Evaluating Jet Pump Turbulizers in Double Tube Heat Exchangers: A Preliminary CFD Study (2024)

Nowadays, Computational Fluid Dynamics (CFD) is increasingly being used to study heat transfer intensification and optimise the performance of heat exchangers. In contrast to experimental measurements, the use of CFD provides insight into the essence of heat transfer in an exchanger by visualising the flows and knowing the properties of the media in each cell of the computational grid, without disturbing the flow with the presence of sensors. The degree to which CFD is used varies strongly from study to study: from visualising velocity fields and flow direction in the apparatus to show the mechanism of turbulizers [1] by modelling a section of the exchanger wall with turbulent inserts in 2D geometry [2] to obtaining results simulating the operation of a complete shell and tube heat exchanger in 3D geometry [3]. Thanks to the high computing power of personal computers and the spread of commercial software such as Ansys Fluent, simulating the operation of a heat exchanger has become a less costly and less labour-intensive alternative to making and testing prototypes [4].

Heat transfer processes are present in almost every plant in the chemical industry. Typically, thermal apparatuses are characterised by a large surface area to maximise heat transfer between the media. The increasing cost of construction materials and the pressure to minimise environmental impact have prompted an increase in exchanger efficiency. This could be archived achieved mainly by increasing the turbulence of the flow and, consequently, the intensity of heat transfer from the exchanger wall to the fluid. The easiest way to do this is to increase the flow velocity through the exchanger. Such a solution is not preferred because of the higher power requirements for the pumps and electrical energy costs. An alternative solution is to use passive turbulence promoters called turbulizers. Turbulizers come mainly in two forms: surface modifications of the exchanger and inserters [5, 6]. Regarding surface modifications, many researchers studied the many shapes of grooves and ribs [7-11]. The main disadvantage of surface modification method is the problematic cleaning of the exchanger equipped with the above-mentioned devices. This problem/difficulty is solved by the second class of turbulizers: Easy to disassemble turbulizer inserts. The characteristics of twisted tape turbulizers have been well studied [12-14] as classic turbulizers. Many authors have described their modifications [3, 15, 16] and completely different types of turbulizing inserts. The effects of Białecki rings were studied by Dziak and Ratajczak [17], nodular turbulizers by Charun [1], rotating turbulizers by Fodemski and Staniszewski [18], vortex generating turbulizers by Saraç and Bali [19]. In the course of the above mentioned studies, an increase in the achievable Nusselt number of several tens to more than one hundred per cent was obtained compared to exchangers without turbulizers, especially for low values of the Re. On the other hand, the use of some turbulizers increased pressure drops by up to 50 or 100 times. Consequently, the use of turbulizers led to a reduction in the size and cost of heat exchangers while requiring more efficient and more expensive pumps.

At the same time, in the chemical industry, jet pumps are successfully used for pumping corrosive fluids, dispersing suspensions and emulsions and as components of mixers. Their hydrodynamics has been described in full, among others, by Karambirov and Chebaevskii [20] and Winoto et al. [21]. The operating principle of a typical jet pump is shown in Figure 1. The high-pressure working fluid enters through the main inlet (A). In the nozzle (B), the decreasing diameter forces the flow to accelerate and form a jet (C), which simultaneously reduces its static pressure. The moment the pressure drops below the pressure on the side inlet (D), the pressure difference starts to push the secondary fluid through the side inlet and into the mixing chamber. There, the high velocity difference between both fluids creates a high sheer stress and turbulences which start to mix both fluids. The mixing continues inside the diffusor (E), where the diameter gradually increases to decrease the fluid’s velocity and recover some of its original static pressure. The mixed fluid flows through the jet pump outlet (F). Many researchers focus on the optimisation of the jet pumps for specific applications and operating conditions [22, 23] or determining the influence of individual geometry features [24, 25], but their studies concerned typical applications of jet pumps – as pumps or ejectors.

The authors therefore set out to test whether the use of a novel jet-pump-like structure, composed of a nozzle and diffusor parts placed inside the heat exchanger pipe will reproduce suction and mixing behaviour typical for jet pumps. The use of this new type of turbulizers, called jet pump turbulizers or ejector turbulizers, could result in increased turbulent flow and jet mixing accelerating the heat transfer deep into the flowing stream, and consequently noticeably increases the amount of heat exchanged. In this study, Computational Fluid Dynamics (CFD) was used to investigate the performance of jet pump turbulizers, compared to classical strip turbulizers in a double tube heat exchanger, whose experimental data of pressure drops and heat transfer coefficients are available in the literature.

Figure 1. Schematics of typical jet pump with the main flows marked

The modelled exchanger with an effective length of L=585 mm consists of a 16×2.5 mm inner tube and a 32×2 mm outer tube, made of brass. In the inner tube of the exchanger, the tested jet turbulizer was placed in such a way that the outlet of its nozzle was located exactly halfway along the exchanger (Figure 3). The influence of the basic dimensions of the jet was investigated by simulating four variants of the turbulizer with diameters and angles collected in Table 1.

Table 1. Geometrical configurations of the tested jet pump turbulizers

For use in simulations, the exchanger geometry, including the jet pump, was simplified to the axisymmetric 2D case. Only the flow through the inner pipe was modelled on the hydrodynamics of the turbulizers tested. The numerical mesh had a variable cell size depending on the exchanger region, ranging from 300 μm in the free space of the pipe to 30 μm at the outlet of the jet nozzle (a total number of cells for hydrodynamic calculations was from 150,411 to 297,576, for thermal calculations – from 203,323 to 340,201 depending on the jet pump). In addition, a 7-layer boundary grid was used to increase the resolution of the calculations on the walls. During grid validation, it was confirmed that there was no effect of its further refinement on the results (for a grid that was twice as dense, the pressure drop differed by 0.0661% and the ejection coefficient by 0.687%).

2.3 CFD simulation parameters

Simulations were carried out using the Ansys CFD package. A realizable k-ε model was adopted to describe the turbulence, as confirmed in literature to give good results with modelling of both jet pumps [22, 23] and double tube heat exchangers equipped with the turbulizers [1, 3]. Simulations were performed as steady-state, with pressure-based solver, Semi-Implicit Method for Pressure Linked Equations (SIMPLE) pressure-velocity coupling scheme, least squares cell based gradient discretization and second order “upwind” properties discretization along with hybrid initialization. All other settings, including model calibration constants and relaxation factors, were left as default values. The temperature dependence of density, viscosity, thermal conductivity and specific heat of water was approximated by polynomials based on experimental data available in the literature [28] while the properties of brass were assumed to be constants due to insufficient data for approximation [29]. The boundary condition velocity inlet was used at the inlet of the device, whereas the boundary condition pressure outlet was used at the outlet of the apparatus. In hydrodynamic studies, inlet velocities v_in were 0.112÷0.775m/s and medium temperatures T_in12÷50℃. In heat transfer tests, the velocities and temperatures were, respectively: for the cold medium (in the inner pipe) v_in,c=0.054÷0.727m/s and T_in,c =8.7÷17.1℃, for a hot medium (in the outer pipe) v_in,h=0.187÷0.196m/s and T_in,h =30÷60℃.

Figure 6. Comparison of heat transfer coefficients

Figure 7. Comparison of reduced ratio of heat duty to mechanical power

This paper deals with the idea of jet pump like construction of turbulizers never before described in literature. CFD simulations made it possible to test the performance of the proposed new turbulizer design. The existence of suction and mixing of the fluid from the laminar boundary layer into the jet was confirmed, with a consequent significant reduction in the heat transfer resistance in the area beyond the turbulizer. Partial flow reversal was also demonstrated, increasing mixing and temperature uniformity, as well as rapid flow in the gap between the nozzle and the exchanger wall, further disrupting the boundary layer.

The preliminary study found that the proposed jet pump turbulizers show considerable potential for use in double tube heat exchangers. CFD simulations suggest achieving 5÷10 times higher Euler number values Euand 1.7÷3 times higher values of overall heat transfer coefficient k_Twith the use of jet pump turbulizers compared to an empty heat exchanger, as well as 1.2÷4 times higher Euler number values and 1÷2 times higher values of overall heat transfer coefficient against strip turbulizers. Of the geometry variants considered, especially the one with a small nozzle and diffuser inclination angle and a large nozzle diameter (geometry No.2) promises the most favourable increase in heat load of the exchanger in relation to the increase in pressure losses, especially at low flow rates of Re<4,500, as indicated by the value $\varepsilon_r$>1.

Preliminary numerical research has yielded promising results for the proposed jet pump turbulizers, which justifies further work on their development. Construction of the experimental installation is planned to confirm the results and validate the simulation model for further studies on determining the optimal jet turbulizer geometry.