4. Experimental studies for influence on flow and heat transfer

One of the fundamental research topics about offset strip fin is its geometry. The effect of the geometry on the performance of the fin and heat exchanger is the main concern of the investigators for decades. For this purpose, some researchers kept working on the effect of parameters on the performance of offset strip fin experimentally along the same path developed [5, 6]. One of these studies was released by Dong et al. [10]. Experiments are carried on air side of the heat exchanger to figure out the evolution of j and f for 16 different offset strip fins by flat tube heat exchanger. The flow of the air is ranging in 500<Re<7500, while the water flow is kept constant at 2.5m3 /h. The effect of fin geometry is presented for: space, height, length and flow length. In addition to the conventional approach to the fin structure, few researchers try to implement a new aspect to the studies. One of these investigations was reported by Peng et al. [11], which discusses the performance of a compact heat exchanger by an innovative offset strip fin array (Figure 7) at a range Re =500–5000. With regard to the results, when fin pitch increases, the heat transfer reduces while the friction factor reaches to higher values. In addition, the fin length effect is presented as well, with regard to that heat transfer will be enhanced by the shorter lengths, while the friction factor increases as well. One other valuable analysis that can be pointed out is the effect of fin bending; according to them, by the increment of fin bending, the heat transfer increases and friction factor reduces.

Plate fin heat exchangers are commonly used in the air to air or air to gas applications, but not so appropriate to the liquid to liquid applications, which also have crucial importance for some applications. In order to fulfill this requirement, brazed plate heat exchangers are emerged. Commercially available ones are made up of stainless steel that is brazed by copper or nickel. Even though they meet the higher pressure and temperature conditions, they are not sufficient for the applications where corrosion issues appeared. In order to overcome this deficit, Fernández-Seara et al. [12] studied titanium brazed plate fin heat exchangers with OSF. The corrosion resistance lets titanium to be used in liquid-liquid applications, unlike the typical fin

Figure 7. Scheme of the innovative offset strip fin [11].

A different performance evaluation aspect is employed to the OSF heat exchangers by considering second law of thermodynamics [9]. A new parameter that is called as relative entropy generation distribution factor is proposed by the researchers. This new parameter represents a ratio of relative changes of entropy generation. The effect of parameters, which are commonly used nondimensional parameters for the OSF studies, is discussed. Optimization for the investigated parameters is carried out, which could provide sufficient information about the conditions that could maximize the performance. The proposed performance evaluation parameter bases on the entropy generation number (Ns1) and can be written as in the

1/3 ratio vs. Re for different plate channels [8].

40 Heat Exchangers– Advanced Features and Applications

<sup>ψ</sup>� <sup>¼</sup> <sup>ð</sup>Ns1, <sup>o</sup>,Δ<sup>T</sup>−Ns1, <sup>a</sup>,Δ<sup>T</sup>Þ=Ns1, <sup>o</sup>,Δ<sup>T</sup> ðNs1, <sup>a</sup>,Δ<sup>P</sup>−Ns1, <sup>o</sup>,Δ<sup>P</sup>Þ=Ns1, <sup>o</sup>,Δ<sup>P</sup>

obtained data, smaller thickness at lower flow rates gives better results.

where 'a' and 'o' refer to the augmented (OSF) and reference (plain plate-fin) channel, respectively and 'ΔT' and 'ΔP' stand for entropy generation due to the temperature difference and

The given performance is examined for different geometrical parameters in order to estimate the effective ones, by considering varying relative temperature difference. According to the

<sup>¼</sup> <sup>1</sup>−N�

N� <sup>s</sup>1,Δ<sup>P</sup>−1

s1,ΔT

: (21)

following,

Figure 6. j/f

pressure drop, respectively.

structures. Across the heat exchanger, water-water and 10–30% ethylene glycol are used in the experiments. As a result of the investigations, empirical correlation to determine single phase convection heat transfer coefficient, as a function of Re is derived. The importance of this correlations are underlined since the experiments are carried out for water and ethylene glycol, which have higher numbers than the air and tends to provide more accurate predictions than the ones derived by air. The outcomes of the research are presented for varying Re number of water-water test and ethylene glycol mixture, individually. Pressure drop and overall heat transfer coefficients of the particular conditions are summarized with respect to the velocities or mass flow rate that they are tested. It is noted that the pressure drop gets higher as the ethylene glycol percentage increases when the flow regime drives up. In addition, the heat transfer is poor in ethylene glycol when it is compared with water and it gets worse when the mass flow reduces.

Offset strip fins are usually operated at low Reynolds numbers (Re<1000), at which the flows are typically steady and laminar. Researchers have investigated the thermohydraulic performances of the offset strip fins widely at this regime. Unlike the reported studies, Michna et al. [13] tried to find the answer for the thermohydraulic analysis of the OSF at turbulent regime. The friction factor and heat transfer coefficient of the OSF are measured at Re number in the range 1000<Re<10,000 in the experiments. In order to have a better approach, pressure and heat transfer tests are executed by aluminium and stainless steel, respectively, where the constant heat flux is provided by film heaters. For the confirmation of the friction factor, the correlation of Manglik and Bergles [4] (the correlation given in Table 3) is used, even though the correlation is for Re<20,000, the correlation is extrapolated for the higher values of Re. An agreement with the correlation could be observed for friction factor at Re<20,000 and it starts to oscillate beyond that range, while Colburn j-factor consistently has higher values than the correlation of Manglik and Bergles [4]. The evaluation of the modified Colburn j-factor is presented row by row for varying Re values as well. With regard to the vortex shedding and turbulent flow, the highest values of j and f are observed for the higher values of Reynolds [13].

Even though the uniform distribution is expected to obtain higher performances in compact heat exchangers, maldistribution could also be seen, especially for parallel channel heat exchangers. There are limited studies published on this topic for offset strip fins. One of those few studies is released by Saad et al. [14], in which the experimental distribution of phases and pressure drop in offset strip fin type compact heat exchangers is given. The experiments are accomplished in a flat vertical compact heat exchanger in which the flow of air and water are examined. The experiments are carried out for single phase and two phase flows by order. The single phase flow experiments are followed by CFD analysis of the same problem employed in the investigation. Furthermore, a correlation (Table 3) that belongs to single phase flow is developed as well for prediction of friction factor. The nonuniformity of the two phase flow regime is determined by the maldistribution parameters such as standard deviation (STD) and normalized standard deviation (NSTD). The pressure drop profiles and flow rate distribution are used to identify the phase distribution of the flow since it is hard to determine the solely phase distribution without other physical features. The results of the experimental and CFD

#### Reference Correlation

structures. Across the heat exchanger, water-water and 10–30% ethylene glycol are used in the experiments. As a result of the investigations, empirical correlation to determine single phase convection heat transfer coefficient, as a function of Re is derived. The importance of this correlations are underlined since the experiments are carried out for water and ethylene glycol, which have higher numbers than the air and tends to provide more accurate predictions than the ones derived by air. The outcomes of the research are presented for varying Re number of water-water test and ethylene glycol mixture, individually. Pressure drop and overall heat transfer coefficients of the particular conditions are summarized with respect to the velocities or mass flow rate that they are tested. It is noted that the pressure drop gets higher as the ethylene glycol percentage increases when the flow regime drives up. In addition, the heat transfer is poor in ethylene glycol when it is compared with water and it gets worse when the

Offset strip fins are usually operated at low Reynolds numbers (Re<1000), at which the flows are typically steady and laminar. Researchers have investigated the thermohydraulic performances of the offset strip fins widely at this regime. Unlike the reported studies, Michna et al. [13] tried to find the answer for the thermohydraulic analysis of the OSF at turbulent regime. The friction factor and heat transfer coefficient of the OSF are measured at Re number in the range 1000<Re<10,000 in the experiments. In order to have a better approach, pressure and heat transfer tests are executed by aluminium and stainless steel, respectively, where the constant heat flux is provided by film heaters. For the confirmation of the friction factor, the correlation of Manglik and Bergles [4] (the correlation given in Table 3) is used, even though the correlation is for Re<20,000, the correlation is extrapolated for the higher values of Re. An agreement with the correlation could be observed for friction factor at Re<20,000 and it starts to oscillate beyond that range, while Colburn j-factor consistently has higher values than the correlation of Manglik and Bergles [4]. The evaluation of the modified Colburn j-factor is presented row by row for varying Re values as well. With regard to the vortex shedding and turbulent flow, the highest values of j and f are observed for the higher values of Reynolds [13].

Even though the uniform distribution is expected to obtain higher performances in compact heat exchangers, maldistribution could also be seen, especially for parallel channel heat exchangers. There are limited studies published on this topic for offset strip fins. One of those few studies is released by Saad et al. [14], in which the experimental distribution of phases and pressure drop in offset strip fin type compact heat exchangers is given. The experiments are accomplished in a flat vertical compact heat exchanger in which the flow of air and water are examined. The experiments are carried out for single phase and two phase flows by order. The single phase flow experiments are followed by CFD analysis of the same problem employed in the investigation. Furthermore, a correlation (Table 3) that belongs to single phase flow is developed as well for prediction of friction factor. The nonuniformity of the two phase flow regime is determined by the maldistribution parameters such as standard deviation (STD) and normalized standard deviation (NSTD). The pressure drop profiles and flow rate distribution are used to identify the phase distribution of the flow since it is hard to determine the solely phase distribution without other physical features. The results of the experimental and CFD

mass flow reduces.

42 Heat Exchangers– Advanced Features and Applications


Table 3. Heat transfer and friction factor correlations for offset strip fins.

Figure 8. A comparison of test and numerical results of coolant polyalphaolefin [15].

simulation are confirmed by Manglik and Bergles [4] correlation (Table 3). A similar trend of that correlation, which is the descending of the friction factor by ascending Re number, is observed in the reported results. The correlated friction factor and the comparison of it with the reported ones are summarized and given in detail in correlations of j and f factors in OSF heat exchangers (HEX) section in order to give a better aspect to the findings.

Although the scope of the study is about OSF compact heat exchangers, which mostly deal with air, this is not the only fluid used with OSF. So, it is important to note the thermal hydraulic response of the fin structure with different Pr numbers. The effect of Prandtl is presented in a series of investigation by Hu and Herold, either experimentally [15] and numerically [16]. The experiments are corresponding to a liquid (water and polyalphaolefin) cooled offset fin compact heat exchanger. As for the given liquids, the Pr number that is used in the study ranges from 3 to 150. The obtained results compared by the air cooled models and a noticeable difference is observed for varying Prandtl values. A numerical analysis is performed to investigate uniformity of heat flux and temperature distribution and also for the validation of the tested conditions, where a good agreement could be seen (Figure 8). The model results were used to guide data reduction procedure.

#### 5. Numerical studies for influence on flow and heat transfer

One of the leading numerical investigations about the OSF heat exchangers is carried out by Joshi and Webb [3]. The laminar and turbulent flow regimes are analytically modeled to predict the j and f factors of OSF. Moreover, the experiments are employed to visualize the flow structure, transition regime in particular. The ultimate purpose of this endeavor is to classify the regimes as the laminar, transition, or turbulent as best as it could be. The necessary equations for Nu and f are derived from energy and momentum balances. The character of

Comprehensive Study of Compact Heat Exchangers with Offset Strip Fin http://dx.doi.org/10.5772/66749 45

Figure 9. Flow patterns observed in visualization experiments [3].

simulation are confirmed by Manglik and Bergles [4] correlation (Table 3). A similar trend of that correlation, which is the descending of the friction factor by ascending Re number, is observed in the reported results. The correlated friction factor and the comparison of it with the reported ones are summarized and given in detail in correlations of j and f factors in OSF

Although the scope of the study is about OSF compact heat exchangers, which mostly deal with air, this is not the only fluid used with OSF. So, it is important to note the thermal hydraulic response of the fin structure with different Pr numbers. The effect of Prandtl is presented in a series of investigation by Hu and Herold, either experimentally [15] and numerically [16]. The experiments are corresponding to a liquid (water and polyalphaolefin) cooled offset fin compact heat exchanger. As for the given liquids, the Pr number that is used in the study ranges from 3 to 150. The obtained results compared by the air cooled models and a noticeable difference is observed for varying Prandtl values. A numerical analysis is performed to investigate uniformity of heat flux and temperature distribution and also for the validation of the tested conditions, where a good agreement could be seen (Figure 8). The model results

One of the leading numerical investigations about the OSF heat exchangers is carried out by Joshi and Webb [3]. The laminar and turbulent flow regimes are analytically modeled to predict the j and f factors of OSF. Moreover, the experiments are employed to visualize the flow structure, transition regime in particular. The ultimate purpose of this endeavor is to classify the regimes as the laminar, transition, or turbulent as best as it could be. The necessary equations for Nu and f are derived from energy and momentum balances. The character of

heat exchangers (HEX) section in order to give a better aspect to the findings.

Figure 8. A comparison of test and numerical results of coolant polyalphaolefin [15].

44 Heat Exchangers– Advanced Features and Applications

5. Numerical studies for influence on flow and heat transfer

were used to guide data reduction procedure.

flows on fins and in wakes is illustrated (Figure 9) with regard to the data attained from the numerical analysis, which are demonstrated with respect to velocity profiles. Since the effect of the fin length, thickness and spacing on the wake have not revealed in earlier communications, it is focused to find out this by either visualization or numerical analysis. The findings of the experiments are demonstrated by Figure 9, which starts to oscillate when the regime turns from laminar to turbulent (from (a) to (d) in Figure 9). Parametric study is included as well and the results of the friction factor are summarized. It is noted that when the thickness to length ratio ascends, the friction factor increases [3]. A correlation (Table 3) that bases on the wake width is developed to predict the Re at the transition points from the Re numbers of 21 different heat exchangers reported in literature.

Along the same path, the effect of thickness for the flow regime and performance is examined numerically [17]. And the flow pattern is investigated as presented in Figure 10 (fin thickness ascends from top to bottom cartoon). With regard to that results when the fin is thin enough no particular difference could be observed from the low Re number. With the increment of the fin thickness recirculation zones emerge next to the plate.

Another study about the effect of parameters on the performance of the offset strip fin was reported by Kim et al. [2]. Even though the ultimate purpose of the study is to provide a better correlation for the offset strip fins at higher blockage ratios and different fluids, the parametric effect is also observed. It is remarked that the j and f factors descend when the spacing and length increases and in contrary it rises when the thickness increases. By the means of pressure drop, a similar trend could be observed with the j and f factors. It is highlighted that the thickness effect is more prominent on the performance than the spacing and length.

Even though most of the studies employed for the flow at horizontal heat exchangers, Suzuki et al. [18] operated the flow at vertical flat tubes used in the free-forced-mixed convection at low Re. The investigation has proceeded for a staggered array using numerical and experimental approaches. Three rows of flat plates are arranged in a staggered pattern where the air flows upward at a low speed.

Figure 10. Flow patterns for different plate thickness [17].

In the numerical computations, two methods are used: the first one is α-ω method, the second one is U-V-P method (which solves axial and normal velocity components U and V together with pressure correction ΔP); in order to see the difference, both methods are run for the same case where it is seen they are almost identical. The effect of thickness is observed along the flow

Figure 11. Comparison of the experiments by London and Shah [18].

length in the described array and any significant impact could be observed at the slower flow regimes. In all those trials, the highest heat transfer is obtained at the entrance of the array. A comparison between the published data from London and Shah [6] and the numerical data of the study is presented (Figure 11). It is noted that there is a good agreement between the calculated results and London and Shah for their offset geometry Core 104. Finally, it is concluded that the thicker fins are not necessary for the slower regimes since there is no distinctive effect on the performance. Even though the shorter fins give better heat transfer performance due to its higher production cost it is not efficient to replace with longer fins.

The physical phenomena for the heat transfer mechanism in offset strip fin geometries for the self-sustained oscillatory flow at high Re range is also considered by the researchers. Saidi et al. [19] noted that, due to the nature of the compact heat exchangers, in most of the studies, the flow regime is at lower Re range; in addition, in some rare applications, some researchers examine the flow at higher Re range in order to see the response at the turbulent. The flow regime at the intermediate Re range is not very common and there are unrevealed parts for this flow type. A numerical study based on finite volume method computed by a CFD code is given in the communication [20]. Friction factor and Colburn j-factor are compared by DeJong

In the numerical computations, two methods are used: the first one is α-ω method, the second one is U-V-P method (which solves axial and normal velocity components U and V together with pressure correction ΔP); in order to see the difference, both methods are run for the same case where it is seen they are almost identical. The effect of thickness is observed along the flow

Figure 10. Flow patterns for different plate thickness [17].

46 Heat Exchangers– Advanced Features and Applications

Figure 11. Comparison of the experiments by London and Shah [18].

Figure 12. The comparison of Colburn j-factor of different heat transfer surface types [21].

et al. [20] for the confirmation of the findings. Then, the velocity field around the fin during a complete period of oscillation is presented by contour plots for various time steps in the sequence of development. Moreover, unsteadiness is presented in the same manner. The second moment correlation of fluctuating velocity components and second moment of temperature velocity fluctuations are demonstrated. It is noted that heat transfer and moment transfer are dissimilar.

Orientation of inlet and outlet headers plays a major role in performance of compact heat exchangers, especially in aerospace applications where the orientations are not straight. In order to understand the maldistribution at these components, Ismail et al. [21] studied these phenomena numerically. Three headers are developed to modify the distribution, which are combined with three offset strip fins and sixteen wavy fin geometries used in the heat exchangers. The computational results for wavy and offset strip fin structures are validated by analytical results and a good agreement is observed for j within –2% and for f within –9%, as it is illustrated in Figures 12 and 13. As for the headers, the flows are analyzed for either real (without modified headers) or ideal cases (with modified headers) and in all these conditions, the pressure drop is higher at real cases than at the ideal case. Among those, the biggest real case-ideal case difference is observed for the heat exchanger with baffle plate by 34%.

Figure 13. The comparison of friction factor of different heat transfer surface types [21].
