1、OR-05-1 0-2 Thermal and Hydraulic Characteristics of Brazed Plate Heat Exchangers-Part I: Review of Single-phase and Two-Phase Adiabatic and Flow Boiling Characteristics Joachim Claesson ABSTRACT This paper reviews the status of research reported in the open literature on plate heat exchangers, oper
2、ating in single- phase and two-phase adiabatic andflow boiling applications. The review has been made as apart of an ongoing study aimed at enhancing the eficiency of heat transfer in plate heat exchangers used as evaporators in small heatpumps. It consti- tutes the background for the report of the
3、activities within the project, given in part two of the present article. In order to obtain a complete picture of the performance of a plate heat exchanger operating as an evaporatol; it is as important to have knowledge of the single- phase characteristics as it is to have knowledge of theflow boil
4、ing characteristics. The review shows that single-phase flow is quite well understood, even if the influence of the many geometric parameters is still under discussion. It also shows that there is still a controversy concerning which mechanism is dominant in flow boiling: convective evaporation or n
5、ucleate boiling. INTRODUCTION Plate heat exchangers (PHE) are very effective devices for transferring heat between two fluids. The PHE was initially used for single-phase applications, e.g., in the dairy industry and district heating systems. Used in these applications, the PHE had an advantage ovcr
6、 traditional geometry-the heat transfer surface could readily be cleaned. The ease of cleaning was due to the use of gaskets to seal the fluids from each other and from their surroundings. The plates were held together, with a gasket in between each plate and two outer thick, flat end plates support
7、ing the heat transfer plates, forcing them together to obtain metallic contact. A further aspect of these gasketed plate heat exchangers was the possibility of increas- ing the heat transfer area simply by adding more plates. The gasketed plate heat exchanger could also be used in two-phase applicat
8、ions; however, the gasket material limited the temper- ature and pressure range at which the PHE could be operated. The introduction of compact brazed plate heat exchangers (CBE) overcame this limitation, allowing higher operating pressures and temperature levels. A standard brazed plate heat exchan
9、ger from one of the leading Swedish manufacturers is certified for pressures up to 30 bar (a) and a temperature of 155OC. The ability to withstand higher pressures made the brazed plate heat exchanger interesting for heat pump and refrigeration applications, both as evaporators and condens- ers. The
10、 brazed plate heat exchanger consists of a number of (often) identical pressed plates. Every second plate is rotated 180, assembled in a pack with a thin copper sheet between each plate, and then put into a hot furnace. The thin copper sheet melts and the capillary forces moves the liquid copper to
11、the contact points of the two opposite plates. Then, the temper- ature is decreased and the copper solidifies, creating a plate pack with many strong contact points (see Figure i). The complex flow geometry created in plate heat exchangers facil- itates high heat transfer rates but also leads to hig
12、h pressure drops. The single-phase heat transfer and hydraulic character- istics of a plate heat exchanger have been found to depend on several geometrical parameters, e.g., chevron angle (v), corru- gation pitch (A), and corrugation depth (b), defined in Figure 2. In addition, another important geo
13、metrical parameter is the enlargement factor (I$), i.e., the ratio between actual heat trans- fer area and projected heat transfer area. Joachim Claesson is pursuing his PhD degree in the Department of Energy Technology at the Royal Institute of Technology, Stockholm, Sweden. 822 02005 ASHRAE. Figur
14、e 1 Plate assembly. DIMENSIONLESS PARAMETERS In the treatment of plate heat exchangers in the literature, at least two different definitions of the hydraulic diameter are used. Perhaps the most common definition used is similar to the definition of two wide parallel plates, with a distance of b betw
15、een the plates, hence, de=2.b. (1) B The other definition, perhaps more “physically” correct since it is defined according to the noncircular tube definition of the hydraulic diameter, is In the following, we distinguish between these two by the use of different subscripts, e for effective diameter
16、and h for hydraulic diameter, as suggested by Shah and Focke (1988). Now, the Reynolds and Nusselt numbers may be defined as G . de P Re = - and a . de NU = - h (3) (4) ,_.-._ -. I Figure 2 Plate assembly and characteristic geometry parameters. work is defined, based on the effective diameter and th
17、e projected length between the inlet and outlet ports, as P AP . de f= 2 2.L;G However several different definitions may be found in the literature (see Table 1). In the references of Table 1, there seems to be no clear explanation for choosing the developed flow length over the distances between th
18、e ports. Wanniarachchi et al. (1995) mentioned that it is “more reasonable” but does not offer any explanation. Bogaert and Blcs (1995) mention the difficulties of defining an appropriate flow length in a three-dimensional flow geometry and develop a plate specific, dimensional, fric- tional functio
19、n to avoid the difficulties. Muley (1997) and Muley and Manglik (1999) used the distances between the ports, but Muley et al. (1999) instead chose to use the devel- oped flow length. No explanation or reason for the change was given. One may think of a physical benefit of using the actual fluid velo
20、city rather than the mass flow velocity used in Equa- tion 3. The actual fluid velocity, assuming a two-dimensional flow, would be the velocity the fluid has as it flows in the corru- gations. The flow would then resemble a fluid system with many parallel channels. The relation between the actual fl
21、uid velocity, u, and the mass flow velocity, u, would be (6) um cos cp u=- where the effective diameter is used. When using correlations nitions used in the original texts. Especially confusing is the definition of the friction factor, where an extra geometrical from the literature, it is very impor
22、tant to adhere to the defi- parameter, the flow length, is added. Two different definitions of flow length can be found in the literature: developed flow Or (7) u=- G p. coscp length and the length between the inlet and outlet ports. These have been mixed freely with different definitions of hydraul
23、ic diameter. The friction factor of a plate heat exchanger in this The appropriate Reynolds number, based on the actual fluid velocity, would then also be a function of the chevron ASHRAE Transactions: Symposia 823 Table 1. Definitions of Friction Factor from the Literature Reference Definition Leng
24、th Scales Martin (1 996) Edwards et al. (1974) L =A / W(flow length) g, = conversion factor (= 1 in SI units) 2. Ap. d, =p.u: .L, (10) Lp = port center-to-center distance I (8) Bogaert and Blcs (1995) L = A / W (flow length) Wanniarachchi et al. (1995) Focke (1983) Focke et al. (1 985) Talik et al.
25、(1995a,b) Muley and Manglik (1 999) Muley et. al. (1999) angle. In this sense, a use of an actual flow length, which would be a function of the plate length and the chevron angle, similar to Equation 6, would be rational rather than the devel- oped flow length used in Equations 9, 1 O, and 13. In ad
26、dition, using this concept would facilitate the use of the analogy between heat transfer and pressure drop. One major drawback with this approach is that the flow may not be following the corrugations from side to side but, rather, take a spiralling path along the main direction of the flow, as disc
27、ussed by Shah and Focke (1988). The actual flow length and the actual number of changes in the flow direction is, therefore, not known. Thus, it is simpler to consider the plate heat exchanger as a rough channel, using the mass flow velocity, the effective diameter, and, finally, the length between
28、port centers as char- acteristic length to correlate friction pressure drop and heat transfer (see Equations 1 and 3 through 5. In Table 1, “core” indicates definitions where the pressure drop in the heat transfer section of the plate heat exchanger only is used. The others may include the pressure
29、drop in the ports. The reader is referred to the original references for specific details. Plotting the friction factor and the Nusselt number vs. the Reynolds number shows one benefit of using plate heat exchangers in single-phase flow: the thermal and hydraulic characteristics for different Reynol
30、ds numbers display no abrupt change in either of the two (see Figure 3) in the transi- tion from low to high Reynolds numbers. SINGLE-PHASE FLOW The heat transfer and pressure drop in plate heat exchang- ers have been investigated for several years, and the amount of work has become rather extensive
31、. On the other hand, the possible combinations of geometric parameters are almost infinite. Hence, there does not exist a general theory or corre- lation covering all geometrical combinations. Each investiga- tion should rather be regarded as a special case and the results only applicable for the sp
32、ecific geometry tested. Unfortu- nately, all geometric parameters are seldom reported in detail. Overall Thermal Performance Bounopane et al. (1963) presented an overview of the thermal performance of plate heat exchangers and suggested values for the correction factor for the logarithmic mean tempe
33、rature difference. For a two-channel heat exchanger (one thermal plate), pure countercurrent heat exchanger may be assumed. For all other single-pass heat exchangers, the correction factor was given as 0.967 and 0.942 for even and odd number of thermal plates, respectively. Kandlikar and Shah (1989a
34、, 1989b) later conducted a more refined analysis, where correction factors for LMTD and effectiveness-NTU correlations were given for a number of different flow config- urations. Bond (198 1) presented data from the APV Company. He states that mixed angle plates may be used to extend the range of av
35、ailable chevron angles. The performance of the mixed plate may be estimated using the average chevron angle, 824 ASHRAE Transactions: Symposia 100 100 O00 - 10000 - 10 Yi 1 O0 10 - . 1 1 Re 1 10 1w 1000 1ow)o 10000010w0 Figure 3 Single-phuse thermal und hydraulic characteristics as function of Reyno
36、lds numbez according to the author. This was later confirmed by Walton (2001), who used mass transfer tests to determine local mass transfer coefficients in mixed angle plate heat exchangers. Walton states that even though the individual plates have very different mass transfer characteristics, the
37、integral mass trans- fer performance is very similar to a plate pair having a pure chevron angle equal to the chevron angle arithmetic mean of the two mixed plates. One of the major contributors to the knowledge of single- phase heat transfer and pressure drop of plate heat exchangers is Focke. Toge
38、ther with Co-workers he has published several papers concerning, e.g., effect of chevron angle (Focke 1983; Focke et al. 1985), asymmetrically corrugated plates (Focke 1985), flow visualizations (Focke and Knibbe 1986), and a method of selecting optimum plate heat exchanger surface pattern using hea
39、t transfer pressure drop analogies (Focke 1986). Film Heat Transfer Coefficient The film heat transfer coefficient for plate heat exchang- ers has been investigated by several researchers. Most of them correlate the heat transfer coefficient using a modified Dittus- Boelter type of equation, where t
40、he constants and exponents are changed. Focke (1995) investigated several heat transfer analogies. He concludes that all the transport analogies tested approxi- mately hold for plate heat exchangers. He discussed the possi- bility that the Calderbank equation has a more fundamental validity compared
41、 to the other analogies tested. Bogaert and Blcs (1995) investigated experimentally heat transfer and pressure drop of plate heat exchangers from SWEP. This paper is interesting since it is almost the only one where the exponents on the Prandtl number and on the viscos- ity ratio are not given as co
42、nstants and are also explicitly given. A similar expression for the exponent on the viscosity ratio for calculation of the friction factor has also been reported by Shah and Focke (1988). The Nusselt number was given by Bogaert and Blcs (1 995) as where BI and B, are empirical constants specific for
43、 a certain plate and a certain Reynolds number range. This kind of corre- lation is recognized as a modification of the well-known Dittus-Boelter correlation. It is the absolutely most common type of heat transfer correlation for plate heat exchangers found in the literature, even though the constan
44、ts are different and the exponents on the Prandtl number and the viscosity ratio are often constants. A similar correlation was suggested by Muley (1 997). He developed empirical correlations for the parameters BI and B, in Equation 14, both being functions of the chevron angle and the area enlargem
45、ent factor. The same correlations are also reported in Muley and Manglik (1 999). The correlations are based on data from the literature as well as on their own exper- imental data. The final correlation for the Nusselt number at low Reynolds numbers was stated as Nu = 0.44 . (5)038 . . Pr 1/3()0.14
46、 30 5 p 5 601 30 I Re I400 and for higher Reynolds numbers, (0.2668 - 0.006967 . p + 7.244. 20.7803 - 50.9372 . Cp + 41.1585 . $2 - 10.1507. (p3) . 9,) Nu = I ( (16) Crozier et al. (1964) described the use of plate heat exchangers with non-Newtonian fluids. However, perhaps the most interesting fact
47、 with this paper is the statement that the heat transfer in plate heat exchangers may be calculated using the Lvque approximation under certain conditions. This idea was exploited independently by Martin (1996) who developed a semi-theoretical correlation of heat transfer and pressure drop in plate
48、heat exchanger. Martin used the hydrau- lic diameter in defining the Reynolds and Nusselt numbers; thus, Re = Re, . I$ (17) and ASHRAE Transactions: Symposia 825 Martin developed his correlation by extending the Lvque theory into the turbulent region. The final expression for the Nusselt number was
49、given as The range of validity of the correlation is not explicitly given in the original reference; however, it seems that the Reynolds number was varied between 400 to 10 000. For the specific range of validity, the reader is referred to the sources of the original reference. Gaiser and Kottke (1 998) also investigated the chevron angle and the effect of corrugation pitch on heat transfer and pressure drop. With a large chevron angle, small corrugation pitch gave lower heat transfer coefficients than a wider corru- gation pitch. For small chevron an
