Heat Exchanger Design Handbook PDF

2. heat exchangers Shell and Tube Exchangers ‘A heat exchanger is a device in which two fluids flow against the opposite sides of a solid boundary wall which separates them while permitting heat to pass from the hot to the cold fluid. Of the various types of heat exchangers the shell and tube type is the most widely used in oil refineries and chemical plants. It consists of a number of tubes enclosed in am outer circular shell. One fluid flows inside the tubes and the other fluid flows outside them. This discussion will be confined to applications in which ‘one fluid flows inside the tube and the other flows outside perpendicular to it, ‘The fluids considered here are those which do not change state—that is, all of the heat trans. ferred goes into changing’ the temperature of For Detall Of Joint See Fig. to Chonnet cover Gasket Bottles Tube Tobe Sheet Outtet Impingement Plote Tie Rods & Spacers. the fluid. The term fluid includes both liquids and gases. Units in which one oil gives up its heat to reheat another oil are known ss exchangers and those in which oils are cooled by means of water are known as coolers. The following discussion applies to both. Construction Tn heat exchanger design it is desirable to place the maximum number of tubes in the enclosing shell. Basically, a shell and tube heat exchanger consists of (1)'a tube bundle which is circular in ‘eros: section end contains meny closely spaced tubes, (2) hollow shell which encloses the tube bundle and (8) baffles equally spaced slong the tube length which have segmental cutouts at the opposite sides of successive baffles to cause flow back and forth across the tube bundle. ‘The simplest type of construction for a shell and tube heat exchanger is known os the fixed tube sheet type (Figures 2-1 and 2-2). In this type of unit the distance between the tube bundle and the inside circumference of the shel is at a minimum. ‘The construction is simple, because the tube bundle is welded to'the shell. ‘Tube Sheet ‘Chane! Cover Inter Outet Gosker: toler Figure 2-1. Fixed tubosheet exchanger.Figure 2-2. Detail of joint at J in Figure 2-1. Heat Exchangers 29 Shatner ; Sie Se Sott Being saggenent te anal covery + soa cow Tie Rote 8 Spec vans / | caer feta sate veae d weet ot Figure 2-2. Typical lostng head, removable tubebune en die exchanger fernten at Sai ; Fat Tae thet tt Construction features which provide for expan- sion and for cleaning outside the tubes are incorporated in the floating head, removable tube bundle type heat exchanger illustrated in Figure 2.8. This type is standard for eil refinery use. Oil flowing through the shell of this exchanger enters the shell inlet nozzle and, after being directed back and forth actoss the tube bundle by means of segmental baffles, leaves through the shell outlet. Cooling water enters the exchanger through the channel inlet and flows through the tubes in the lower half to the floating head cover. Where, after reversing direction, it flows through the tubes in the upper half into the channel and leaves through the channel outlet, The exchanger shown in Figure 2-3 has a single-pass shell and two-pass tubes. A cross-sectional view of this unit near the middle of ‘he shell is shown in Figure 2-4. ‘The inside diameter of the shell is used as the reference dimension for heat exchanger design. For example, the unit shown in Figure 2-4 has a shell with an outside diameter of 20 inches and an inside diameter of 19-1/4 inches. These are the dimensions of 20-inch, Schedule 20 pipe, which is cheaper and easier'to use than flat steel plate which would have to be rolled into a cylinder and welded together to form the shell. Figure 2-4. Cross section of 20-in. exchanger showing shell, tubes, baffles and tie rods.30 Equipment Design Handbook After the exchanger size is set, the shell must be filled with the maximum number of tubes. in ‘accordance with the process and mechanical requirements. Figure 2.5 shows the tube bundle ‘which, except for the stationary tubésheet (E), is inserted in the shell. The exchanger consists of the twogubesheets (D and E), tubes (F) and batfles (G) Tube ends sre rollerexpanded into each tubssheet. The baffle is held in place by means of tie rods (H) and spacers (3). ‘The floating tubesheet closure is shown in Figure 2-6. Its function is to ditect the water flow from the first to the second tube pass while preventing contact of cooling water with the fluid in the shell. The floating hesd cover must be removable for two reasons: (1) to allow the tube bundle to be pulled out of the shell and (2) to permit access to the tubes for cleaning. A typical Bolted connection consisting of the floating tubesheet cover (C), spacer ring (A), backing ring (B) and gasket (0) is illustrated. ‘As shown in sectional view DD of Figure 2-6 the backing ring is in two halves to make it easy to remove it from the tube bundle. The use of this type of floating head and gasket means that the ‘area of the floating tubesheet covered by the ‘gusket is not available for tubes. ‘Figure 2-7 shows the unit partly assembled. The ‘tube bundle has been placed within the shell by introducing the floating tube sheet (D) at the channel end of the shell, which has the No.3 ‘Because the tube bundle fits in the shell closely, it is clear that the shell diameter is too small to ‘accommodate the floating head cover and backing ing. Therefore, the shell terminates just before the floating head, and a shell cover (K) is used to cover co reece i Figure 2-5. Details of tubebundle. at Up cover. js ne Tone SHCET ETON Figure 2-8, Details of channel.the floating head. The diameter of the No. 4 shell ‘flange has been increased so that it can be bolted to the No.5 shell cover flange. The channel is shown bolted to the No.3 shell flange at the stationary tubesheet (E) in Figure 2.8. ‘Tube Pitch Triangular tube pitch is suitable only where it will not be necessary to clean outside the tubes either manually or mechanically, since this type of tube spacing does not permit access to the tubes. However, the tubes may be cleaned chemically by circulating a suitable solvent through the shell without dismantling the unit. Square Pitch ‘Square tube pitch provides for the insertion of a flat cleaning tool between the outsides of the tubes. This may be done continuously in either of two directions. In one arrangement, the tubes are in line with the flow. In the diagonal, square pitch, the tubes are staggered, causing the flow to divide as it meets each tube. Partition Plates The velocity of the fluid inside the tubes is a function of the number of tubes in parallel. To reduce the number of tubes in parallel, partition plates are used. For example, in the exchanger shown in Figure 2-3 cooling water flows through only half the tubes in parallel. Dividing the total number of tubes into two halves is accomplished very simply by welding a flat horizontal partition plate in the middle of the channel. Baffle Opening ‘The baffle opening location is fixed by the fouling tendency of the fluid being handled. Figure 29a shows a vertical baffle cut, Figure 2-9b shows ‘a horizontal cut, and Figure 2-8e shows an oblique (45°) cut. In each of these figures, the segmental area bounded by the solid line represents the baffle ‘opening, and that bounded by the dotted line representa the opening in the next succesive le. ‘The baffling arrangements of Figures 2-90 and c are suitable for fluids which do not deposit dirt or sediment. If sediment were deposited in the exchanger shell, these types of baffling would retain it, build it up and clog the shell. The arrangement of baffle openings shown in Figure 2-90 should be used with dirty fluids because if silt or sediment drops out in the shell, it tends to be flushed with the flow. For dirty fluids, then, the flow should be from side to side (horizontal), and for nonfouling liquids it may be in any direction — horizontal, vertical or diagonal. leat Exchangers 31 OS@ Figure 2-9. Location of baffle opening, Heat Transfer ‘Thermal Design, Basic Equation ‘The relation governing heat transfer in exchangers is Q@ = UAAt where: A = Effective external tube surface, sq.ft. Q Heat transferred, Btu/hr. U_ = Overall coefficient of heat transfer, Btu/hr. sq.ft. °F. At = Effective temperature difference, °F For a liquid being heated or cooled (@) the duty required to be performed by the exchanger is known at the outset, Q@ = wWo,(r-t where: C, = Specific heat of fuid at average temperature, Btu/lb. °F t= Lower temperature, °F Tr Higher temperature, °F w Rate of flow, Ib./hr. Temperature Difference For two fluids in either parallel flow or counterflow, the effective temperature differential is the logarithmic mean temperature difference (LMTD). _ Oty At TMTD ~TT(At [Atay where: 4t = Temperature difference of fluids at one end of exchanger (larger value) At, * Temperature difference of fluids at other end of exchanger (smaller value)32 Equipment Design Handbook Or, if the temperature of the incoming hot stream is T, and exits at T, and the cooling water enters at ¢, and exits at £2: (Ty = th) (Te - ta) h-hh . Tt LMTD = ‘A nomograph is shown in Figure 2-10 that will solve the above equation. To use the nomograph, simply connect the appropriate values on the outer scales by means of a straightedge and read LMTD on the middle scale. In baffled heat exchangers such as the one in Figure 2-3 flow is not true counterflow since the shell-side fluid continually passes back and forth across the tubebundle. For multipass tubes the relation is ret ak R-1_8i-PR @Py-1-R +R + Ip 108 (gyP)-1- R- (RF + where: F = Factor by which LMTD must be multiplied to obtain the effective temperature difference P= (tg- (Tr - fh) R= (T,~ Tea ~4h) ty = Inlet temperature of tube-side (or cold) fluid | tz = Outlet temperature of tubeside (or cold) fluid 7, = Inlet temperature of shell-side (or hot) fluid T, = Outlet temperature of shellside (or hot) fluid ‘Values of F which have been calculated for various values of P and R are plotted in Figures 2-11 and 212, For twoshell passes and multipass tubes, the factor by which the LMTD is multiplied is expressed in the equation in the box below. Log Mean Temperature Difference ‘A simple method of obtaining LMTD for any given temperatures of T, and 7, follows. (Let Ty be greater than T; ): Step 1. Find 7; /Ta. Step 2. Read the value of multiplier M from Toble 2-1 or 2-2. _Siep 8. Obtain LA'TD by milipiying 7 with M where M=((T,/T2)— 1] {ln (71/791. Example. Calculate LMTD if the temperature drops measured at the ends of a liquiddiquid cooler are 60°F and 17°F, respectively. Step 1. T, /T = 60/17 = 3.53 Step 2. From Table 2-1, M = 2.01 Step 8. LMTD = M Ty = 2.01 (17) = 84.17 Example. Calculate LMTD if temperature drops measured across a condenser ends are 193°F and 18°F, respectively. Step 1. T, /T = 193/18 = 10.7 Step 2. From Table 2.2, M = 4.05 Step 3. LMTD = MT, = 4.05 (18) = 72.90 Effect of Clearances on F Values for F (the mean temperature difference correction factor taking into account leakage in the system) were derived assuming that the shell fluid is perfectly mixed at the cross sections taken perpendicular to the general flow direction. ‘Actually, clearances around batfle tube holes and ‘between’ the shell and baffles prevent perfect mixing. (Text continues on p. 36.) qt say* | Ae P R= 1) R 2 tog HPD 2 = B+ GIP) [- Py - PRY + (R? +1)" 108 (a/P)- 1- R+ (2/P) (PMA - PRY — R? +1ty 1004 200 4 10o~| 20 unto 54000 [100 ty + Estar terminal tompmratue ate ference i tg = Other terminal tempat its F ference LamTD = Logarithmic mean vemperature diffrence Heat Exchangers 33 dig F- 400034 Equipment Design Handbook ae) a two or more even-numbered tube passes). Figure 2-11. MTD correction factor (one-pass shel Lach Figure 2-12. MTD correction factor (one-pass shell, two or more even-numbered tube passes).Heat Exchangers 35 Table 2-1. M Values for 7,/7, 1.00 to 9.99 T oe + eB el wl ela Varela eieidharg i [ig | is | |e | im | ie SLE EL Ea aL Ei : aan mee erate te | |e BUELL EEG |) BLE srerirealdl glad am | ai | i | de | St | os Peete eased Table 2-2. M Values for T,/T; = 10.00 to 99.9 eet a ae atte teeta ete B/E ETE} HEE eee = Ree Ge uae eee 4 We GE TRE [i ane aearar arte SHE ER BB Be | ee | ae Sa BLE RUE aR rie (ee ae tee ee Tae aaa ela at ete i eS ue a ae vie ele gla BY ARB | HB | 23 | EB [BS | Be ETL RT RL RL ETA The effect of leskage for one-pass shell exchangers with any even number of tube passes can be shown by a serles of curves. The temperature efficiency factor of an exchanger, P, can be calculated a8 follows: Ps (te tT- t) where: T, = Temperature of the by-passing shell- side stream, °F t, = Tube fluid temperature entering a baffle section, °F t, = Tube fluid ‘temperature leaving baffle section, °F Figure 213 shows values of F vs. P for an ideal system. Figures 2-14, 2-15 and 2-16 are for a two-baffle system with leakages of 20%, 40% and 60%, respectively. Figures 2-17, 2-18 and 2-19 represent exchangers with 60% fluid by-passing in systems containing 4-, 8- and 16-baffle sections, ly. In Figures 2-17, 2-18 and 2-19, observe that as the number of baffles is increased, the correspond- ing values of F, for the same condition, increase. For systems containing four, eight and sixteen baffies, only plots for 60% leakage were drawn because the effect of low leakage rates on F tends to decrease as the number of baffles in the heat exchanger unit increases. Figure 220, which illustrates this effect of low leakage rates on F, consists of a composite set of graphs in which curves for two, four, eight and sixteen baffles are plotted for an arbitrarily chosen P of 0.5 and R of 1.0 against varying fractions o' leakage (R is the ratio of stream heat capacities). As the number of baffles is increased above sixteen, the curves of F vs. leakage fraction tend to become progressively flatter, reflecting a declining dependence on the leakage rate. A question which may arise is, “Where does all this lead?” There must be a compromise between the improved performance resulting from chosen clearances and the added cost of reducing these clearances. Baffle tube hole and shell-to-baffle clearances are currently shown in the Standards of the Tubular Exchanger Manufacturers Association (TEMA). An engineering evaluation of the compromise between closer clearances and higher ‘manufacturing costs should be made. (Text continues on p. 40.)36 Equipment Design Handbook 10 os] os] o7| 08 @ 08) 03] 02| ou Figure 2-13. F vs. P, ideal system, 0.0 leakage. Figure 2-14.-F vs. P, two baffles, 0.2 leakage.Heat Exchangers 37 Figure 2-15. F vs. P, two baffles, 0.4 leakage. Figure 2-16. F vs. P, two baffles, 0.6 leakage.38 Equipment Design Handbook Lo og 0} o7| os = 08} oa] 03] oa} ou Figure 2-17. F vs. P, four baffles, 0.6 k Figure 2-18. F vs. P.Heat Exchangers 39 BAFFLES 2 4 86 LEAKAGE FRACTION Figure 2-19. F vs. P, 16 baffies, 0.6 leakage. Figure 2-20. F vs. leakage fraction.40 Equipment Design Handbook Example. Given: Shell exit temperature, Tow = 100°F; tube fluid temperature, f, at of tubes = 71°F; overall heat transfer coefficient, U, multiplied by the surface area of the exchanger per baffle section, A. (eq.ft.) or UA, = 2,000; shell flow mate, W, (Ib./hr.) multiplied by the specific heat of the shell fluid, C, (Btu/lb. °F) or WC = 10,000. Ratio of stream heat capacities, R, equals the tube fluid flow rate, w (Ibs./hr.) times the specific heat of the tube fluid, c, (Btu/Ib. °F) divided by WC equals one, or R = wo/WC = 1; magnitude of the by-pass stream expressed os @ fraction of total flow, ! = 0.6. ‘Find: (1) The temperature profile for a one shell ‘pass, two tube pass heat exchanger with four baffle plates and (2) F. Solution. _we+ UAs/4 _ 10,000 + 500 | UA./4 500 we- UAg/4 _ 10,000 - 500 _ TAA 500 XY (1- I) + (X+ Y)R = 21 (19)(1 - 0.6) 19 Ye * + (21 + 19)(1.0) = 199.6 XY (1-)- (X+ YR (1-2) = 21(19)(1 ~ 0.6) - (21 + 19) (1.0) [1 - 200.6)] = 167.6 ‘Average shell-side temperature at preceding baffle: 199.6 Tous = 2ROL= Oth + Vt . 167.6 __ 199.6 (100) - 2(4.0)(1 - 0.6){(21)71 + (19) 71) 1676 = 105.540°F = G2 T + Tow oe a1 _ (1+ 2(0.6)] 105.54 + 100 1-06 = 197.23°F ‘Outlet and inlet tube side temperatures at preceding baffle: 1 Vtg #20 aq, 19(TL) + 197.28 a ty x 197.28 Shell-side temperature at preceding baffle, before mixing: IT, _ 100 - 0.6 (105.54) _ Tey [roe 7 81 6eF AL = Ty - t;'= 105.54 - 68.09 = 87.45°F A2 = T,- t,'= 91.69 - 68.09 = 23.60°F Ratio (A1/A2)nuet_ = 37-45/28.60 = 1.59 < 2.0 Outlet ratio = AL=7, - ty’ = 108.54 - 73.63 = 31.91°F A2= Ty ~ ty'= 91.69 - 78.63 = 18.08°F Ratio (A1/A2)eutet= 31,91/18.08 = 1.77 < 2.0 ‘Therefore, replace T, by Ty, t2' by te and ty’ by t. The following values are obtained: T, = 112.18°F = 16.51°F hl = 64.33°F Inlet ratio = 47.85/81.25 = 1.58 <2.0 Outlet ratio = 35.67/19.06 = 187 < 2.0 Using this calculation method, the temperature profile in Table 2-3 may be constructed. Vetues in parentheses are for the same system calculated 96 though it were ideal, i . = 13.63°F = 68,09°FHeat Exchangers 41 Table 2-3. Temperature profile one Shell Temp., 7 Tube Temp., to Tube Temp., t, 4 105.54 (106.46) 78.63 (74.01) 68.09 (67.61) 3 112.18 (114.44) 176.61 (77.53) 64.88 (63.10) 2 120.21 (124.35) 79.72 (81.52) 59.61 (67.17) 1 129.77 (136.74) 88.26 (86.19) 53.48 (49.45) 0 143.62 (152.31) 88.84 (91.75) 44.72 (89.44) ‘Note: Values in parentheses are for the same system calculated as though it were ideal, i.e., 1 = 0. ‘The actual mean temperature difference corec- tion factor, F, may then be calculated: = We (te ~ t1)_10,000 (88.34 - 44.72) UA (n+ 2000 (4+ 1) = 43.62°F (52.31°F) Atiou mean = Tt, = 100- 44.72 Atimean = 55.28°F (60.56°F) 88.34 - 44.72 P= (ta - GT- 4)" 59 gano = 0.4410 (0.4634) and 43.62 F = Atwean! Atos mean * 3e pg ~ 0-789 (0.863) ‘Temperature Gradient ‘The temperature drop from hot fluid to cold fluid is proportional to the resistance to heat flow. At, = (/R)(th ~ te) ‘The temperature at the inner tube wall is expressed by the relation we Ft2) ates where: r= Any component resistance Yai = Resistance of dirt or scale on inside of tube 1 = Resistance of fluid inside tube R= Total resistance Average temperature of cold fluid Average temperature of hot fluid ‘Average temperature of fluid inside tube t, = Temperature at inner tube wall At = Effective temperature difference in unit At, = Temperature drop across r Weighted Mean Temperature Difference For the more complicated shel! and tube exchangers (condensers with superheat and/or subcooling, vaporizers with supetheating) the usual methods of calculating At or MTD do not give correct answers. In counterflow or multitube pass exchanger, it is not difficult to calculate the true At if only sensible heat is involved. If there is an appreciable amount of latent heat in the system and the overall heat transfer rate is not fairly constant, the usual method of calculating MTD is not valid. Then it is necessary to break the heat transfer calculations into zones and calevinte what is commonly called a weighted MTD. ine normal sensible heat At is usually called a straight line MTD. Some services that usually require weighted MTD’s are: 1, Overhead condensers with steam and hydro- carbon condensing 2.Feed-effluent exchangers with change of phase ‘8, Amine ovethead condensers 4. Condensing steam out of a noncondensible 5. Pure component condensers with subcooling 6. Condensers with large desuperheating zones (a) Refrigerant ot chemical (b) Steam 7. Pure component vaporizers with superheating.42 Equipment Design Handbook Heat Release Curves Usually, a prerequisite for determining a weighted MTD is a heat release curve. It is used for a guide in determining where to zone the heat, transfer calculations. A heat release curve is a plot of heat load vs. temperature of both shell and tube flue. An example of this can be seen ln Figure 1. ‘The heat release curve is then broken up into zones. Temperatures for these zones are selected 80 that straight lines drawn between the temperature points will approximate the curve. The greatest care should be taken when the shell plot and the tube plot come the closest together. Usually, three to five points are sufficient if a curve of this nature is done by hand, If the MTD is being calculated on the computer, more points are usually used. Shell-Side Condensation, Multitube Pass Figure 2.22 illustrates a temperature vs. length plot of a typical example for two-tube passes. As can be seen, there are five unknown tube-side temperatures’ when the plot is divided into only three zones. Solving for these would be a tedious trial and error solution. If there were just a small curvature for the condensing fluid, the zones would be figured as a counterflow exchanger and then the normal overall correction factor would be used in each zone. This assumes the correction factor for each zone is the same as the overall. With a large hump in the condensing curve, you cannot apply the same correction factor to each zone; nor can you take each counterflow zone and calculate a correction factor from the counterflow temperatures. This would give an inaccurate At, particularly in the low At zone next to the shell outlet. ‘This method gives good results and requires a minimum of time, Here the terminal tube temperatures (t) ond fj) are used with each zone to calculate both the MTD and the correction factor. For the limiting case of single phase, linear fluids it will give the same result as applying the customary conection factor to the overall counterflow MTD. ‘The method checks closely with the long trial and error method, even when the hot and cold ‘outlet temperatures are the same. ‘The method assumes that there are no downflow {crossflow) components in a horizontal exchanger. ‘This would tend to give a higher correction factor. It would be more exact for vertical exc ‘where the liquid level is at right angles to the tubes, See Table 2-4 for calculations. w TEMPERATURE WEAT LOAD Figure 2-21. Heat release curve. TEWPERATURE TenoTe Figure 2-22. Temperature curve for one shell pass, two tube pass exchanger. asweighted MTD Qrocat @ @ @ G).* @).* Ge). 8,870,000 i ” 1g80M | 100m” 650m ~ 91°F 136.5 95” 348 This is the overall corrected MTD. It is not as important as being sure to use the correct temperature difference in a given zone with the rate for Heat Exchangers 43, that zone, This method is used when the exact tube-side terminal temperatures are known. There- fore, if it is used with shells in series, the interim temperature between shells has to be predetermined. This can be done graphically. Graphical Solution for Shells in Series When calculating weighted M7D's it is important to remember the following points: 1. If curves bow out (convex), the MTD factor will be greater than a straight li 2. The comection factor will be less than a straight line if the curves bow in (concave). Table 2-4, Calculations for Sh I-side Condensation, Multitube Pass Zone Heat Loads, Btu/hr. 1 T, =261~- 215 = 7, 1,880,000 2 T, =216- 170* 75 1,110,000 a T, =170- 110= 7, ‘690,000 Cooling water enters at 80°F and leaves at 110°F. Zone 1 Hot Side Cold Side Difference ty = 251 Hot Temp. t, = 110 ty te 2141 fog 215 Cold Temp. ty = 80 fog > fy = 185 At = 36 At= 80 ate 6 R= 30/36 ~ 0.833 X= 36/171 = 0.21 At, = 188°F, corrected = 136.6°F Zone 2 Hot Side Cold Side Difference fen = 215 % =110 tea ~ te = 105 tee =170 4 = 80 toa t= 90 At = 45 ate 30 Ate 15 R= 0.665, X = 0.333 ‘Ate = 97.5°F corrected = 95°F Zone 3 Hot Side Cold Side Difference tea = 170 t =110 tea - t= 60 t =110 4 = 80 teh = 80 at = 60 At= 30 At= 30 ‘Aty = 48.2°F corrected = 34.8°F44 Eg ment Design Handbook If the normal method of determining the correction factor is used, it will give too high an MTD. Unless the unit had a safety factor, it would be underdesigned. If there is a temperature cross, where shells in series are required, the heat curve can be used to determine the minimum number of shells. This method is based on setting a minimum correction factor of approximately 0.8 or where the hot stream outlet temperature is the same as the cold stream outlet. Refer to Figure 2-23. Line T; 7 is the hot fluid line. Line tts is the cold fluid line. Draw a horizontal dashed line through temperature point t until it intersects the 7; Tz line. This intersection is shown ss temperature T, . From this point drop down vertically and intersect the cold fluid line. This is shown as f. ‘The four temperatures T,, Ta, ty ond te reptesent the terminal temperatures of the hottest exchanger. This procedure sets the hot fluid inlet temperature, T,, equal to the cold fluid outlet temperature, t,. Setting T, lower would give a temperature’ cross in this exchanger. ‘Temperatures T, end t, will also be the hot terminal temperatures of the exchanger under the hot one. To find the cold terminal temperatures, proceed as before by drawing a horizontal tie line from Ty to intersect the hot fluid line. This will be at T,. Then other temperatures will be below this, at ty. Continue until a horizontal tie line intersects Tz or comes out lower than T;. If the tie line inter- cepts T,, the minimum number of shells is the number of horizontal lines. If the tie line comes cout lower, the last line is still counted even though it misses 7. In Figure 2-23, the minimum number of shells is three. ‘This procedure is quite similar to that for finding the minimum number of shells in series. ‘The differences are as follows: 1, The dashed tie line between the hot side and cold side outlet temperatures usually has slope instead of being horizontal. 2.t is necessary for the last tie line to intersect the hot stream, final outlet temperature, 73. Geometrically, set the angle between the tie line and a horizontal line the same for all shells in series. This angle is determined by trial and error. Example. Figure 2-24 is a heat release curve for ‘a combined feed exchanger in a Platformer. For the first trial, run through the procedure for finding the number of shells in series. If the last tie line intersects T,, you need no more trials. ined Figure 2-23. Minimum number of shells dete ‘graphically If the last tie line just misses T,, only a small adjustment is required, Start again by giving the tie line a slight slope. Use the same procedure as before except give all tie lines this same slope. Keep adjusting until the last tie line intemects 7. ‘A reconimended procedure for giving all tie lines the same slope is as follows 1. Place two triangles on the curve with the edge of one on the slope line. Hold the other triangle firmly on the curve and slide the triangle on the slope line down to a new position. The triangle that is being firmly held may be rotated until a tic Tine can be drawn through ¢,. Continue. 2. The vertical dash lines indicate temperatures between shells. As can be seen from Figure 2-24, the number of shells in series required is four (it counterflow is not used). 3. Pick out one of the exchangers represented on the curve. It is best to pick out a middle exchanger and one where the heat release curve is the easiest to read if the heat release curve is quite erratic. Choose four adjacent temperature points as represented by intercepted points on the curve. All will give approximately the same correction factor. For example, the terminal temperatures of the hottest shell will be T,, T,, tz and t,. For purpose of illustration, we will use these to calculate the correction factor.HORIZONTAL LINE $38 8 TEMPERATURE °F 8 01020304580" 60708090, 01 ‘HEAT LOAD, 108 BTU/HR, . Tube Dimensions a Sat |S ae ogee "ER oS seat) Bae [eee (ton, | He) 2S [ease TT be u fe we | a4 [oot|notefese i ee Vo Far paetey reterred to water at 60°F, Floater Fe iBea, Rate of Flow, bse per Tube Using the four points, determine F as you would normally. RB = (984 - 822)/(803 - 632) = 0.948 P= (803 ~ 632)/(984 - 632) = 0.486 F = 0849 Weighted MTD will be (181) (0.849) = 153.5°F. Heat Transfer Surface In a shell and tube exchanger the heat transfer surface is the total surface of all the tubes ‘separating the two fluids. In practice this surface is 45 Heat Exchangers i ws 20 10 OO 169 10180” Bo taken at the outside of the tube. For example, a Linch tube of any gage has a peripheral length of ‘"D inches = 3.1416 inches. In a 1-foot iength, the surface is 87.8 square inches = 0.2618 square feet. For the exchanger, A = total number of tubes (tube length, ft.) (a: of surface) ‘Table 2-5 lists dimensions of commonly used tube sizes, Coefficient of Heat Transfer As heat flows from hot fluid to cold fluid through the metal tube wall it is opposed by five resistances: ro = Resistance of fluid outside the tube 2.rao = Resistance of dirt or scale on outside of tube 3.%m = Resistance of tube wall 4.rg; = Resistance of dirt or scale on inside of tube 5.7, = Resistance of fluid inside the tube ‘These five resistances are added to obtain R, the overall resistance to heat transfer. Rr + rao Fm (DiDrm ) + Fa: (D/D)) + % (D/Di) where: D = Outside diameter of tube D, = Inside diameter of tube Dy = Logarithmic mean diameter of tube DD ind)46 Equipment Design Handbook Of the five component resistances, Fm has little significance when 1/R is less than 125, as is usual fn oil refinery exchangers. Values of Tao and fai, which allow for the amount of fouling expected Outside and inside the tube within a reasonable period, cannot be predicted with any degree of certainty because of the different types of fouling deposits. These values are best obtained from experience with other units in similar type of uty. Values of r. and 7, depend on the physical properties of the fluid and its velocity. Since ‘drop and velocity are interrelated, it is Customary 10 start with the allowable pressure drop as a base and determine the maximum fluid Velocity permissible without exceeding this pressure drop. The higher the fluid velocity the lower the values of r, or 71, ‘Often, the term coefficient of heat transfer (U) is used instead of resistance. One is merely the reciprocal of the other. U_ = Overall coefficient of heat transfer hy = Coefficient of heat transfer of fluid outside the tube hao = Coefficient of heat transfer of fouling outside the tube Im = Coefficient of heat transfer of tube wall hu, = Coefficient of heat transfer of fouling inside the tube hy = Coefficient of heat transfer of fluid inside the tube ‘Tube Wall Resistance » Bt &) where: k= Thermal conductivity, Btu/hr. sq. ft.°F/ft. t = Tube wall thickness, ft. Fouling Factors For flowing cooling water, a fouling factor of 0.003 is generally used. For clean liquids and clean. condensing vapors, a fouling factor of 0.001 is used. For liquids, the fouling factor may range upward from 0,001 depending on the service. Values of rz, must be referred to the external tube surface by multiplying by (D/D,). Coefficient of Heat Transfer for Flow Inside Tubes ‘The Sieder and Tate relation involving coetficient of heat transfer, mass velocity, physical properties of a fluid and inside tube diameter is thown in Figure 2.25 in terms of dimensionless groups with Reynolds number as abscissa. It will be Seen from Figure 2-25 that there are three distinct zones of flow. The first is the streamline region for values of Reynolds number of 2,100 and less. The Series of parallel lines is expressed by the equation shown in Figure 2-25. “At Reynolds number values of 10,000 or more, turbulent flow occurs and the equation of the straight line is 2 som C2)" (8) (F In the intermediate region where Reynolds number varies from 2,100 to 10,000, the relation ‘does not follow a straight line. Specific heat, Btu/lb. °F. ‘1D, = Inside tube diameter, ft. G, = Mass velocity inside tube, Ib./hr.sq ft. hh’ = Coefficient of heat transfer, Btu/hr. sq.ftoF. k= Thermal conductivity, Btu/hrsqftF./ ft. = Viscosity at average fluid: temperature, Ib. fhr ft. du’ = Viscosity: at tube-wall temperature, 1b... ‘To determine whether flow inside the tube is in the streamline, turbulent or intermediate region, it is necessary to evaluate the Reynolds number. This is done using Figure 2-26, which is based on the equation Ne = 6.31 (w/p'Dy') where: ‘Dy = Inside tube diameter, in. Npe = Reynolds number w= Rate of flow, Ib./hr./tube y’ = Viscosity at average fluid temperature, centipoises BeyHeat Exchangers 47 100 © + Speci Heat, Djelnaide Tube Diemeter feMees Velocity mslde Tobe, 1 eCoettcer OF Heat Transfer, ¥ © Theemol Conductivity, BTUMIneXea M NPE) L sLenat, Feat 1» sViscosiy Ab Avaroge Fld Temperate, (it) Visas At Tube: Wall Temperate, —(o/(or tt) Bru nee) Feet Ie ediontd BTU/(o sa tPF) Jac 7 Figure 2-25. Coefficient of heat transfer, fluids inside tubes. Instructions for using Figure 2-26 and an example are given in Figure 2-27. Water flowing inside a tube is a most frequently encountered condition. For water, the equation for ‘AD,[k is applicable to Reynolds number values of 2,100 and higher. Figure 2-28 is an alignment chart for the solution of hD,/k after multiplication (Di/D,) to refer the coefficient of heat transfer to the external tube surface. Instructions for and an gzample of the use of this chart are given in Figure Coefficient of Heat Transfer for Flow Outside Tubes For industrial heat exchangers using segmental baffles with unbored shells, the coefficient of heat transfer for the shell-side fluid is expressed in the following equation: ee) AD 7 082 16000 796,500 where: G, = Weighted mass velocity = (G,G,) Ib fhr.fsq.tt. G, = Crossflow mass velocity, Ib./hr.sq.ft. G, = Mass velocity through’ baffle opening, based on the area of opening less area of the tubes passing through it, Ib./hr./sq.ft. Alignment charts permit rapid solution of this equation for fluids commonly encountered. When water flows through the shell, Figure 2-30 is used to obtain the coefficient of heat transfer. To use the chart, simply connect the appropriate values on. the two outside scales with a straightedge and read the value of h, at the point of intersection of this, Tine with the inner scale. When a petroleum oil flows through the shell, Figure 2-31 is used to determine the value of’ the coefficient of heat transfer. The use of this chart is illustrated in Figure 2-32.pcae ls side diameter, tube or pipe, inches. Ngo" Reynolds number, dimensionless fee Rate of flow, Wo tm tbe) ul= Viscosity, centipoises. Figure 2-26. Reynolds number (inside circular cross section). : 4 R 8 8 . % E 8 8 : F-20000 F t 6 W Eyggre Epo Eveco Es z Fem § oo fee 5 bspoo g t E E P2000W.Nre Heat Exchangers 49 DYRECTIONS FOR USE; in 1: Cone the snap sige ee and g!acies wih sighedpe vot so thou elk obec ine tes he snared i Setersects Dy’ = 0.62. 3—Reod Nav = 12,700 Figure 2-27. Directions for using Figure 2-26. Graphical Solution to Heat Transfer Coefficient Rating hest exchangers without the use of a computer is time consuming, because it must always be done on a trial and error basis, The task is even more tiresome if the rating method in question is a laborious step-by-step computation, 4s is required for the correct results. ‘A graphical method has been developed which allows prediction of the overall heat transfer coefficient within a limit of 43%. This suggested short cut will, in many instances, make a laborious calculation unnecessary. A study of exact methods showed some remark- able interrelations. In Figures 2-334 and b, for instance, the Reynolds numbers and shell-side heat transfer ‘coefficients (h,) are plotted at constant pressure drop against the ratio of baffle spacing to shell diameter, ['p/d,, for various shell sizes and for various fluids, all with 16-footong heat exchanger tubes, The Reynolds numbers are corrected to the cross-sectional flow of the fluid. All calculations are carried out at a window cut ‘of 20%. This seems to be an optimum figure as it produces the maximum heat transfer coefficient at given pressure drop for all tested heat exchangers. It is further assumed that TEMA standard clearances would exist between the tubes and baffles and between the shell inside diameter and baffle outside diameter. The various numbers in Figures 2-334 and b denote the fluids, pressure drops and ‘exchanger dimensions given in Table 2-6. Figure 2-33c shows that the Reynolds number ‘on the shell side of a heat exchanger is practically constant at constant pressure drop, independent of the baffle spacing. This means ‘that in a heat exchanger, the throughput of shell-side fluid can be altered proportionally to the baffle spacing with- ‘out having the pressure drop changed. ‘The shellside heat transfer coefficients, on the ‘other hand, increase with a growing baffle spacing as seen in Figure 2-336. Because of the constant erossflow Reynolds number at various Ly /d, ratios, the velocity through the window sections rises as the baffle spacing increases.50 Equipment Design Handbook | fawGe Tube 9058, Sirangham wre goo | DI = Oulsice te siometer, inches. yj © Coefficient of heat torsfr referred fo OUTSIDE of ube, TU./ (RAAMDITF) | 1 = Bvarage woter temperature, * = Rote of tow, fe / Wwituved Figure 2-28. Coefficient of heat transfer, water inside tubes {N»,>2100).oO BWs, DIRECTIONS FoR Us ih athe ne Lite 1: Connect values on Le when 2,000 The sat date tide @ 1" QD. 16 BWG tube. 1—Conect » 2000 Figure 2-29. Directions for using Figure 2-28. ‘This velocity rise increases the average value of the shellside coefficient, as could be expected. Furthermore, the small leakage rate at a large baffle spacing improves the heat transfer conditions to a certain extent, The increase of the heat - transfer coefficient, however, is remarkably constant for very different kinds of fluids and for different shell diameters. In caloulating the values for Figures 2-83a and 5, * it was assumed that sealing strips would be Piprided at both sides within the crossflow section. e calculations show that Heat Exchangers 51 the heat transfer conditions to such an extent that the effect of the actual tube it (triangu- Jar, square or diamond pitch) is of minor impor- tance, In fact, the heat transfer coefficients are practically the same for the same pressure drop if the vessels ate equipped with sealing strips, independent of the tube arrangement, All these facts allow the development of a direct and ‘comparatively simple equation of the shell-ide heat transfer coefficient. If we assume that a baffle pitch in inches (Lp) divided by the shell diameter in inches (d,) is 0. Le. Ly/d, = 0.3 —a constant of 1,825 can be used in the following equation: 0.8 6 0.88 20.87 7 = 1395 tense Gp where: h, = Shellside heat transfer coetficient, Btufhragft.°F 4 = Ratio of absolute viscosity of the fluid, 4, Ib,/ftfhr, to the absolute viscosity of the fluid at the heat transfer surface, n,, 1b./ft. fhe. 5 = Specific gravity of shellside fluid (water =1) Ap, = Shell-side pressure drop, psi C,' = Specific heat of the shellide fluid, Bulb. K = Thermal conductivity of shell-side fluid, Btu/hr. °F = See definition under “Coefficient of Heat ‘Transfer for Flow Inside Tubes” above. If the baffle spacing is different from Ls /d, = 0.3, an additional correction to the shallside coefficient, h,, must be carried out using Figure 2.84, The heat transfer coefficient of the tube side for Reynolds numbers above 20,000 can be expressed in a similar way: he [yap] 6,088 5287 Fm 2860 ee where: hy = Tubeside heat transfer coefficient, Brufhr.sq.tt.°F Ap, = Tubeside pressure drop, psi Figures 235 and 2-86 represent the above equations in nomograph form. Ap, is the pressure drop on the shell side with 16-foot-dong tubes. If the tube length differs and is L ft. long instead, the permissible pressure drop Ap must be multiplied by a factor 16/L: Ap, = Sp (16/L) in order to use the shell-side equation or Figure 2-35. {Text continues on p. 55.)t ho Ge (400 Hy 4,000 eB [300 -s,000 F-s0o b f-s00 F 250 c }-2,000 F400 F200 [300 [200 150 r 500 F100 Feo L 20 [400 [-s0 80 E 40 E 10 c 30 f 60 [200 Lee r 30 r | 40 I r | L | fee b 0 Lio NOTE:For bored shells increase ho 14% Chart based on %" O! tubes For other sizes multiply chart reading by oe Nias MULTIPLIER] 1076] 10 [089|(075/0') he Mass velocity through boftle opening, Iby(see)lsqtt) Gk= Crossflow moss velocity, Ib/(sec)(sa.ft) G4=(GLGb)"2, geometric mean moss vetocity, Ib/ (sec)(sq.ft) ing? Shell=side heat tronsfer coefficient, BTU. / thr(satty(*F) D's Outside tube diameter, Inches. 1 = Average water temperature, °F. Figure 2:30. Shell-sde coefficient of heat transfer for water in exchangers having sogmental bafflesHeat Exchangers 53 2p) Gro, deqees A Gh exe velocity troup ole opein, Tay teeieathy Gt* ccnteow mats vein, We /tecieatt = 16E6)1, guometnc meon mess voy, te /secisat, ns Salsa eat hanetr ceettoet, aU /esaatac) 1° vacouy of aerogetenowatur, centgoaes. Die Wacaiy ot tube-wat Yonge ot ceniposes NOTES 1 satgty ng,oblaines above, by (27s) Hq 8 3 si 2 8 8g Ltn epeenenteereg wen age p nip a pga Wrste eee chet o io eio ie Oe eta giants 2 chart based on 24°00 tubes, for other szes dinde chort reading by on | |] Dwisor [Gas [+ ‘For bored shes increase hy 13% Figure 2-31. Shell-side coofficiont of he: transfer for petroleum oils.54 Equipment Design Handbook son Gand scales out paint of eet with appro ult is read at the point where Line ai" 40 obtained should be the mater at the bottom of the DIRECTIONS FOR USE: Li Sitka straightedge, Line 2: Peat strois “ection of Line twith the Freesat AP ele, The he cole. The vole rived int ‘te Figure 2-32. Directions for using Figure 2-31 200,000 2,00 1,000 “eof «aol 100,000 80.009 ‘0.000 sec} 0000} 0009 20.009 209 Noogs os as 10 ute a Figure 2-33. Reynolds numbers (a) and heat transfer costficients (b) of various heat exchangers and fluids at Table 26. Fluids, Pressure Drop and Dimen: ‘of Heat Exchangers for Figures 2-332 and 2-335 stereree| ed \ ane ———g3 3080 Figure 2-34, Correction factor of shell-side est constant pressure drop plotted against baffle spacing. transfer coefficient.000 KEY: SUA py)-ek fom Gp PREF LINE = hy Figure 2-35. Shell-side heat transfer coettic ‘4p; represents the pressure drop of one single- tube pass 16 fect long, including the tube entry and exit losses. If several tube passes, cae necessary, the permissible pressure drop (Ap) must then be divided by n and multiplied by a factor F, from Figure 2-87 to allow for a tube length other than 16 ft: Ap, = ApF;, /n. After finding the heat transfer coefficients for both shell and tube sides, we know the overall coefficient. This, in tum,’ allows us to find the number of tubes and the shell diameter. The approximate baffle spacing Ly can also be determined by a simple formul @ = 270,000 [s(Ap,)]}* where: G "= mass velocity, Ib./hr./sq.ft. ly = ¥ - 270,000 S,,* [s(Ap,)1* * where Ly = baffle piteh, in., W = weight flow of ftuid, Ib./hr. sm? = the minimum area, in square feet, for crossflow near the centerline of’ the ‘exchanger per inch of length. Heat Exchangers 65 Sly) -=k-+p +6, — REF LINE hy Figure 2-36. Tube-side heat % ere snecrcet UGE LeNéTH, Fr Figure 2-37. Correction factor of tube-side pressure drop. ‘This equation allows the determination of the ratio Ly/d, within limits of +10% so that a decision of whether to /, can now be made. Figure 2-34 shows the correction factor of the shell-side heat transfer coefficients for the baffle spacings Ly/d, = 0.2- 1.0.56 Equipment Design Handbook Example. The following data are given: Tubes: 3/4-inch OD on 1-inch triangular pitch, 16 BWG 8 feet long. See Table 2-7 for shell-side (fluid ammonia) and tube-side (cooling water) data. Heat is 785,000 Btu/hr.; fouling resistance is 0.0015 + 0.0015 = 0.0030. LTD = 40°F. ‘Assume various tube passes and find h, and hy from the equations in Figures 2-85 and 2-36, L=8tt, F, from Figure 2-87 (at L = 8 ft.) = 1.58 Shell side: 6(Ap,) = (Ap)(16/L) = 0.00835 (1)(16/8) = 0.0067 Tube side: h, = 40.1 A passes: s(Ap,) = s(AP)(Fi/n) = 1,0(6.5)(1.58/4) = 2.49 hy = 1,225 Ato = 1,013 where hy, is the tube-side heat transfer coefficient corrected to the tube OD, Btu/hr.ft.°F 6 passes: s(Ap,) = s(4p)(F 1 /n) = 1.0 (6.5)(1.53/6) = 1.66 fy = 1,024 hyo = 847 8 passes: s(Ap,) = s(ApF,/n) .0 (6.5)(1.53/6) = 1.245 hy = 917 Igo = 158 Compute the overall heat transfer coefficient, U exchanger surface area, A, and number of tubes and fix the shell diameter. See Table 2-8. Now the tube side figures must be checked to make sure that the right velocity and pressure drop ‘are used to obtain the right heat transfer coefficient. This may be done using the following equation, 1 (WV an iL . aon iin( ) 2 foams G7 0.0053} psi Permissible pressure drop, psi Weight of fuid flow, Ibs./hr. Flow area through the tubes, sq.ft. Number of tube passes Specific gravity Friction factor ‘Tube length, ft. = Inside tube diameter, ft. Substituting example values shows that in this instance eight passes should be used, which would comply with the permissible pressure drop of 6.5 psi. Four and six passes would produce too low a pressure drop and, therefore, too low a coefficient. Tube Side Rating One of the two heat exchanging fluids has a smaller film coefficient than the other. It could be called the weaker of the two fluids. Its coefficient is, therefore, the governing factor of the overall heat transfer coefficient. To reduce the exchanger size to an absolute minimum, the weaker of the two fluids should consequently be passed through that side of the exchanger, resulting in a smaller heat resistance value at the given pressure drop. In the turbulent flow region and at the same pressure drop the tube side often offers better heat: transfer conditions than the shell side. ‘This depends, however, on various factors: tube length, number of tube ‘passes, tube diameter, physical properties of the fluid, etc. To arrive at the optimum solution in a minimum of time, use of the following calculation method is suggested. The method is best explained by examples. The equations used have been derived from well-known heat, transfer principles. ‘Example 1. The weaker medium, liquid propane, is to be cooled from 135 to 65°F. ‘The average physical properties of propane are ‘Thermal conductivity, k = 0.068 Btu/ft./hr.,°F Specific heat, Cp = 0.605 Btu/lb.°F Absolute viscosity, Specific gravity, 8 ~ 0.471 (water = Weight of flow, 'W = 400,000 Ib./hr. Heat duty, Q'= 400,000 (0.605) (135-65) = 16,950,000 Btu/hr. ‘Average temperature difference between the two heat-exchanging fluids: At = 39.2°F. Tubes: 9/4-in, OD on Lin, triangular pitch, 14 BWG (d, = 0.0487 ft.) ‘Length of one tube pass, L’ = 16 ft. Heat transfer coefficient of shell-side fluid hy ~ 950 Btu/hr. sq. ft.°F Heat resistance, C’ = 1/hp + fouling factors = 0.0020. Permissible pressure drop through tubes (excluding the pressure drop through the shell nozzles) ‘Ap = 3.5 psi Solution: Ap (s) W (At) __ 3.5 (0.471)(400,000) 39.2 Q 16,950,000 = 1.525. LHeat Exchangers 57 Table 2-7. Shell and Tube-Side Example Data Permissible fluid Ibjhr, Coe a e 3 AP Shell side: fluid ammonia 9,872 0.017 0.53 0.029 1.0 0.00335 1.0 ‘Tube side: cooling water 78,500 0.36 10 199 10 10 65 Table 2-8. Calculations of General Heat Transfer Coefficient, Surface Area, Number of Tubes and Shell Diameter Tube Passes 4 6 8 vu cp - = = 34.55 34.3 34.2 EK + Fouling Resitance 35 +7 gpg 0.008 785,000 ———— , mft. 568 572 573, A345 a0)’ ‘Number of tubes 362 365 366 Internal shell diameter, d,, In. 23-1/4 25 25 Minimum area for cross flow near centerline of the exchanger per inch of length, S,,, sq. ft. 0.0468 0.0590 0.0590 Approximate Ly 9872 ‘ 270,000 (0.0468) (0.0067)"* . . . Lsid, 0.413 0.304 0.304 ‘Tubes per pass 91 61 46 @ per pass, sq. ft. 0.191, 0.128 0.0965, Water velocity in tubes 78,500 [ec 1.83 2.73 3.62, 3600 (62.5) (0.191)68 Equipment Design Handbook From Figure 2-88, c = 0,252 where c is a factor related to physical properties. From Table 2-9, at d, = 3/4 in., 14 BWG: b = 3,045 (where b is a constant factor which mainly depends on the tube pass length, L', and the tube diameter) ‘The tube-side heat transfer coefficient related to ‘tube outside diameter is A = bcF, where Fy is a correction factor for h. Initial heat transfer coefficient: = 8,045 (0.252) = 767 Btu/brag.ft.°F (uncorrected). U_ = 1/@jh+C')=1/(1/767 + 0.0020) = 303 Btu/hr.sq.ft.°F U/h = 803/767 = 0.395 From Table 2-10 at U/h = 0.895 (by interpolation): F, = 0.784. Corrected h: ‘767 (0.734) = 563 Btu/hr. sq. ft.°F. Corrected U: 1/{(1/563) + 0.0020] = 265 Btu/he. oq. f1.°F Dimensions of tube bundle: @ __ 16,950,000 (An (29.2) 265 ~ 169 sat. Find constant ¢ from Table 2.9 at d,, = 3/4 in., 14 BWG: € = 0.419. Hence, A __ 1630 (0.419) _ [as] = Seam “1400 Find a (total flow area of one tube pass) from Figure 2-39: a = 0.50 sq. ft., from Table 2-9: a’ = 0.00186 eq. ft, r= 0.01215 ft. Number of tubes per tube pass: N = a/a’ = 0.50/0.00186 = 269 tubes. ‘assumed in this example, thus allowing for the thickness of the tubesheets. Find heat transfer factor because of the tube arrangement, C,, from Table 2-11 at 3/4-inch OD on Linch triangular tube pitch and L'= 16 ft: Cy 1,470; assume the heat transfer factor due to baifie spacing, Fp = 1.0, and find from Figure 2-40 with (Ap) s = 8.8 (0.471) = 1.65, and W(At) _ 400,000 (89.2) @ 16,950,000 « [mee] x = 0,925 @ where ¢ is a combined physical property factor _ 40279, 836 fp(g)] 822 —— ‘Therefore, the efficiency ratio of the tube to shell side coefficients E is oan p= [@ ] CFs Q 3045(0.734) ra10q.0) ~ 585 With the assumed Fy = 1.0, the shell-side heat transfer coefficient at the permissible pressure drop would have been h/E = 563/1.535 = 367 Btu/hr. sq. ft°F. This would give U = 1/((1/367) + 0.0020) "= 212 Btu/hr. sq. fF and a heat exchanger surface of A = 2,040 sq, ft., instead of 1,680 sq. ft., representing a vessel of 31-in, internal shell diameter (D,). ‘At afinal check with the constant based on tube arrangement, Cp = 1,750 (Table 211), and the minimum area for croseflow near the centerline of the exchanger per inch of baffle spacing, Am, = 9.25 sq. in, (from TEMA standards), obtain a baffle distance of w Total tube length L: b= Se i An (Ap@)ae/LV* 1-4t (¢ ) __ 1680 (0.01216) 0.0487 400,000 @ \a, 10.50 (0.0625) 1750 (0.28) [3.5 (0.471)46/16)]* : = 19.2 in. = 80.8 ft. and a factor Fy = 1.06 (Figure 2-41) instend of ao LO, so that the actual value of E would be 1,463, Neate, crete Pee eae bp or h, = 63/1.463 = 385 Btu/hr. eq, ft.°F; U= 218 ‘Two tube passes, each 16 fect long, can safely be Btu/hr. sq. ft°F and A = 1,980 sq. ft. (Text continues on p. 63.) cIdpeerWeat Heat Exchangers 59 100 10 10) jon 10 oor lo or om © oat 00 0001 SEEM gy cep enee ine Figure 2-38. Nomograph to find factor ¢.60 Equipment Design Handbook Table 2-9. Tube Data for Heat Exchangers a, Tube Outside Dismcter dg = 3/4 in. = 0.0625 ft i mate # a us ig ga | ae | iB 36 | $88 | 1 18 PRES 12 ae 382] S888] 1 1B) am(?)] FL 5 S| te a | on. » Fug | RS 3 | ae a 588 | 258 20 | 2a oe aoe | gee aS rey He | Bae Bie ae 328 | Bae i | aie ae | Se | 3a ah ite | Be a0 | 3390 ot az | Bae ‘gag | ga - Hi | fe #8) 8 a ie | Base he | 30 os ("= low of single tube? = total inside perimeter of one tube pas, in feet) Table 2-10. Correction Factor Fy Uh La Uh Ufh Pe oe | | Table 2-11. Factors for Shell-Side Rating ens Amantcomnt a [fe peepee mses | HB) Pa Hesse sent Bae as] 1a) 2a a : eamegerny maim pee | HR) Ea i eRe ARE so I t Hoosier aiaeee me. |] BAR] HB HBHeat Exchangers 61 10%, 10 10%} 0 if wo ons 10 Tt 0.001 Figure 2-39. Nomograph to find total flow ares, 9, of a one-tube pass at L' = 16 ft. tube length per pass, 2 to be mu! tiplied by correction factor f at other tube lengths L'.62 Equipment Design Handbook 100) 10 KE a oot 001 S00 0001 ea KEY Ap ee Ke Om EE a (Ha OF Figure 2-40. Nomograph to find e[WiAt)/a] °"* oa) oar of aivelTIn this example, it is quite obvious that the exchanger size can be reduced by about 20% if propane is passed through the tubes and not through the shell side. Correction of L. If, in the above example, At is 49°F instead of 39.2°F, the obtained data would then read: Initial ratio U/h = 0.373, Corrected h = 606 Btu/hr. sq. ft.°F, (corrected). U=274 Btuphr. sq. ft.°F; A= 1,260.4. ft, a= 0.456 sq. ft., L= 26.2 tt. with the number, of tube passes resulting in an odd figure of n = L/L’ = 26.2/16 = 1.64. 1, Retaining the gage of 14 BWG, the tube pass length, L' could then be reduced to 18 ft. 6 in., allowing for the tubesheet thickness. The number of tubes in each of the two-tube passes at an outside surface of 2.57 sq. ft. per tube within the tube sheets gives 1,260/[2(2.57)] = 245 tubes. 2.1 the length of one tube pass (L' = 16 tt.) must remain constant, ond if a smaller wail thickness is acceptable, the same pressure drop and the same heat duty can be obtained by using the conversion figures shown in Figure 2-42. In this example the total (effective) tube length has to be increased from 26.2 to 30.8 ft. (exchiding tube sheet thicknesses), ‘resulting in a length correction of 30.8/26.2 = 1.176. At this ratio and with Ush = 0.373, find from Figure 2-4 multiplier of d, = 1.115 or d; = 1.115 (0.0487) = 0.0548 ft. (= 18 BWG) Multiplier of A: 0.970 or A: 0.970 (1,260) = 1,220 sq. ft. ‘The number of tubes in each of the two tube passes at an outside surface of 3,02 sq. ft. per tube within the tube sheets: 1220/[2(8.02)} = 202 tubes. Example 2. The weaker medium: Butyl alcohol to be heated from 50 to 150°. Average physical properties of the fluid: k= 0.088 Btu/hr. ft.°F C, = 0.60 Btu/lb. °F H'= 4.8 Ib,/t br, 8 = 0.786 (water = 1.0) W = 400,000 Ib./hr. @ = 400,000 (0.60)(150 - 60) = 24 (10°) Btu/hr. Heat Exchangers 63 4 02s os 08 RATIO BAFFLE SPACING/SNELL. DIAMETER TO Lg Figure 2-41. Heat transfer factor Fe unso7s 025 to ‘wuurpiies ‘oF SSA uc os o7 06 8 10 12a ‘CORRECTION OF TUBE LENGTH L. Figure 2-42, Multipliers of d; and A at alterations of L, retaining the same pressure drop and heat duty.64 Equipment Design Handbook Average temperature difference between the two heat exchanging fluids: At = 50°F. Tubes: 3/4-in. OD on Lin. triangular pitch, 16 BWG, L’=16 ft. fag = 1,600 Btu/sa. ft. br., °F @ = ij + fouling factors = 0.0020. Permissible pressure drop through the tubes (excluding the pressure drop through the shell nozzles): Ap = 6.0 pai. Solution, Following the same procedure as in Example 1, find F, = 0.914, at a corrected h = 123 Btu/hr. sq. ft., °F. A comparison with, the shell-side coefficient shows e(W(Aty/@)]°*'? = 0.244 (Figure 2-40) and a baffle pitch ratio of Ly/D, = 0.284 (shell diameter 37 in. Ap = 10 ins Ly = 10.5 in., resulting in F, = 0.845 (Figure 2-41), With b = 3192 (Table 2-9) and C, = 1,470 (Table 2-11) we have oF, Cra 3192 (0.914) 1470 (0.845) (e) (w(anales? (0.244) = 0.578 ‘The shell-side coefficient would therefore be hy = 123/0.573 = 215 Btu/hr. sq. ft. °F at the permissible pressure drop, and in this example the heat exchanger size could be reduced by about ‘85% if the weaker medium is passed through the shell side of the exchanger. Heat Transfer Coefficient — Petroleum Fractions ‘A. simplified equation for the heat transfer coefficient of petroleum fractions in horizontal tubes is: Pee ae ve nmnaitanre's (af)* (3 where: hi, = Natural convection film coefficient, Btu/ hr. sq. ft°F k, = Thermal conductivity, Btu/br. ft.°F p’ = Specific gravity = Specific heat, Btu/lb.°F Z,, = Dynamic viscosity, cp Ap’ = Difference in, hot and cold specific aavity, ec - ¢'n Lo, Outside tube diameter, in. B Table 2-12. Average Value of Constant between 0 and 400° Using well-known values fork, p’ and ¢ of petroleum fractions, {k" (p’)* c* ] can be Considered a constant for a given degree API gravity, Table 2-12 shows the average values of these constants over the API gravity range. Selecting a 14° API oil and a 3/4-inch OD tube ‘as reference values, the above equation becomes 0.104 he = 894-7 ee) Zz" Ne" Figures 2-43 and 2-44 are nomograms based on the latter equation. Table 2-13 gives correction factors or multipliers to h, for tube diameters ther than 8/4-inch and API gravities other than 14°. For liquids other than degree API oils, multiply hg by 9.6 [B® (p')*e]*. ‘The maximum error in the range of the nomograms is 14.5%. But the significant value in this approximation is (Ap'/p')" which provides a ‘maximum error of 3.4%. Example for Horizontal Tubes. The heat losses from a kerosene storage tank are estimated at a maximum of 3,000,000 Btu/hr. Automatically controlled exhaust steam at 212°F is available for hheating. A horizontal bayonet tube bundle consisting of 7/8inch OD tubes on 1 1/8.inch square pitch, 12 fect long is to be inserted through an Opening at the tank bottom. The kerosene has & gravity of 40° API at 60°F, and its viscosity is 1.47 Ep at 10°F and 0.68 cp at 210°F. The kerosene ‘temperature is to be maintained at 10°F. Determine the required heating surface.Heat Exchangers 65 TentonTaL Tu8ee NATURAL CONVECTION FILM COEFFICIENT FOR PETROLEUM FRACTIONS (law Viscosity Ranged () 7 Tonizonrat, Tw8es WATURAL CONVECTION FILM COEFFICIENT FoR PETROLEUM FRACTIONS. (high Vacoeity Range) Yl = 40 * " : :| Figure 2-43. Natural convection film coefficient for petroleum Figure 2-44. Natural convection film coefficient for petroleum fractions in the low viscosity range—horizontal tubes. Solution, The available temperature difference = At=212~- 100= 12°F. Assume a temperature drop across the kerosene film of Atmum ~ 100°F. ‘Then the wall temperature, t = 100 + 100 = 200°F, and the mean film’ temperature, t; = 100 + 200 7 10°F. By interpolation on an ASTM viscosity chart, Z, = 0.98 cp at 150°F. From physical property data Ap! = p'yo0°r ~ 6'200%r = 0.810 - 0.771 = 0.089 and Se'fo! 0.039/0,771 = 0,0505 fractions in the high viscosity range—horizontal tubes, Using Figure 2-43 and Table 2-13: h = 48(0.962)(1.06) = 49 Assuming the inside film coefficient referred to tube OD, hy. = 1,500, the clean overall coefficient is calculated: 49 (1,500) oes: bane Checking the assumed At film: Us 415 Arum =e At Gg (112) = 108.8°F > 100°F assumed Using an overall fouling factor of 0.0015, esc Gera 0.0088 and the fouled overall coefficient, U, = 44.5.66 Equipment Design Handbook Table 2-13. Correction toh. Table 2-14. Correction to A, for Oil Gravity Muteiper % 8 : @ e ibs Possible interference with natural convection currents in tube bundles makes some safety factor in design mandatory. Using a 26% safety factor, the required heating surface is 3,000,000 445 (112) = 753 cu. ft. 183 12(0.229) = 274 tubes. A= Zo (catety factor) = wat (1.25) ‘and the required number of tubes is Vertical Tubes For vertical tubes over 2 ft. high, the flow can no longer be regarded as laminar. Using a procedure similar to that used to develop the simplified equation for horizontal tubes, the natural convection film coefficient for vertical tubes is: 141 @'dp'® aa where: = " Natural convection film coefficient, Btu/ hr. sq. ft°F fp! , = Specific gravity 4p’ = Difference in hot and cold specific gravities, P'e-P’y Z, = Dynamic viscosity, cp Bese: Note: This equation is based on 10° API oil as a reference. For other specific gravity petroleum fractions, the coefficient is corrected by a multiplier given in Table 2-14. For liquids other than degree APT oils, multiply he by 8.0 (ke) Figure 2-46 is a nomogram of the above equation of h, for vertical tubes. Example for Vertical Tubes. ‘The subcooling zone of a vertical naphtha condenser-subcooler ‘condensing and cooling 15,400 Ibs./hr. of 60° API naphtha has a calculated duty of 2,460,000 Btu/hr. cover a cooling range from 380 to 100°F. Cooling water in the tubes has a cooling range from 80 to 90°F and an average temperature of 86°F (in the subcooling zone); the water film coefficient has been calculated as fy, = 800. Determine the required clean surface in the subcooling zone. No baffles are used so that the naphtha can be considered in free convection. The exchanger is a single-pass counterflow unit. ‘The naphtha has a viscority of 0.50 cp at 100°F and 0.18 ep at 30°F. Solution, The log mean overall temperature difference in the subcooling zone is and the arithmetic mean temperature difference is At= 240 - 85 = 15°F. Assume a temperature drop across the naphtha film of Atti 7 140°F. Then, the wall temperature, ¢,, = 240 - 140 = 100°F ‘and the mean film’ temperature, ty = 2 240 + 100 _ y70°R. By interpolation, Z; = 0.35 2 cp at 170°F.verricaL Tuses NATURAL CONVECTION RUM COEFFICIENT FoR PETROLEUM FRACTIONS, i % dag’ wo ‘ a 20 20. 109 i 10 3 40 ee #2 6. © 0 100 “ ey a 200 ” 300 = 40 = $0 woo al 20 008 4 » 08 7 0 f 30 s 00 Figure 2-45. Natural convection film coefficients for Petroleum fractions—vertical tubes. ‘From physical property data 40" = p'ro0%r ~ P'240%r = 0.722 - 0.659 = 0.063 B= p'r70%e 0.690 p'dp’ = (0.690)(0.063) = 0.0435 Using Figure 2-45 and Table 2-14: h = 70 (1.80) = 91.0. ‘The clean overall coefficient is 800(91) 891 = 81.7. Heat Exchangers 67 Checking our assumed Atiam : tenia * (Ue/he) At, = (81.7/91.0) 185 = 139.1°F ~ 140°F assumed. Again using a 25% safety factor, the required clean surface is @___ 2,460,000 Ac "ULLMTD ” 81.7(101) (1.26) = 372 5q. ft, Heat Transfer Coefficient: Condensing Liquids ‘The heat transfer coefficient by condensation may be found using the following equation: 352, aa \ a (4 52 gA(Cos a)! ud, (Ate) where hh, = Heat transfer coefficient by condenss- tion, Btu/sq. tt. hr. F k = Thermal conductivity of the condensed fluid, Btu/hr. oq. ft°F 5 = Density of the condensed fluid, Ib./eu.ft. & = Acceleration of gravity, ft./hr.” A = Latent heat of condensation, Btu/lb. @ = Angle of inclination of tubes compared to the horizontal axis = Absolute viscosity of condensed fluid, Ib./ft. d, = Outside tube diameter, ft. At, = Temperature difference across the con- densate film, °F For horizontal, vertical or inclined tubes, hes he* (Fu) where h,! is the uncorrected heat transfer coefficient by condensation and F,* is the correction factor. ‘To find h,’ use the nomograph, Figure 2-46. In using the nomograph, note that the heat load, Q,, for the particular tube position must be calculated according to the appropriate equation for Q, shown in Figure 2-48. ‘At = Mean temperature difference between the two heat exchanging fluids, °F. L_ = Tube length, ft. All other terms are shown above, Table 2-15 shows valves of (Cos a)°-* for angles between 5 and 80°. To find F.™ , use Figure 2-47. Here you will need the value of h.! from Figure 2-46 plus theHEAT TRANSFER COEFFICIENT, ng , BTU/FT? HR *E ' 0 ‘00 14000 onowtal TUBES: = AT VERTICAL TUBES: 9, = 222% = Neos a)? |NCLINED TUBES:0)= Aviad git on 0 100 HEAT LORD, Q) , BTU/LA °F FT Figure 2-46, Heat transfer coefficients by condensation outside tubes. 10,000 1000Table 2-16. (cos « }°* for Angles 5-80° —eee 2 8 & | i E ee | & - i = eo Pg eae Seer od 44 HRS n \ \ Figure 2-47. Correction factor F plotted against h.' value of a constant, C, which is C = 1/(1/hy + fouling factor), where’ hy is the heat transfer coefficient of the cooling fluid corrected to the tube outside diameter, Btuhr. sq. ft. °F. Physical properties of various liquids are given in ‘Table 2-16. The data in parentheses are extrapolated values. The data are given at the boiling points t, and at 100°F above and below the boiling points. These data are needed to solve Figure 2-46. Heat Exchangers 69 Example, Find the heat transfer coefficient by condensation of methyl alcohol outside horizontal tubes. Given: k= 0.118 Btu/hr. £1.°F bh = 08 Ibs,/tt. br. + = Specific raniy of the condensed Muld d= 473 Btu. At = 60°F d, = 0.0625 ft. hg = 650 Btu/br. sq. f1.°F Fouling factor = 0.0016 Solution. Using the equation for horizontal tubes, Figure 2-46 soak 413 "easy ~ 60(0.0625) = 126 Btu/lb. °F/ft, From Figure 2-46 at Q, = 126, find h,1 = 430. C = 1h, + fouling factor) = 1/(1/650 + 0.0016) = 329. From, Figure 2-47 at h,! = 480 and C = 829, read F,“ = 128. ‘Therefore, ig = he’ (Fa ) = 480 (1.28) = 550 Btu/br. sq. °F Rating by Computer Hand rating of exchangers is unnecessary if you have at least a medium-sized digital computer. A ‘computer memory storage of approximately 1,000 words will solve simple, liquid-to-liquid exchanger rating problems. A computer with a 4,000 word memory will solve most of the usual heat exchanger rating problems. Block Diagraming ‘A major portion of solving a problem by computer is in block diagraming. A block diagram is an outline or flow chart of the problem. From this diagram and the necessary equations, a step by step procedure can be written that the computer ‘can follow (this is a program). This procedure is ‘written in the language of the computer at hand. ‘There are two main problems in programing: 1, Converting engineering data to a form that can be used by the computer 2. Including all possible situations the computer will encounter.70 Equipment Design Handbook of Various Liquids | de me = # siete | & | fete 7" ete ai cei ae | Tele #i2 e(2/ 8/8) 818 3 & © |g | | = a = a) B\e a| =| 8|@| 2 g | B | & @ | in| & | 2 | co Ba = ate eee = an Ei: a\e g 2\2\2 2 |B Gta | |B & | S| 3 & Seen =e * a a be ae eee 2 | @ € a | a | & 3 | Ea a\ 8 e 2 2/2 &| & ees | 8 8 2 8 |S tS | te ‘Dichloromethane. [38 is we ue 199 088 ron oat O88 Bete a | 8 S| |B | cae e #| 2 | B| @| @ ae | ae a | i mB | | es a8 | BS eee) ae | 78g ele) e| ele) S| ele 18 2 + 00 mm 182 8 “ oe one 0.86 oar 020 ele - are be | a ele § | ap | oe |e | ie | oe ee) el al 4 & BB] eR 8 | me |B | os | oo | | al@lel@lé = sa | 38 ge) EG) 8) &la|s| @leal Bs Pee ee ee er ban Dee | as w| el 2] e\ se e| | 8| 88 ee ee ee rr S| Be tee ae be ae | Teint a |e) #\8|s S|) @l fle) s es a a | 2 || f | | & Exeenai | 21 8) S| |e |e | OB | ae | ee ie | | Stem tae bie | ae | aa als) e| =| 8) B) ey ie] | Water i ae 19. 1 a ati 36 333 ‘ee ‘0.63 o i | | & 2 | ae | a ie | Note: The data in parentheses are extrapolated values. ‘Sample Program ‘A program which will rate condensers, vapor. izers and exchangers with two-phase flow in several different types of equipment is quite complex. The ited here will be a relatively simple ‘be modified to include these refinements. Also, it is limited to a floating-head type exchanger with two or more tube passes and & single-pass shell. ‘The nature of the programing will depend on the computer size. With small computers it may be necessary to be repetitive to conserve memory space, For example, on a small computer the Thachine tries a new shell size each time it starts With two tube passes and increases in passes until the pressure drop is met. On a large computer it wrould not be necessary to start the program with two tube passes. With extra instructions it could start with the number of tube passes used on the previous shell size. 7 TTInput ‘The input format can be seen in Figure 2-48. For computers smaller than 2,000 words this form will be slightly changed. For these smaller computers the physical properties, heat load check and corrected log mean temperature difference are calculated by hand and included in the input. Fluid Identification and Properties ‘The first computer step after loading the input is to identify the fluids on the shell and tube sides. A block diagram of the fluid identification and properties subroutine is given in Figure 2-49. After ‘identification, the properties are assembled and transferred to where they will be used in the . To save valuable memory space this same routine is used for both the shell and tube sides. A fluid is identified by a search of the input storage. For example, if there is a zero in the flow input for water, the computer knows that it is some other substance. Thus, in effect, it answers the question, “Is it water?” ‘There are three sources from which this routine can assemble properties. They are (1) through input data, (2) by calculation and (3) within the subroutine. Properties which require for calculation more memory capacity than the computer has available included in the input after they are calculated; an example of this is liquid viscosity, Properties constant over a wide temperature range are stored within the subroutine. The specific heat of water isan example. If the process stream is identified as oil, the specific heat is calculated from equation ‘1 in Figure 2-49, where C= Specific heat, Btu/lb. °F S.G. = Specific gravity t, = Average temperature, °F. This equation is based on a characterization factor of 11.8. If the oil has a significantly different factor, a multiplier must be used with equation 1, The multiplier can be obtained from the expression 0.055K + 0.35, where K is the characterization factor. It is not necessary to supply a thermal conductivity value to the input for the previously mentioned oil. By means of equation 2 (Figure 2-49), where = Viscosity, op ke = Thermal conductivity, Btu/hr.°F sq.ft./ft. Heat Exchangers 71 SHELL TUBE Inlet Temperature, cee enene Outlet Temperature, °F. are aaa Flow Liquid (excluding water) Ib.for, Water, 1b. as Gas, Ib./hr. —_ — Condensing steam, tbr. —_— Specific Heat, Btu/lb, ee Visconty (ag. ep. ee ‘Viscosity (wall), ep. —_— —_ Oil Specific Gravity @ 60°F. — Liguid Density (avg), tb.feu. ft, = Gas Molecular Weight nena Compressbility Factor —_— Operating Pressure, psla Saoney eae Allowable Pressure Drop, psi _ — Film Coefficient (if known) eee Heat Loed, Btu/br. cee Max. Overall Transfer Coefficient ‘Tube 0. in. aaeeaaees Tube BHG Tube Length, f. ‘Tube Pitch (1) 15/16, (2) 1, (3) 144, or (4) 1-4 Overall fouling Shell Design Pressure, psl Corrosion Allowance, in. HII Figure 2-48. For computers having 8 capacity of less than 2000 words, this input form will be changed slightly ‘The function K (cu/k)*"* for heat transfer can be calculated. Tt is necessary to supply only the viscosity. ‘This relationship is based on thermal conductivity data in the TEMA standards. The coefficient of 0.15 used in the equation is an average value, The lighter hydrocarbons through hexane reflect a coefficient of 0.148. From hexane through 50° APT oll, the coefficient is 0.152. The coofficient decreases with gravity to a low of 0.145 for 37° API and heavier oils. This use of an average coefficient is justified when the large variation of thermal conductivities in the literature is considered. ‘After the properties are assembled they are transferred to the position where they will be used. For instance, the specific heat is sent to the heat, load cheek and also the heat transfer coeticient routine,72 Equipment Design Handbook \ [ds Cond. Seam) [send & 1 = 890 To 0 Check| [ene Temp. Renae) rant owt 108 08 c= 0.681 - 0.3088.G. + t, (0.000816 - 0.000306 5.G.) (1) ke (culk)!9= 0.15 42 (2) MW. Po Gos o "F907 Z (460 +t) @) (Shell) At, = ty. - (Tube) At, = fer - @ Q=WCAtor Q= WAH 6) {SSSI ee Figure 2-49. This program will check the input heat load. Other equations solved in the Figure 2-49 subroutine are equation 3 for gas density where: e M, Density, Ibs./cu. ft. Molecular Weight Operating pressure, psi Compressibility factor, Average temperature, °F Po Zz ta and equation 4 for the shell and tube temperature differences where: = shell inlet, outlet temperatures, °F channel inlet, outlet temperatures, °F teas tea tersten and equation 5 for the heat duty, @, in Btu/hr. where: W__ = flow rate, Ibs./hr. AH = enthalpy change, Btu/lb. ‘This program is based on the premise that the heat load is supplied by an external source. ‘The machine will rough check the input heat, load rather than supply it. Only a slight modification in the program is necessary for the machine to ‘calculate and supply its own heat load. After the heat load is calculated by means of equation 5 it is checked against the input heat load. If the differ- fence is more than 10%, the calculated heat load is printed out. The flow’ or temperature range can then be changed in the input accordingly. With a large computer this change in input could be programmed to be done automatically. ‘When the heat load for condensing steam is checked, an average latent heat of 890 Btu/lb. is tused. This value will check within 10% of the heat load for steam pressures ranging from atmospheric ‘0 800 psi. Corrected Mean Temperature Difference ‘The simplest way of obtaining a mean temperature difference (MTD) correction by hand is to use the TEMA curves. The machine procedure, as can be seen in Figure 2-50, is not so easy. Unless you ‘ean spare the memory space, the advisability of putting it on the machine is debatable, Fortunately, we do not need six different equations for one+to-six shell pass correction factors. We can use one equation if we modify the X variable for each case. Usually, we can calculate X for a desired number of shells in series by using ‘equation 12 (Figure 2-50). This X is inserted in equation 14 to obtain F. ‘The terms for equations 6 and 7 (Figure 2-50) are defined above.[oats Terminal A” @) & [ene -ientDe aioe Figure 2-60. The program for corrected mean temperature difference should not be used if you are limited in memory space. Heat Exchangers. 73 © a MTD = iy (GID/LTD) 8) R= At,/At, (9) Ate * Tha tal 0) When R= 1 qa) 2) yh) Bayi R+ VR \ aay In a 1-R- VRPea Xa-2n Fran VRe+1 1-Xn- an ma? (ioe =) Bote VRP \ sy in { Seam yok: Weed MTD, = MTD (F) In equation 8 (Figure 2-50), MTD is the mean temperature difference, °F; GTD is the greater temperature difference, °F, and LTD is the lesser temperature difference, °F. In equation 9 (Figure 2-50), R = parameter ‘or MTD correction. In equation 10 (Figure 2-50), X = variable for MTD correction,74 Equipment Design Handbook In equation 12 (Figure 2-50), 7 = the number of shell passes. In equations 13, 14, 15 (Figure 2-50), F = MTD correction factor. In equation 16 (Figure 2-50), MTD, = corrected MID. ‘As can be seen in Figure 2-50, the first time the correction factor F is equal to or greater than 0.8 ‘an exit is made end the corrected MTD is ulated. From the X,. 9, that is used in computing tthe correction factor it is known how many shells in series are to be used. For example, if Xea is ‘used, then four shells in series with a total of eight or more tube passes are required. ‘The complexity of this routine partially stems from a built-in prevention of division by zero. If a division by zero occurs the machine will overflow. For example, in equation 14 a division by zero will occur when (2/X)-1-R = (R? +1)" or R= 1. In tthe first case this eannot occur for a one-shell MTD correction factor greater than 0.8. Therefore, when the machine discovers this case it develops an X parameter for two shell passes and (2/X)-1-R will Ro longer equal (R? + 1)* . When R= 1, Xy.2n 8 calculated by equation 11 and used in equation 13. Selection of Size and Arrangement ‘The function of the routine shown in Figure 2.51 is to keep track of the exchanger arrangement and the shell size. ‘The preliminary surface is calculated by means of equation 16 where: A = exchange surface, sq. ft. Z = heat duty, Beu/hr. Umae = maximum overall heat transfer coefficient, Btu/hr. sq.ft.F ‘The first surface selected will be either the same or smaller than the final selection to eliminate the possibility of making a selection that satisfies all Conditions but is oversurfaced, a possibility if we started with a larger surface than the final selection. Starting with a smaller curface is accom- lished by specifying in the input # Uma, which is the maximum acceptable overall heat transfer ‘coefficient. If in doubt about what to specify, use 2 U that is abnormally high or use the reciprocal of ‘the total fouling factor. Next, the surface per exchanger is calculated from equation 17, where: Ne Neo = number of exchangers in series = number of exchangers in parallel (Go Up 1 Sha Se A “MTD. Una (16) Sq. Ft/Exch = A/Nes Nep an N, = A[(Sq. In./Ft.) (L) (18) (P) (rs) Pee ge as) ae ( Pipe size + thicknes ) a ‘TEMA shell LD, Bundle dia., (20) Figure 2-51. The function of this routine is to keep track of the exchanger arrangement and the shell size, Points A, 8, ‘and C also refer to Figure 252. ThIn the first trial, one bank of exchangers is used. Nyy (number of exchangers in parallel) will be one. Ne, (umber in series) will be as determined from the MTD program. ‘Then, the number of tubes is caleulated by equation 18, where N, = number of tubes L' = tube length, ft. ‘The machine will search the standard two-tube pass counts until a shell size with the same or more ‘tubes is found. ‘Some companies have the equipment to handle larger exchangers. Thus, we must have some way of limiting the shell size. This is done by storing in the memory the two-tube pass count for the largest acceptable shell size. Then, whenever the shell size is changed, the machine tests to make sure that the number of tubes is less than this maximum count (Ni max). If the tube count is greater than N; mux» @ parallel bank of exchangers is added, thus reducing the shell size. ‘Exchanger shells that are smaller than 25 in. ID. ‘are manufactured from pipe. The tube counts used in the computer are based on the shell thickness shown in TEMA. If the shell thickness calculated by equation 19 exceeds the standard, then the tube circle is reduced accordingly. In equation 19, T, = Shell thickness, in. P= Baffle pitch, in. 7, = Shell inside radius, in. When the tube circle is reduced, the new tube count is obtained by equation 20. For shells larger than pipe size, keep a standard tube circle and shell ID for all presrures and Iet the OD fall where it will. ‘Tube Side Pressure Drop ‘The program for the tubeside pressure drop is illustrated in Figure 2-52. As can be seen, it involves a great deel of logic. Equation 21 calculates the flow rate W in bs.fhr. where: W, = flow rate in one tube, Ibs./hr. N, = number of tube passes N; = number of tubes Nep = number of exchangers in parallel Equation 22 calculates the velocity, V, in ft,/sec., where: d, = tube ID, in. p = density, Ibs,/eu.tt, H it Exchangers 75 Each time the tube count or number of tube passes is changed, the velocity is checked to make sure it is not too high. The maximum allowable velocity is calculated from equation 23. If the fluid is clean, the value to be used in the numerator will depend’ on the tube material used. A value of approximately 8,600 can be wed for admiralty ‘ut ‘The Reynolds number is calculated by equation 24 to see if the flow is turbulent or viscous. In ‘equation 24, 1 is the viscosity in centipoise. Equation’ 25 calculates the effective tube length, L', where L is the tube length, ft.; N,, is the number of exchangers in series. This equation takes into consideration pressure drops on the tube side other than that inside the tube. If the Reynolds number indicates viscous flow, equation 27 (where = viscosity at the tube wall, cp) is used. If the flow is turbulent, equation 26 is used. Both of these equations are based on the Fanning equation. The friction factor incorporated in the viscous equation is simply f = 16/Re. The friction factor used for turbulent flow is f = 0,048/Re°? . A safety factor of 20% is used to take care of fouling and the effect of possible overtolerance of the tube wall. ‘This program is set up on the basis that 12-tube passes are the maximum. If the calculated pressure drop is still under the allowable with 12-tube ‘passes, the computer proceeds to the tube side heat, transfer coefficient section. ‘Tube Side Film Coefficient For purposes of heat transfer we will consider the turbulent zone to be for a Reynolds number greater than 10,000. In this zone, equation 28, (Figure 2-58), is used where: Iyo = inside film coefficient Btu/hr. sq.ft. °F Re = Reynolds number d, = tube OD, in, K = characterization factor C= specific heat, Btu/lb. °F H = viscosity, ep Hy = viscosity at the tube wall, ep. When the Reynolds number is between 2,100 and 10,000 it is called the transitional zone. In this zone one equation is not sufficient. Here we must resort to breaking the heat transfer data into intervals to fit the TEMA curves as closely as possible. ‘Equation 29 is used when the Reynolds number is less than 2,100. Each rating that is computed in this zone should be examined carefully. This ‘equation is in some instances conservative, because76 Equipment Design Handbook | cue. Wan) ate. v.22) Cale. Vmax 23) 2 Go Toh it doesn’t include the effects of natural circulation and turbulence at the tube ends. ‘Simplified equations are used for water and gas. ‘These have the advantage of using relatively few physical properties, which saves time on compiling input data. Equation 80 is used to calculate the inside film contiint for water where t,“ average tempers W=W, Ny INeNep ca) V= 051 W/d,?p (22) Van, = V (To Be Inserted) ve (23) A Re=6.81 Wid (24) Li = (L+ 5.85 d)) Ny New (25) wh 5! yo? AP ub, = watson (26) 4.08 Wub! AP as = 10%dj*p (1.1 Hl) (27) Figure 2-62. A relatively large amount of logic is required for calculating the tube-side pressure drop. Equation 81 is used to calculate the inside film coefficient for gas. Shell Side Pressure Drop First the computer assembles the information necessary to compute the croseflow velocity and the baffle window velocity. From storage it brings the net free distance (NFD) for crossflow. The netaim Geie, Transition be ‘Dranstion Zone i Curve Fitted Figure 2-53. The tube-side film coefficient depends on the type of flow within the tubes. free area (NFA) for baffle window flow is calculated from equation 32 (Figure 2-64) where: 7, shell inside radius, in. NY, = number of tubes tube OD, in. &. 1 Equation 33 is used to calculate the shell longitudinal mass velocity, G,, in Ibs./sq. ft. sec., where W, = shell flow rate, Ibs./hr. Equation 84 is used’ to calculate the shell exossflow mass velocity, G., in Ibs./sq. ft. sec., where P = baffle piteh, in. ‘Equation 86 calculates the low flow velocity, Vi, in ft,/sec., where p, = shellside fluid density, Tbs.feu. #. Equation 36 calculates the maximum shell velocity, V mass Equation 37 calculates an MTD correction factor, R,, to compare it with the tube pitch shown on Figure 7-54, where 4, = shell-side fluid viscosity, cp. Equation 38 calculates the tube pitch, |, in. Equation 39 calculates the turbulent, shell crossflow pressure drop, AP., psi, where N= number of rows in crossflow and B= number of baffles. Equation 40 calculates the viscous, shell crossflow pressure drop, AP, , psi. Equation 41 calculates the shell longitudinal pressure drop, AP;,, psi. Heat Exchangers 77 ‘Turbulent Zone 0.486 Re* (Cu 8 fw \ or ue =e, = (28) Viscous Zone 13.1Re™ (a,\*®_ (cu\ 4 (_w\ 028 ees () a ¢ He (29) ‘Transition Zone Is Curve Fitted Water hie Ae oy verge (30) Gas (31) Equation 42 calculates the total pressure drop in the shell, psi. ‘The program is designed to discover special cases promptly. The special cases stem from: 1. Low operating pressure on a vapor 2. Abnormally high flow 3, Abnormally low flow ‘The machine starts with maximum baffle pitch and maximum percent cut, The maximum allowable baffle pitch is that specified in TEMA and will vary with tube OD. For most designs, the 7) :xi- mum cut is approximately 45%. The machine siut.5, with this baffling arrangement and calculates the pressure drop. If the pressure drop is excessive the machine will go into the divided flow portion of the routine. For the same pitch and cut, divided flow will reduce the shell-side pressure drop by a factor of approximately eight. ‘Thus, the computer immediately tests for the first and second cases just discussed. In testing case 3, the machine starts reducing the baffle cut in 5% increments, calculating the pressure drop each time. When the baffle cut is reduced to 10%, the machine will add baffles one at a time. Baffles are added until minimum baffle pitch is encountered (1/5 of the shell ID down to 2 inches). When minimum baffle pitch is reached, the78 Equipment Design Handbook aaa TE eR DLB For = 028 do? "hig > 6.98 For = 0.33 dof Rc NPA = 1, - 0.785 Ny do? (32) G,= 0.04 W,/(NFA) (%Cut) (83) G, = 0.04 W,/ (NFD) (P) 4) Vi= Giles (38) Vmax” 40/ Pr (36) By = dy Gx /ity @7) I= Tube Pitch - dy (88) ‘Turbulent AP, = o.ooo12s N@+i) — (89) Viscous AP, = o.0000s22 42 1 (+1) (40) low AP, = G,?B/5950 p, a1) AP, = AP, +P, (42) Figure 2-54 (left). This program for shell-side pressure drop is designed to discov ial cases promptiy. Point D refers to a supplemental program shown in| Figure 2-85. low flow case is discovered without trying every possible combination of pitch and cut. Without modification, this procedure normally gives a low percent cut and wide baffle pitch. This type of baffling will also give a relatively low heat transfer coefficient for a given pressure drop. For small computers, where memory space is at a premium, it is necessary to approximate the region of pitch and cut which will give optimum hheat transfer for a given pressure drop. This is done by limiting the ratio of G,/G, to less than three. In large computers, where memory space is not a problem, it is possible to choose a pitch and cut, combination which gives the highest heat transfer. To amive at the optimum heat transfer coefficient, use is made of the fact that the shell film coefficient is directly proportional to the product of G, and G,. There is one combination of pitchHeat Exchangers 79 ‘Siore DF In Shall Paw or Bate Store 29-71, But Ure Figure 2-55. If the pressure drop in the shell is ex- ‘cessive, the machine will go into the divided flow portion of the routine shown and cut which will give « higher product of G, and G, than any other. After the computer finds all the baffling combinations that meet pressure drop, the one which gives the highest velocity product is selected for use. Tt is necessary to place a restriction on the shell-side velocity to eliminate velocities that would set up damaging vibration. This is more of a problem in U-tube exchangers than it is in floating heat and fixed tubesheet types. The velocity around the baffle is the most critical and usually has a higher value than the cross velocity. The velocity around the baffle is limited by equation 36. if the shell pressure drop is excessive, the program will use the divided flow portion of the Toutine shown in Figure 2-55. Shell Film Coefficient Figure 2-56 is the block diagram for the subroutine of the shell film coefficient. Equation 43 is used to determine the geometric mean velocity, Gm Ib./eq, ft. sec. where G, and G, are found by equations 33 and 34, Equation 44 is ‘used to find the shell film coetficient for oll, h, Btu/hr./sq. ft. °F, where: Gm = Geometric mean velocity, Ibs./aq. ft. sec. d,’ = Tube OD, i w= Shell fluid viscosity, ep K = Characterization factor He = Viscosity at the wall, cp VGIG) cu [ y\ox « (2) (4) as) (1+ 0.0067 t.) d,0% poe Ont (48) 84 G28 “4,04 ge Water h = 247 pa (45) Gash = 29 4s) Figure 2-56. The subroutine of the shell film coaf- ficient. When the shell fluid is water, equation 45 is used to find the shell film coefficient, h, Btu/hr. sq. ft. °F, where: i, average temperature, °F p= Density, Ibs./cu., ft. When the shell fluid is a gas, equation 46 is used to find the shell film coefficient, h, where C = specific heat, Btu/lb., °F.80 Equipment Design Handbook Calculation of U and Surface This subroutine is illustrated in Figure 2-57. The overall U is calculated from the sum of the heat transfer resistances shown in equation 47 (where hig is the outside film coefficient, Btu/hr. sq. ft. °F; ‘hyo is the inside film coefficient, Btu/hr. aq. ft. °F; and r, is the fouling factor). The calculated surface equation 48 (where @ is the heat duty, Btu/hr.; MTD, is, the corrected mean temperature dif. ference, °F) is compared with the actual surface equation 49, where: N,, = number of exchangers in series; number of exchangers in parallel number of tubes and tube length, ft. Specification Sheet When specifying a heat exchanger for process use, an excellent check list is the TEMA Specifica- tion Sheet shown as Figure 2-58. Air Cooled Heat Exchangers Economics is the deciding factor on whether to use air cooled or water cooled exchangers. Three sets of curves, Figure 2-59, will allow you to decide which is more economical ‘The following equations must be solved before using Figure 2.59. ‘The ratio for 316 stainless steel is Lu c 88 [Q(LMTD,. ) Uy 1°-72 + 1.72 (107° 88 [Q(LMTD,) U,]°" +(1.27 Q/LMTI where: C, = Annual cost for cooling water Cy ‘Annual cost for air cooling @ = Heat transferred from one media to another, Btu/hr. LMTD = Logarithmetic mean temperature difference, °F (where subscript w represents water and a represents air) Overall heat transfer coefficient, Biu/hr. sq. ft., °F (where subscript w represents water and a represents air) For carbon steel, the ratio is 40 [Q/LMTD,, ) Uy J°-* + 1.72 (107 *Q 52 [QALMTD,) U,]°? +1.27 Q(LMTD,) Uy Cais, Over AN UTD Gate. Rea. Ave 48) ale. Actual Area (49) etanlie A To Output Ur Ake + Ulig +e «7 Q Ace MTD, 7 (48) Actual A = (Ney Ney Ne L) (Sq. In./Ft.) (49) Figure 2-57. The overall U is calculated from the sum of the heat transfer resistances. Point A refers to Figure 2-51 Figure 2-59 can be used to analyze any process to be cooled or condensed. The limitations of the curves are (1) process pressure must be less than 150 psig; (2) process temperature must be less than 300°F; (3) cooling water temperature rise is from ‘85°F to a final temperature of 105°F; (4) dry bulb air temperature is 100°F; and (6) air face velocity is 675 fect per minute. When process variables fall outside these limits, separate calculations should be made. ‘The first design step is the determination of the process design conditions. For a particular service the design flow, concentration, inlet temperature and condition, and outlet temperature and condition should be fixed. Heat Duty ‘The heat to be removed from the process is calculated from the design conditions by standard heat balance equations. | iHeat Exchangers 81 assumption cen be improved if heat transfer coefficients for similar services are available, After the bundle size is assumed, the face area and the inside surface area are calculated from values tabulated in Table 2-17. The total air flow through the bundle is calculated from the face area and the design face velocity. The average air temperature change required to remove the desired heat load is calculated using the equation for Q,. Since the design inlet temperature is 100°F, the alr temperatures for the assumed bundles are set. The LMTD is calculated for true countercurrent flow by (Ty = te) ~ (Ta = th) Figure 2-58. Heat exchanger specification sheet. Q, = WC, (Ti T2) @. = Ww(R,-H) Giese Ha) where: a = Sensible heat transfer, Btu/hr. @ = Condensation or latent heat transfer, Btu/hr. Q& = Total heat transfer, Btu/hr. W = Flow mate, Ibs hr. Cy = Specific heat, Btu/lb. °F 7.7, = Inlet and outlet process temperatures, °F H, = Vapor enthalpy, Btu/lb. H, __ = Liquid enthalpy, Btu/lb. H,,H, = Entering and’ terminal enthalpy, Btu/lb. Bundle Selection ‘The set of exchanger bundles required is determined by a trial and error procedure. A set of ‘bundles must be assumed to determine the heat transfer coefficient. The accuracy of the initial MTD = In (Ty - t2)MTa - 4) where: IMTD = Logarithmic mean temperature difference, °F T,,T_ = Inlet and outlet process temperatures, oF ty,t2 = Inlet and outlet air temperatures, °F Since the air flow is mixed and the process ‘unmixed, an LMTD correction must be applied. ‘An overall heat transfer coefficient, U, for the assumed number of bundles is determined by substitution of the calculated values into the equation U= Q,/A(LMTD) where A = surface area for heat transfer, sq. ft. ‘The U value obtained is compared to the expected or historical U value. If the U values do not match, another set of bundles is assumed and the above’ procedure is repeated until the best match possible with the bundles available is obtained. Heat Transfer Coefficient Inside film coefficients based on the number of bundles calculated are determined from ke (ad? gn \ 18 Iie ort a & (aa\* (cyw\ o* 9.0388 | ( : ) 2 Tue = Inside film coefficient for condenss- tion, Btu/hr. sq, ft. °F hy = Inside film coefficient for subcooling, Btu/hr. sq. ft.°F k = Thermal conductivity, Btu/hr. sq. it, °F d, = Inside tube diameter, ft. p = Density, Ib./cu. ft. e Acceleration of gravity, 4.18(10*) ftfbr®82 Equipment Design Handbook 14 ~ eas basis TERMINAL TEMP, TERMINAL TEMP, yee Tet ore ish L 210% ng 100) COST RATIO, Cy/C, o7| 24108, rye100 os Basis: ‘TERMINAL TEMP, Teese favo m4 200 an, os ar) 30 Figure 2-69, Air cooling is more economical than water cooling when the Table 2-17. Typical Air Cooler Dimensions EFECTYE, SURFACE aye n _ = Number of tubes in parallel L Length of heat transfer surface, ft. = Viscosity, Ib./hr. sq. ft. w Flow rate, Ibs. hr. G Mass velocity, Ibs./hr. = Specific heat, Btu/lb. °F ‘Note: Do not use h,, if the Reynolds number is less ‘than 2,100. ft, a INITIAL PROCESS TEMPERATURE. T “oes ~~ «OO ratio shown here is greater than one. ‘The outside film coefficient based on the inside surface area for the finned tube bundles described is 180 Btu/hr. sq, ft. °F at a face velocity of 675 ft/min. An outside fouling factor of 0.001 is generally used, The inside fouling factor depends ae ‘The overall theoretical heat transfer coefficient is calculated from the film coefficients and other resistances by the equation ae ie U, = Overall air heat transfer coefficient, Btu/hr. sq. ft. °F hy = Process film coefficient for heat transfer, Btu/hr. sq. ft. °F hg = Air film coefficient for heat transfer, Btu/hr. sq. ft. °F= Tube wall thickness, in. = Thermal conductivity, Btu/hr. ft. °F = Fouling factor, hr. 9q. ft. °F/Btu = Inside film coefficient for heat transfer, Btu/hr. sq. ft. °F fy = Inside fouling factor, hr. sq. ft. °F/Btu If the service is all condensing or all subcooling, the theoretical U value is compared with the U value required for the number of bundles used. If the U values are different, the bundle calculations are repeated with the theoretical U value. The theoretical U value calculations must be rechecked if the change in U value causes a change in the number of bundles required. This procedure is reiterated until the U values match. When a service calls for condensation with subcooling, the calculations become more compli- eated. The U values for the condensing and subcooling sections of the bundles required are calculated by the equations for 1/U,, hic and hy. ‘The equation for the heat transferred by condensation, Q., is obtained from the equation for U above: pas @ = UsAc(LMTD). (corrected) where: U. = Overall heat transfer coefficient for condensation A. = Surface area for heat transfer by condensation (LMTD)e (eorreetea) "Corrected LMTD by condensation @ = UA, LMTD 4 corectea) for the condensing and "subcooling sections, tespectively. Each equation contains two unknowns. An analysis of the unknowns 4., A,, LMTD, and LMTD, reveals that they are related by the equations for heat removal by sir. If the equation for the condensing section is divided by the equation for the subcooling section, the following ‘equation is obtained: ()- ( a (am eons A, a ) U.)} \EMTD. (eons Upon substitution of the predetermined values of (Q./Q,)(U,/U. ) into the equation, the following equation is obtained: Ac. _EMTD, (cn A, LMTD, (cor) ‘The latter equation can readily be solved by a trial ‘and error method. ‘A ratio of the areas required for condensing and subcooling is first assumed. Since the inside surface Heat Exchangers 83. area is directly related to the face area, the face areas of the condensing and subcooling sections of the bundles assumed can be determined. The air flows are calculated from the face areas and the design face velocity. The equation for Q, is used as before to calculate the temperature difference for ‘he air flow through the condensing and subcooling sections. The LMTD’s for the sections are calculated from the established temperatures and the correction factors are determined and applied. The comected LMTD's are substituted into’ the equation for A,/A, to check the validity of the assumed area ratio. The trial and error procedure is, continued until the equation for 4,/A, is satisfied. ‘The LMTD’s determined by the solution of the equation for A./A, were based on the bundles assumed. To check this assumption, the areas required for condensing and subcooling are determined the appropriate forms of A = QULMTD)U. The total of the ateas required for condensing and subcooling should equal the total inside area of the bundles assumed. If the areas do not match, additional calculations determine the apparent overall U value to be used in the next gaumption. Assuming the areas do not match, « balance between the changes necessary in the A, and A, against the resulting changes in LMTD, and IMTD, is necessary. Once this balance has been made, ‘the new values should be tested in the ‘equation for A./A,. ‘Using the newly calculated total ares, calculate the average air outlet temperature and determine the overall corrected LMTD. The overall apparent U value is calculated and is used to check the number of bundles required. If the bundle selection is changed, the complete procedure must be repeated. It is necessary to apply engineering judgment when analyzing the product of this design procedure. When the process requirements do not conform to any combination of standard sizes, it is necessary to decide between an increase and a decrease in air cooling requirements. Example. A vapor stream from a distillation tower is to be condensed and subcooled. Design rate is 52,000 pounds per hour of 99 weight percent acetic acid and 1 weight percent water. Design temperature of the saturated inlet vapor is 109°C. It is to be subcooled by 26°C. The material of construction required is 316 stainless steel. Process data: (H, - Hy) at 109°C = 178 Beu/lb, , at 96°C = 0.66 Btuy/lb.84 Equipment Design Handbook Heat duties: Correction factor, F Q = WC, (T:- Te) ees (4) (62,000 tb,/hr.)(0.56 Btujib., °F46.8°F) | R= 1 RP. ae al 1,368,000 Btu/hr. tp (BPa1Rt (n+ 1)" ' = eH) 2/P-1- R- (R?+ 1)" a = (62,000 Ib./hr.)(178 Bty/lb.) = 9,256,000 Btu/hr. EMTD(cor.) = (84.1) (0.97) = 82.2°F @, = 10,619,000 Btu/lb. & 10,619,000 u = : ie LMT! é Determine number of bundles. From historical (AVLUTD or) 98362 2) data, U, = 90 to 100 Btu/hr./sq. ft. F = 76.1 Btu/hr, sq. ft. °F First assumption: Two 8 x 24 ft, bundles Since the required U, is lower than the historical Fane nN aed wa 900 sa. Fe values, another set of bundles is assumed J Air data: Second assumption: ‘Density at 70°F and 14.7 psia = 0.0750 Ib/eu. ft Bundles Face Area, sq.ft. Inside Surface _| C, = 0.24 Btu/lb., °F Area, sq. ft. Air Flow= (675 ft-/min,) (60 min, br.) (362 One- 8x24. it 849 8q. ft.) (0.0750 Ib./cu. ft.) ‘One- 4.x 24 ft. nO 378 | = 1,099,600 lb fr. : oo cat ee oe See the equation in the box at the bottom of the eee See nste 000 page. We, ~ (,099,600)(0.24) ump = (228.2 168.0) -(181.4- 100) = 40.2°F In (70.2/81.4) i T, = 228.2°F melas % 2 isia’r LMTD correction = Bee 0 Ro = 58.0/46.8=1.24 = t 7 140.2°F Pi oo sce (Ty ~ tz) ~ (Pp-) F = 093 LMTD = 2 In (Ty - Ta - 4) MTD coor. = (15.1) (0.98) = 10.4 °F : = 880-814 gy zoe y,~ 10:819,000 Btu/hr. 1m (88.0/81.4) oat aq, f(70.4°F) LUTD comer ‘This U, is too high. The first assumption seems Gare to contain the most economical set of bundles, Z i because two 4.foot bundles are more expensive ~| than one 8-foot bundle. However, since two 4 x 24-foot bundles are available from idle equipment, P it is economical to install two 4x 24-foot bundles _ with one 8 x 24-foot bundle. 110,619,000 Btu/hr. - = 58.0°F (t2 ~ ty bev. > (40,500 ft /hr )(261 69. ft.)(0.075 Ib./sq, ft.)(0.24 Btu/lb.°F) 58‘Third assumption: Bundles Face Area, aq. ft. Inside Surface Area, sq. ft. One - 8 x 24 ft. 181 349 Two- 4x 24 ft, 140 1568 821 1605 See the equation in Box A at the bottom of the age. (228.2 - 145.4) - (181.4 ~- utp = § 5.4) - (Li 100) Heat Exchangers 85 Heat Transfer Coefficient Calculations 1, Inside Film Coefficients ‘a, Condensing section (See the equation in Box B at the bottom of the page.) 'b. Subcooling section (See the equation in Box C at the bottom of the page.) c. Fouling factors (1) Inside fouling factor. .0001 hr. sq. ft. °F/Bta (2) Outside fouling factor, fo = 0.001 hr. sq. ft. °F/Btu d. Metal Resistance eee (B16 88, Lin, OD for 14 BWG tubes) = 82.1°F ne, = #4, (0.083)(0.834) Tem dmean — (12)(0.918) LMTD correction: coe = 0.0068 har. sq. ft. °F/Btu P= 0.365, Fo = 095 where: Tmi = Inside metal resistance to heat LMTD (cor. ~ (82.1) (0.95) = 78.0°F transfer x ‘Tube wall thickness, in. ‘This U, is close to the expected range. The set a ‘Tube ID, in. of bundles in the third assumption gives the best ky = Metal ‘thermal conductivity, check possible. The theoretical U, calculation is Btu/hr. ft. °F based on this set of bundles. mean = Mean tube diameter, in. BOX A. 10,619,000 Btu/hr. t- tiers. = Tay 600 RJbn)GDL ca BNGOTE Tee NODE EI = 48.4 °F (a fee, 0,600 fe for (821 sq. ft.)(0.075 Ib./ou. #t.)(0.24 Btu/lb. °F) BOX B. Ihe = 0.76 (0.060 Btu/hr. ft. °F) (0.0695 £t.)* (60.5 Ib./eu. ft.)?(4.18 x 10° ft./hr.2)(2)(314 tubesy(24 ft.) ] 2 0.0695 ft. (62,000 Ib fir.)(1.19 Ib. fi. ft.) = 833 Btu/hr, sq, ft. °F BOX C. 0.060 (0.0695)(52,000) ru = oonas [298] [ 0.0695 = 46.5 Btu/hr./sq. ft. °F -# 7(0.56)(1.16) ] °4 (1,19)(314)(0.00379/2), 0.06086 Equipment Design Handbook 2. Heat transfer coefficients 56.5 : a, Condensing i IMTD, == Tosa ~ 92°F , us | © Chie] * tgs + (tm a) * fet fo P = 10 ? where hy is the inside film coefficient for ait yp _ (228.2 - 119.4) - (181.4 - 100) and a reasonable value = 180. . = ——gosayeiay OC co 1 = 94.5°F i © * (4/188) + (1/180) + 1,0068 + 0.001 + 0.001 = = 89.0 Btu/hr. sq. ft. °F F = 099 ’b. Subcooling ae =< = 2.998 42.564 : Ay 4 u= (fh) + hoi) ttm * f+ fo = 2,883 #2.468 1 i sian see ante epee eaeeeeeao SE 4 (9/465 + (1/180) + 0.00068 + 0.001 + 0,001 Second assumptic = 33.6 Btu/hr. ft.? °F 71.0% of total area is used for condensation. 29.0% of total area is used for subcooling. aoe ee ee Air flow, = (0.71¥(975,000 Ib./hr.) J : (2ssaue 38.6 ( A, \1,868,000/ (89.0 phen eiabes i = 2.564 2UTucory ir flow, = 8,000 Ib. fr. - oo Air flow, = (0.29)(97 fr.) First ascumption: : = 282,000 lb/hr. 10% of total area is used for condensation. 80% of total area is used for subcooling. (= tein =,85:7°F i Airflow, = (0.70)(40,500 ft,/ht.) (821 sq, ft.)(0.075 Ib,feu. ft.) ‘= 682,500 Ib,fhr. (0.30)(975,000 Ib. /hr.) = 292,500 Ib fhe, (a> tetony = 20.1°F | Atrftow, LMTD, = 97.7°F and F=1.0 LMTD, = 94.1°F and F=0.99 9,256,000 Btu/hr. i (94.1)(0.99) 91.7 (>t Je com "(969,500 Ib fbr O24 Biull "F) gay coe a, 7 280 1,863,000 . (a-t sow * Gap,s00y0.24) ~ *°4°F = 2.448 = 2.446 2.448 & 2.564 [Area required with above ratio and LMTD's: . 9,256,000 Btu/hr. 189.0 Btu/hr.jsq. ft.°F(97.7 °F) = 1064 sq, ft. Ae 1,863,004, AL > Geayoan 7 A8bea te Ay = 1,496 5g. ft. ‘The total area calculated, 1,495 square fect, is 110 oq. ft less than the total area of the set’ of bundles assumed. This is acceptable. Additional subcooling will be expected. To obtain the true overall area required, the LMTD, and the LMTD, must be balanced against ‘A; and A,. The results obtained from this balance are as follows: A = 1818 aq. ft. IMTD aicor j= 16.8 °F U, = 91.1 Btu/hr/sq.ft., °F Actual data on the designed fin-fan condenser Reflux rate = 38,500 Ib, /hr. Distillate rate = 13,470 1b.for. Total overhead rate = ‘51,970 Ib fir. Top tower temp. = 108°C Condenser outlet temp. = 62.5°C Composition 99.0 Wt. % acetic acid 1.0 We. % water Air data Flow Rate, ft./min, Outlet Temp., °C No. 1 fan 164 No. 2 fan 807 60.2 46.2 Air Inlet Temperature = 26.0°C Absolute Pressure = 680 mm Hg/Abs. Heat loads, process side: Q. = (51.970 Ib./r.)(178 Btuflb) = 9,261,000 Btu/hr. it Exchangers 87 = (61,970 lb./hr.)(0.85 Btu/lb., °F) (46.5 °C)(1.8 °F/°C) = 2,392,000 Btu/hr. @ = 11,643,000 Btu/hr. Air side face velocity correction for No. 1 fan at 70°F and 14.7 psi = (764) 530°R \ /680mm ts) ) (Feo mm Hg)” 672 ¥/min. for No, 2 Fan at 70°F and 14.7 peig = (807) 530°R_\ /630 mm Hg (sara) rooms)” 70 Qno, 1 fon = (672 ft,/min.)(60 min,/hr.)(321 sq. ft.) (1/2)(0.075 Ib./sq.ft.)(0.24 Btu/lb., °F) (84.2°C)(1.8°F/°C) = 6,990,000 Btu/hr, 2 fon ~ (710 ft,/min,)(60 min,hr.)(321 2q. ft.) (2/2)(0.075 Ib./sq. ft.)(0.24 Btu/Ib., °F) (20.2 °C)(1.8 °F/°C)= 4,475,000 Btu/hr. Quotat = 11,465,000 Btu/hr. This is a good check. Heat transfer coefficient calculations: __(228.2- 126.7) - (144.5 - 78.8) LMTD, in (101.5/65.7) LMTD correction: Ro = 47,9°F/83.7°F = 0.572; 88.7/149.4 = 0.560; F = 093 MTD aqeor) = (82.8) (0.98) = 76.5°F 11,643,000 Btu/hr. (76.5°F (1605 sq. f.) ls Btu/hr. sq. ft,, °F This is a very close check with the U,, 91.1, calculated for design. The exchanger performed adequately.

Heat Exchanger Design Handbook PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Heat Exchanger Design Handbook PDF

Uploaded by

Copyright:

Available Formats

You might also like