Lesson’s Learnt For Heat Exchangers:

Note: Smallest Shell ID with Highest Performance is the most preferable Design.

1). The Length to diameter ratio (L/D) for exchanger should not be less than 5.

2). When we reduce the shell diameter , it results in the decrease of number of tubes for that ID of shell,
which interns results in increase in pressure drop on the tube side.

3). Baffle spacing is a very important parameter in exchanger design. The pressure drop on the shell side
can be adjusted by increasing or decreasing the baffle spacing.

4).For double segmental baffles, in case of horizontal cut and shell side nozzles being on top and bottom
the end baffles should be of outer type and the cross passes should be even.

5). For double segmental baffles, in case of vertical cut and shell side nozzles being on top and bottom the
end baffles should be of inner type and the cross passes should be even.

6). Max baffle spacing generally equal to shell ID, actual it depends upon tube material , diameter
and thickness of tube

7). For multi tube pass the inlet nozzle for shell side will at the rear end of the shell. This is because we
want to make the flow in the first tube pass counter current. But this is specifically for single shell. In multi
shell exchangers the out nozzle of the 1st shell and the inlet shell nozzle of the 2nd shell should be at the
opposite end.

8). If the number of baffles is odd, it means that the shell side inlet and outlet is on the same side and if the
number of baffles is even then the inlet and outlet nozzles are on the opposite sides of the shell (* which is
normally the case).

9). Max allowable sealing strips are the key to adjust the C stream flow.

10). If the fluid on the shell side is two phase , the we always take vertical cut baffle.

11). ??? with the increase of tube passes the B-stream flow increases dramatically , find the
reason???????????

12). Minimum # of tube rows passing through overlap portions of double segmental cut baffle should be at
least two.

13). The max limit of baffle cut for single segmental baffle is 49% and for double segmental baffle is 30%.

14). The pressure drop in the J type shell is 1/8 th of the pressure drop of the E-type shell, Because in J type
shell the flow is divided in to two equal parts, and pressure drop is directly proportional to velocity square,
Length and density. SO when the velocity and length is divided in to half , the overall factor becomes 1/8th.

15). For this project we are instructed that the difference b/w the inlet and outlet baffle spacing should be
max 50, But not for J-type shells.

16). Also we r instructed that the out side tube limit i.e. OTL should not be greater than 1200, it means that
the diameter of the shell will not exceed 1250mm in any case.

17). With the increase of number of tube passes , the velocity in the tube increases, and with the increase of
shell ID, which results in the increase of number of tubes , the velocity decreases.
18). By increasing the number of cross passes or the number of baffles , the ratio of span Length to TEMA
max span decreases.

19). Cross flow amplitude can be reduced by decreasing the inlet and outlet baffle spacing. (But not below
the minimum requirement).
20). Vortex shedding ratio can be controlled with baffle spacing. In case of vortex shedding at the exit , we
should increase the outlet baffle spacing , and similarly in case of vortex shedding at the inlet we should
increase the inlet baffle spacing.

21). For re-boilers we normally prefer J type shells.

22). At 30 tube lay out the heat transfer is max, where as at 90 the heat transfer is min.

23). For calculating the tube sheet thickness , inlet / outlet spacing we take the deign temperature and
pressure , which ever is greater on either side,.

24). For two phase mixture , we will always provide impingement baffle at the inlet.

25). Normally in re-boilers the number of cross passes are max 10,this is from kosakai experienced results.

26). For the utilities (cw, hot oil) that are being through the exchanger , if we can reduce the shell side not
matter the pressure drop crosses the allowable limit, we should do it because the pump head increase may
be acceptable to process designer but not the larger shell with larger cost.

27). If there are more than one shell in series then HTRI assumes that shell as the first one in which hot
fluid is entering.

27). In J-type shells, the value of the inlet or outlet (whichever is in the center) baffle spacing will be half
of the value obtained from the graph.

28). If there are more than one nozzle on the shell side then we will have to give the equivalent higher
diameter to be entered as the shell side inlet nozzle size.

29). Normally we avoid film boiling in the exchangers because in this phenomena the liquid forms a sort of
gaseous film on the surface which reduces the heat transfer because the gaseous film is a poor conductor of
heat.

30). The fluid whose operating pressure is higher will preferably be placed in the tube side because if it is
places on the shell side then we will have to design shell as well as tubes for that high pressure, resulting in
large thickness of shell walls which means greater cost, but if we place it on the tube side then only the
tubes will have to be designed for that pressure resulting in lower cost of the exchanger.

31). If the fluid in the shell side is hot fluid which is to be cooled down then it is to be entered from the top
side , because when it is cooled down it density increases and it becomes heavy so it easily comes from the
bottom and if the fluid is cold fluid which is to be heated then it should be entered from the bottom because
its density decreases as it obtains heat and easily comes out from the top.

32). The ideal central spacing b/w the baffles is b/w 2/5 th to 3/5th of the shell ID. Between this value we
expect to see the best result.

32 a) In case of double seg. Baffle the no. of baffles will always be in even numbers, and the starting
baffle be the center one as shown here
33). For J-Type shells (Only with double seg. Baffle) the number of cross passes should be super even
(Like 8 is a super even number , Because when it is divided by 2, it generates an even number, but 6 is not
a super even number , as 6 by 2 is an odd number). Note : To be discussed with Farooq San , See page 36
of 41 of J-PEP S&T Heat Exchanger.
34). If the baffle cut is horizontal, then the no. Of cross passes must be odd, provided the inlet / outlet
nozzles are on the opposite directions.
35) The baffle cut orientation is relative to centerline of inlet shell side nozzle.
36). The flow in 1st tube pass ( co-current/counter current) depends on the shell side inlet nozzle location ,
i.e we will check whether the shell side inlet location is on the front head side or on the rear head side .and
then check .
37). We may consider installing Shell expansion joints in F-Type shells only , .Shell expansion joints
should not be considered for any other type of shell unless other wise specified by the client.
38). Floating head support plate should always be installed in the floating head type exchangers.
39).In heat exchanger design normally we intend to keep the tube pass max. keeping the shell ID min. , for
which the pressure drop values are in limit.
40). No of cross passes = No. Of baffles +1 ,
41). Increase in Baffle cut is also a good technique to increase the B-Stream flow.
42). The attachments provided to calculate the tube sheet thickness , dead space and inlet outlet baffle
spacing uses design temp, and pressure values, but we will take that value of design temp and pressure
which ever is greater on either side .
43). The tube sheet thickness( for floating head type) to be entered in HTRI is as follows

Tube sheet thickness = [Tube Sheet Thickness (From Graph) + Dead Space (From Graph ) ] - 100

 But for U-tube, the value of tube sheet thickness should be directly entered in the program, No
need to multiply by 2, and no dead space calculation required. But in case of U tube the distance
from the tangent line to last baffle should be mentioned in HTRI (Normally 50mm) or we add 2”
in the tube sheet thickness instead.
43 b.) The tube sheet thickness( for fixed head type) to be entered in HTRI is as follows
Tube sheet thickness = [Tube Sheet Thickness (From Graph) ]

43 b.) Concept of Parallel and Perpendicular Pass Lanes :

To determine that whether the pass lane is parallel or perpendicular , we have to consider two things in
mind
1. Baffle Cut
Y”
2. No. of Tube Passes

Case 1: (Four Tube Pass , Vertical Cut Baffle)

If the baffle cut is vertical , then flow direction will be parallel to XX” axis and perp. to the 3 4
YY” axis , so the pass lane from 1-2 is perpendicular because it is crossing the YY”
and the pass lane from 2-4 is parallel because it is crossing the XX” . X X”
Case 1: (Four Tube Pass , Vertical Cut Baffle)
1 2

If the baffle cut is horizontal, the flow direction will be parallel to YY” and
Perpendicular to XX”, so the pass lane from 1-2 is parallel pass lane and
the pass lane from 2-4 is a perpendicular lass lane. Y

Y”

3 4
X X”
1 2

Y
So normally (except for U tube) the parallel and perpendicular are equal. But in case of U-tube the situation
is different.
In U-Tube the 1-2, or 3-4 is a U bend, so in this case the pass lane width is not same for both parallel and
perpendicular, In this case this pass lane width either parallel or perpendicular is calculated as follows

Pass Lane Width = (2 * U Tube Radius) – Tube OD

Where U tube radius will be given in the project specification, normally it is 2D (D= dia of tube)

And the other pass lane will be same as the normal pass lane width specified in the pj specs.
Y”
It is necessary to specify both pass lane width, when both perp. and parallel pass lane
Exist.
Case 3:

If the tube pass is 2 , then only parallel or perpendicular pass lane will exist
2
so lets see in case of horizontal cut baffle the which pass lane exist , perp . or
parallel. X X”
As the baffle is horizontal cut so the flow is parallel to YY” and perpendicular to
XX”, so 1-2 is a perpendicular pass lane because it touches the XX” line which
1
Is perp. to the flow direction. As the pass lane goes from odd to even , i.e from 1-2 , so
In case of U tube the pass lane width (which is perp. in this case ) will be calculated
Using the u tube radius formula , i.e. Y

Pass Lane Width = ( 2* clearance b/w center of pass partition and Tube Center ) - ( Tube OD) Y”

Case 4: 2
so lets see in case of vertical cut baffle the which pass lane exist , perp . or
parallel. X X”
As the baffle is vertical cut so the flow is parallel to XX” and perpendicular to
YY”, so 1-2 is a parallel pass lane because it touches the XX” line which
1
Is parallel to the flow direction. As the pass lane goes from odd to even , i.e from 1-2 , so
In case of U tube the pass lane width (which is parallel in this case ) will be calculated
Using the u tube radius formula , i.e Y
Pass Lane Width = ( 2* clearance b/w center of pass partition and Tube Center ) - ( Tube OD)

But in case of 4 tube pass in U tube only , the pass lane from odd to even ( either perp. or parallel) will be
calculated using the U bend radius formula , and the pass lane from even to odd ( 2-3) , either perp. or
parallel will the normal pass lane width specified in the project specs.

44). Pass lane width can be calculated as follows (Not For U-Tubes)

Pass Lane Width = ( 2* clearance b/w center of pass partition and Tube Center ) - ( Tube OD)

The parallel and perpendicular pass lanes are equal if they exist both , as in case of 2 tube pass one of them
will not exist i.e. either parallel or perpendicular will exist , but in case of 4 tube pass both will exist. And u
have to enter both of them in tube pass more than 2.
45). The nozzle sizes mentioned in the D/S are nominal sizes , but in HTRI we have to enter the nozzle ID
against the schedule number .

46). Regarding the settings of Inlet , Outlet , and Central Baffle Spacing , 1 st please don’t enter the central
baffle spacing , rather we should control the baffle spacing using the no. of cross passes , once the
exchanger design is ok , then input the central baffle spacing ( To the nearest 10) corresponding the no. of
cross passes , and also input the inlet baffle spacing , keeping the outlet baffle spacing empty and let the
program decide that value , so that’s how we can control these parameters.

47). Pass Lane Concept in U-Tube :

In case of U-tube

In case of U-tube , if the tube pass is 2 , then the width of parallel and perpendicular pass lanes can be
calculated as follows (
Width = ( 2 * U Tube Radius ) – Tube OD

But in case of 4 tube pass of U tube , the width of parallel pass lane will be equal to the clearance b/w
center of pass partition and tube center, and the width of perpendicular pass lane will be

Width = (2 * U Tube Radius) – Tube OD

48). J type shells are of two types, one is J12 and the other is J21 , the 1 st digit in J12 describe the no. of
inlet nozzles and the 2nd digit in the J12 describe the no. of outlet nozzles, so in J12 there is 1 inlet and 2
outlet and in J21 there is 2 inlet and 1 outlet.
49).If the chen number is very high , i.e. beyond the 1300 limit , and there is no other way to reduce it ,
then changing the tube pitch is good option to reduce the chen number .For example if the chen number is
very high at triangular pitch , then by changing the pitch from triangular to rotated triangular (60 degree) ,
the chen number decreases dramatically.

49). If you enter the height under Nozzle (HN) at the shell outlet , then there is a possibility that some times
pass lane center doesn’t match with the shell horizontal center. So we should avoid entering (HN) at shell
entrance.

50). When any of the vortex shedding ratio and turbulent buffeting ratio in the Acoustic vibration check is
less than 0.8, then any value of chen number is acceptable.

 If any of the vortex shedding ratio and turbulent buffeting ratio in the Acoustic vibration check is
b/w 0.8 and 1.2, then Chen number value should be less than 4000.

 If any of the vortex shedding ratio and turbulent buffeting ratio in the Acoustic vibration check is
greater than 1.2, then Chen number value should be less than 1300.

51). The cross flow velocity should not exceed the critical velocity, other vise vibration is probable. The
critical velocity is also mentioned in the vibration report.

52). For bundle entrance / exit, if the vortex shedding ration is less than 0.5 or greater than 1.5, then it can
be ignored, because it will be checked later on VIB higher mode, But if the value is b/w 0.5-1.5, then we
will check the cross flow amplitude which should be less than 25% of the tube gap.

53). If we are using HTRI version 1, then we have to be very careful in case of boiling service, In that case
we have to enter 550F in the boiling point range cell manually, other wise the program will calculate
abnormal values of boiling point range. But this problem has been rectified in version 2.0.
54). If u specify inlet/outlet height under nozzle, then the program calculates the actual value of inlet/outlet
height under nozzle .For example first of all u haven’t input the inlet/outlet height under nozzle, the
program calculates say 34.5 mm height, now when u specify this height say 40 mm the program will
remove some tube rows and will give the actual calculated height under nozzle say 43.4. But some time by
inputting this height the pass lane center doesn’t match with the shell horizontal center .So please checks
this phenomenon after inputting this value.

55). The Min. tube sheet thickness for U tube type exchangers
is 50 mm.

56). Velocity limitations for Al-Br tube is between 0.7 - 2.5 m/s.

57). Bundle Entrance/Exit

Fluid elastic instability ratio < 0.8
* Vortex shedding ratio
if 0.5 < * < 1.0 Crossflow amplitude/Tube gap < 0.1
if * > 1.0 Crossflow amplitude/Tube gap < 0.1 (rev2)

Shell Entrance/Exit Parameters

Velocity (Entrance) < 0.8 X Critical velocity (Inlet)

* If impingement plate exists, this velocity is not taken into account. (rev1)

Velocity (Exit) < 0.8 X Critical velocity (Outlet)

58). Impingement plate size should be at least 50mm larger than the nozzle ID. Therefore, please set the
size of impingement plate at 'shell ID + 50mm' (From Cosmo Pj. Experience). So in HTRI we will
manually input the impingement plate size (i.e. plate width and length , which are equal).

59). Turbulent beffeting vibration message (when it increases from 0.5) can be ignored , when the turbulent
beffeting amplitude is extremely low .

60). In general, for horizontal cut baffles, parallel partition arrangement, pattern 4-1, is most preferable in
view of effectiveness of bundle cross flow
1