Taylor number (Ta) around 80 (Dumont et al., 2000).

The heat transfer coefficient of a C-SSHE in laminar conditions was modeled according to the dimensional analysis approach (Harrod, 1990). Currently, a generalized and microscopic interpretation of the convective and diffusive heat transfer mechanisms in SSHE is not available, and existing models are based on correlations between averaged heat transfer coefficient, averaged axial and rotational Reynolds number (Reax and Rer), averaged Prandtl number (Pr), number of blades (n) and their rotational speed (N).

Regardless of the numerical or experimental approaches adopted, all investigators observed a dramatic change in the thermal behavior close to the inlet and the outlet sections of the exchanger (e.g. Harrod, 1986; Dumont and Fayolle, 2000; Dumont et al., 2000; Stranzinger et al., 2001; Yataghene et al., 2009). As reported in the reviews of Harrod (1986, 1990), this phenomenon is related to back mixing: the rapid displacement of hot fluid moving backward that is replaced by cold fluid moving forward. These fluid displacements create a region characterized by higher fluid velocities and higher shear rates, where viscous dissipations and enhanced heat transfer rates seem to take place. Dumont et al. (2000) found that this phenomenon is also likely to occur in the laminar regime and, based on earlier experimental evidence, they concluded that it is related to disturbances in the flow patterns created by the geometries of the blades and the rotor. Recently, Yataghene et al. (2009) reported that viscous dissipations, occurring in the zones of higher shear rates close to the blade edges and in the inlet and outlet sections of the SSHE, give rise to a local temperature increase with a consequent decrease of the viscosity, which may indirectly enhance the heat transfer coefficient. In particular, the authors showed that the overall heat transfer coefficient for a C-SSHE operated with a shear-thinning fluid with power law constitutive equation, was strongly increased by viscous dissipation phenomena if the flow index of the fluid is higher than 0.5.

Since viscous dissipations and back mixing phenomena may increase the heat transfer coefficient during heating, these can be intentionally induced in an SSHE by adopting new rotor design. A possible option consists in the use of scraping blades shorter than the rotor and specifically positioned on its surface. In this work, we have studied the case of an SSHE with several couples of scraping blades placed along the rotor (Fig. 1). In each couple the blades are set at the opposite side of the rotor (i.e. shifted by 180°); each couple is rotated by 90° with respect to the previous one and two following couples are partially overlapped for 20% ca. of their length. Therefore, in this case, there are parts of the rotor with two blades and parts with four blades. We name this configuration Alternate Blades scraped surface heat exchanger, A-SSHE. This new configuration has been recently adopted on an industrial scale, but has not been studied so far. Experimental tests were carried out on an industrial scale A-SSHE, which was suitably instrumented to allow measures of the overall heat transfer coefficient in real operating conditions. This device is actually used for the sterilization process of hazelnut paste, an intermediate product of the Italian confectionery industry. A CFD analysis with a 3D axial-symmetric model was carried out both to verify the reliability of the numerical methods in determining the heat transfer coefficient in the A-SSHE and to obtain a description of fluid dynamics and heat transfer phenomena in this exchanger. Also according to the indications of Baccar and Abid (1997, 1999), we envisage that the 3D approach is made necessary by the laminar regime and by the geometry of the A-SSHE, even though the computational times are longer than those required for the 2D approach. To allow a suitable comparison between numerical and experimental data, dedicated investigations on physical, chemical and rheological properties of the hazelnut paste were carried out. Finally, due to the absence of any independent reference data on A-SSHE, the analysis of the thermal behavior of this new configuration is based on the comparison with that of a C-SSHE of equivalent design.