Академический Документы
Профессиональный Документы
Культура Документы
c ~
(a) c (b)
2r
~1ode l R/T r/R t/T tn/T r/rp rz/tn
Mor H.Lr Mtr Mob Hi.b Mtb
UA 50.5 0.5 0.5 0.5 0.990 1.00 1. 10 4.62 4.35 16.6 7.00 1. 39
B 40.5 0.5 0.5 0.5 0.998 1.00 1.09 4.54 4.18 15. 1 6.54 1.32
c 20.5 0.5 0.5 0.5 0.976 1.00 1. 05 4.09 3.53 10.8 5.17 1. 2 5
D 10.5 0.5 0.5 0.5 0.955 1.00 1.04 3.60 3.41 5.77 3.28 1. 39
E 5.5 0.5 0.5 0.5 0.917 1.00 1.03 3. 1 7 2.79 3.50 2.64 1. 43
F' 5.5 0.08 0.08 0.08 0.917 6.25 1.04 3. 11 2.12 l. 27 1.36 1.02
SlA 50.5 0.5 0.5 4.34 0.861 0.500 1.04 2.19 2.98 11. 1 2.19 1. 16
B 40.5 0.5 0.5 4.01 0.843 0.500 1.04 2.23 2.87 9.84 2.09 1.16
c 20.5 o.s 0.5 3. 14 0.780 0.500 1.04 2.25 2.43 5.64 1. 51 1. 17
D 10.5 0.5 0.5 2.45 0.705 0.500 1.04 2.07 2. 11 2.81 1.42 1 • 16
E 5.5 0.5 0.5 1. 92 0.623 0.500 1. 03 2.07 1. 70 1. 56 1. 49 1. 1 3
F 20.5 0.32 0.32 2.56 0.732 0.500 1.04 2.36 2.01 2.56 1.22 1.06
G 10.5 0.32 0.32 1.98 0.649 0.500 1.04 2.23 1. 92 1. 43 l. 37 1.07
H 5.5 0.32 0.32 l. 52 0.563 0.500 1.03 2.08 1.68 1.39 1. 37 LOS
I 20.5 0.16 0.16 1. 88 0.646 0.500 1. OS 2.49 1.81 1.22 1. 2 3 1.02
J 10.5 0.16 0.16 1.43 0.555 0.500 1.04 2.38 l. 78 1.26 1.25 1.02
K 5.5 0.16 0.16 1. 08 0.468 0.500 1.02 2,33 1. 64 1.33 1. 32 1.03
L 20.5 0.08 0.08 1.38 0.551 o.soo 1.06 2.52 1.76 1.18 1.22 1.01
H 10.5 0.08 0.08 1. 03 0.459 0.500 1. 04 2.61 l. 7 5 1. 18 1. 21 1. 01
N 5.5 0,08 0.08 0.72 0.391 0.500 1.03 2.63 1.68 1 • 21 1.20 1.02
E'30A 50.5 0.32 0.32 3. 19 0.808 0.500 1.03 2.24 2.30 3.73 1. 19 1.08
B 20.5 0.32 0.32 2.13 0.743 0.500 1.04 2.39 2.13 1.84 1,24 1.08
c 10.5 0.32 0.32 1. 60 0.695 0.500 1. 03 2.27 1.97 1.39 1.32 1.07
D 5.5 0.32 0.32 1.23 0.659 0.500 1.03 2.05 1. 7 3 1.33 1.35 1.05
E 5.5 0.08 0.08 0.533 0.556 0.938 1. 03 2.68 1. 72 1. 19 1. 20 1. 01
(a) c #
2r
= a/ (H/Z r )
(b) Czb = o/(M/Zb)
23 1 S.G. 20.7 1.00 1.00 3.68 8.03 6.8oc 9.33 12.14 1o. 4 9
2 s.G. 12.4 1.00 1.00 2.58 5.35 5 .18c 6.65 8.14 7.38
3 S.G. 7.6 1. 00 1. 00 1.68 3.48 3.55 4.14 4.48 4.36
4 S.G. 5.7 1.00 1. 00 1. 72 2.87 3.20 3.53 3.92 4.53
(c) Maximum and minimum principal stresses have same signs, except for these two cases:
a
max
= 6.68, a i
mn
= -6.80 ; a
max
-5.18, a i = 0.23
mn
connections it only represents a judgmental evalua- Figs. 14(d) and (e) illustrate other possible moment
tion of the effect of the three combined moments. combinations. Fig. 14(e) is the pure run moment case
ANSI B31.1 and the ASME Code both combine for which we do have some data as discussed in Section
stresses by: 4.4.
Fig. 14(d) illustrates the more complex case; the
SE = i[~ + MJ + A(;?jl 12/Z (34)
ASME Code Class 1 method of separating these into
To the extent that i = max(i 0 , ii, i1), which is generally branch moments and run moments is shown. The total
the intent, and for branch connections where io ii and i 1 calculated stress is then obtained by adding the stress-
are different, then Eq. (34) would be more conserva- es due to the branch moments to those due to the run
tive than Eq. (33). Both ANSI B31.1 and the ASME moments. ASME Code for class 2/3 piping and ANSI
Code also use Eq. (34) for run moments. Calculated Codes follow a conceptually different procedure in
values of S E for both the branch end and the run ends that each of the three ends is checked separately.
must be less than the Code allowable stress. Comparisons between these two conceptual methods
Fig. 14 illustrates a problem in evaluating combined is discussed in detail in Ref. 27 so we will not discuss it
moments. Figs. 14(a) and 14(c) show the combination further except to note that:
of moments for which we have i/s for branch moments. 1. The conceptual difference is significant only for
However, there is an infinity of possible run moments the type of moment combinations illustrated by
between (a) and (c) which will balance the branch Fig. 14(d).
moment and which might occur in piping systems, one 2. For a narrow range of branch connection parame-
of which is shown as (b). Fig. 14(b) is of particular ters and moments, the ASME Code Class 1 meth-
interest because 297 20 and Fujimoto 21 analyses are od is more conservative by a factor of up to two.
based on these run end conditions. 3. Neither conceptual method can be demonstrated
If a fatigue test were run with the end conditions to be accurate or even relatively more accurate.
shown in Fig. 14(b), would the resulting ir be different
from (a) or (c)? We do not have any such tests, but 4.6 Branch Connection Description Inconsistencies
would speculate that if r/R is less than about%, the In the quest for more accurate i-factors, a desirable
difference would be small. However, for r/R = 1.0 Code characteristic is that for a given configuration of
there might be a difference in that ir for Fig. 14(b) branch connection the Code should give the same i-
would be less than for (a) or (c). It is the latter that we factors. However, note the following:
have i/s for; hence, in this sense our i/s may represent
upper bounds. The ASME Code, Class 2/3 piping, for a UFT gives:
(a) From Table 10, maximum of c 2 ' for Mar• Mir' Mtr' Maximum is either from Mir or Mtr; where from
Mtr' value is followed by a 10 t".
(b) From table 11, maximum of c 2; for Mar' Mir or 1/2 of c 2; for Mtr" Maximum is either from Mir or
Htr; where from Htr' value Is followed by a "t",
-1Q _ _ L 20-
-10
(d) Branch and Run Combination
2.2.
4.7 ANSI B16.9 Tees, Sweepolets (Bonney Forge Trade
name)
+ -10_1_ 10 In order to keep this report from becoming even
more complex than it is, we have not given data on
0 ANSI B16.9 tees or Sweepolets. There is a fairly sub-
stantial amount of data on B16.9 tees. Data are avail-
(e) Pure Run Moments able for r!R = 1.0 and for r/R = "'-'0.5; but nothing in
10 -10
between. Accordingly, we do not know if there is a
Fig. 14-lllustration of combinations of branch and run moments peak in the SIF for Mob as suggested by Figs. 6, 7 and 8.
At present, plans are being made to fatigue test some 4
x 3, std. wt. ANSI B16.9 tees with Mob loadings. These
tees have an r/R ratio of 0.77 and should give some
indication as to whether a peak does exist.
types of branch connections. This, in the long run, will Sweepolets in sizes 12 x 6 and 14 x 10, both standard
provide improved Code guidance for adequate but not weight, have been fatigue tested with both Mob and
over-costly piping systems. However, there is an in- M,b loadings. The r/R ratio of these two sizes is 0.51
consistency between UFT's and the "Branch welded- and 0.76, respectively. The Mob tests indicate that
on fitting (integrally reinforced)" which merits some there is a peak somewhere around 0.75. The Mib tests
discussion. agree with the general relationship (see Figs. 6-10)
First, footnote 7 tied to "Welded-on" reads: "The that the it for M;b is much lower than for Mob and there
designer must be satisfied that this fabrication has a is no significant peak as a function of r/R.
pressure rating equivalent to straight pipe." Now,
there isn't anything simple about reducing-outlet 4.8 Stress Limit, Sx
branch connections so we ask the question: Which As indicated by Eq. (12), having calculated SE the
straight pipe, the run or the branch? We think the Code then provides a limit; SE.::::: Sx. The stress limit is
intent is the run pipe so that question could be an- an important part of assessing the significance of the
swered by inserting the word "run" before pipe in the accuracy of i-factors. The Codes prescribe the stress
footnote. The question then arises as to how the de- limit, Sx, as:
signer meets the requirement of footnote 7. Presum-
ably, the intent is that the designer orders fittings
from a manufacturer with a designated wall thickness where
(e.g., Sched. 40) with, perhaps, a requirement in his f = cycle dependent factor ranging from 1.0 for
purchase order that the fitting must have a pressure 7000 cycles to 0.5 for >100,000 cycles
rating equivalent to the desired schedule run pipe. Sc = allowable stress at cold temperature in cycle
There appears to be a couple of ways the manufac- sh = allowable stress at hot temperature in cycle
turer could assure himself and his customers that his Ss = sum of longitudinal stresses due to pressure,
fittings, when properly welded into designated run weight and other sustained loads.
pipe, would have a pressure rating equivalent to the
The significance of the stress limit is discussed in de-
run pipe:
tail in Ref. 27. For the purpose of this report, we make
1. Run burst tests. the following observations;
2. Show compliance with paragraph 304.3 of B31.3,
(1) For materials like ASTM A106 Grade B carbon
using designated wall thickness rather than cal-
steel at temperatures up to about 600 °F, with
culated by Eq. (2) of B31.3.
S, and S11 from the ASME Code or from B31.1,
Now the potential inconsistency arises because UFT's there is a margin between failure and Code a!-
For Mz 3:'
Code guidance is good and that k = 1 for branch strength for the cyclic moment, Mob?
means: assume a rigid juncture. He is faced with the We do not know much about the flexibility of a pad
dilemma of changing the piping system in Fig. 15 so it reinforced branch but, since a pad is usually welded to
does meet the Code. He might consider changing the the run pipe at its inner and outer peripheries, the
piping such as sketched in dashed lines in Fig. 15. This flexibility might be estimated by using the equation in
would be very expensive, so the designer might look at Fig. 15 for Mob, but using 2.5T instead ofT. This would
the possibility of using a pad reinforcement. By using a give a flexibility factor of:
pad thickness of 1.5T, he can reduce the SIF to 4.14; kp (for M 0 b) e=: 46.6/(2.5) 2 = 7.5.
his calculated S E is then 33.8 ksi and this might meet
Code Sx limits. Let us suppose that it does and ask Now, from Ref. 28 data, fork of 8, it appears that the
what the designer has accomplished by using a pad. moments would be overestimated by a factor of
First, since this piping system is assumed to go up to a around 3 rather than a factor of 9 for k = 47. This
temperature of 500° F, the pad may cause high ther- means that the pad would cause the moments to in-
mal gradient stresses in the 30 in. pipe and thereby crease by a factor of about 9/3 = 3. Assuming that the
reduce its reliability. Has he improved the fatigue i-factors for UFT and pad reinforced branch indicate
Fig. 16 if
Nominal . a
Type iden. and r/R R/T t/T r/rp l.f
size (R/T) 2 /3(t/T)
group iden.
UFT •A 8
12
X
X
6
10
0.764
0.839
12.9
16.5
0.870
0.973
0.958
0.966
5.84
8.34 2
1.22
1. 32
4 X 4 1.00 8.99 l. 00 0.947 2. 71 2 0.63
Source 1i b t/T
table Hodel R/T r/R t/T i ib t /T ib t/T
r/rp if (a) (b)
number f
UFT
2 4 X 4 8.99 1. 00 l.OO 0.947 2.34 3.68 1. 57 3.68
2 4 X 4 10.6 1.00 1.00 0.955 2.95 4. 15 1.41 4.15
2 4 X 4 22.0 1. 00 1. 00 0.978 6.12 6.91 1.13 6. 91
2 4 X 4 41.8 1. 00 l.OO 0.988 11.0 10.7 0.97 10.7
3 6 X 6 12.0 1.00 1. 00 0.960 3.62 4.53 l. 25 4.53
3 20 X 4 41.4 0. 19 7 1.00 0.942 2.67 6.79 2.54 7. 51
3 20 X 8 24.6 0.375 0.50 0.974 2. 7 5 2.54 o. 92 3.78
3 20 X 14 24.6 0.702 0.60 0.983 3.47 3.51 1. 01 6.27
3 20 X 20 41.4 1.00 l.Oa 0.988 6.90 10.6 1.54 10.6
5 8 X 6 12.9 a.764 a.87a a.958 l. 85 3.36 I. 82 6.01
Weld On
3 4 X 4 8.99 1. oa 1. oa o. 6 3 1.75 2.45 1.40 2. 45
3 4 X 4 8.99 1. 00 1. 00 0.79 1.!36 3.07 1.65 3.07
3 12 X 6 16.5 0.513 o. 7 4 7 0.675 1.28 2.09 1.64 3.51
3 8 X 4 12.9 0.513 a. 7 36 0.79 0.81 2.05 2.53 3.44
Insert
3 14 X 6 18.2 0.466 0.747 0.83 0.98 1. 35 1. 38 2.20
3 12 X 8 16.5 a. 6 71 a.859 a.82 1.52 l. 58 1.04 2.80
3 8 X 4 12.9 a.513 a.736 o. 775 1. oa 1.00 1.aa 1.69
5 12 X 6 16.5 a.513 a.747 0.86a 1.32 1. 33 1. at 2.24
5 12 X 8 16.5 o. 6 71 0.859 a.82a 1.31 1. 58 1. 21 2.80
5 12 X 8 16.5 0.671 a.859 a.874 1.53 1.68 1.10 2.99
ten the equation to cover the probably more common deemed prudent from available fatigue test data.
case of r 2-not-provided. Equations (43) for ib does not
Table 15 summarizes available Mob fatigue test
have the (t/T) factor but that is not really a change
data, previously given in Tables 2, 3 and 5. Calculated
because of (lOc). Note in this respect that the present
values of ib(t/T) by Eqs. (43) or (44) are shown. The
rather complex instructions for reducing outlets leads
right-most column shows the ratio of i6(t/T)/ir. Con-
to exactly the same SEas our recommended note: "For
sidering the scatter encountered in fatigue tests, we
checking branch ends, use i(t/T) in place as i and Z =
consider the correlation to be adequate. In particular,
Zb." By taking the (t/T) out of Eq. (43), this instruc-
the proposed ib adequately solves the Mob inconsisten-
tion applied to all branch connections/tees.
cy. Note that the 8 x 6 and 12 x 10 UFT's are encom-
The change in the equation for ib is intended to:
passed by ib, and the 12 x 6 Weld On is brought into
(a) Provide a single ib, conceptually the maximum of reasonable consistency with the 4 x 4 Weld Ons. Also
i0 b, i;b, i1b, for use with the resultant branch mo- note that an appropriate credit is given for an outer
ment. This is a continuation of present practice, fillet radius, rz; i.e., for the 20 x 6 and 20 x 12 Extruded
but the ASME might wish to consider adopting outlets and all Inserts.
the B31.3 concept of different i-factors; see Rec- While ib provides a good fit to the fatigue test data,
ommendation (12). it seems to pose an anomaly with respect to calculated
(b) Provide an ib that covers the relatively high i- stresses. Assuming that (R/T) 213 is an accurate param-
factors for Mob in the r/R range between about 0.5 eter, then the ib equation (for r/rp = 1) appears as
and 1.0. shown in Fig. 8. If Kzb =La, then we would expect it to
(c) Reduce the over-conservatism in ib to the extent be below the theoretical curve by a factor of 2.0. But
Source ib t/T
table Hodel R/T r/R t/T ib t/T
r/rp if (a)
number f
UFT
2 4 X 4 8.99 1.00 1.00 0.947 2.71 3.68 1. 36
3 20 X 12 9.5 0.635 0.687 0.946 3.9 3.48 0.89
5 8 X 6 12.89 0.764 0.870 0.958 5.84 6.01 1. 03
5 12 X 10 16.5 0.839 0.973 0.966 8.34 8.37 1.00
Weld On
3 4 X 4 8.99 1.00 1.00 0.63 1. 65 2.45 1. 49
3 4 X 4 8.99 1.00 1.00 0.79 1. 72 3.07 1. 79
3 6 X 4 11.33 0.672 0.846 0.63 2.20 3.31 1.50
3 6 X 4 11.33 0.672 0.846 0.63 1. 87 3.31 1.77
3 12 X 6 16.5 0.513 0.747 0.675 3.78 3.51 0.93
5 8 X 3 12.89 0.396 0.671 0.773 3.20 2.69 0.84
5 8 X 4 12.89 0.513 0.736 0.812 3.49 3.53 1.02
5 8 X 4 12.89 0.513 0.736 0.853 3.45 3.71 1. 07
5 8 X 5 12.89 0.639 0.801 0.801 4.20 4.23 1.01
5 8 X 6 12.89 0.764 0.870 0.832 4.73 5.22 1.10
5 8 X 6 12.89 0.764 0.870 0.868 3.95 5.44 1.38
5 8 X 8 12.89 1.00 1.00 0.852 5.19 4.22 0.81
Extruded
6 4 X 4 4.71 0.943 0.494 0.947 1.49 1. 60 1.07
6 6 X 4 5.39 0.703 0.422 0.947 1.65 1. 55 0.94
6 8 X 4 5.50 0.539 0.330 0.947 1.48 1.sob 1.01
6 16 X 4 7.26 0.285 0.230 0.947 1.23 1. sob 1. 22
3 20 X 6 9.5 0.326 0.432 0.935 1.2 1.50b,c 1. 25
3 20 X 12 9.5 0.635 0.687 0.946 2.5 1. 74c 0.70
Insertc
3 14 X 6 18.2 0.466 0.747 0.83 2.64 2.20 0.83
3 12 X 8 16.5 o. 671 0.859 0.82 2.18 2.80 1.29
3 8 X 4 12.9 0.513 0.736 0.775 1.89 1.69 0.89
5 12 X 6 16.5 0.513 0.747 0.819 2.25 2.13 0.95
5 12 X 6 16.5 0.513 0.747 0.860 2.44 2.24 0.92
5 12 X 8 16.5 0.671 0.859 0.820 2.75 2.80 1.02
5 12 X 8 16.5 0.671 0.859 0.800 2.25 2.74 1.22
5 12 X 8 16.5 0.671 0.859 0.874 2.41 2.99 1. 24