Boiler Riser Tubes
Boiler Riser Tubes
Boiler Riser Tubes
Habib1
e-mail: [email protected]
I. Al-Zaharnah
M. El-Shafei
S. A. M. Said
N. Merah
Department of Mechanical Engineering,
King Fahd University of Petroleum and Minerals,
Dhahran 31261, Saudi Arabia
S. Al-Anizi
M. Y. Al-Awwad
M. Hajji
Department of Consulting Services,
Saudi Aramco,
Dhahran 31311, Saudi Arabia
Introduction
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
a quadratic programming optimization to find the optimal reference values for boiler firing rate, steam pressure, and steam temperature during start-up. Wang et al. 7 provided an investigation
on intelligence-based hybrid control for power plant boiler. A hybrid classical/fuzzy control methodology was presented to integrate low-level machine control and high-level supervision for the
steam temperature process of the power plant boiler. The results
showed that the steam temperature can be stably maintained
whenever disturbances are large or small, and a better control
performance as compared with traditional methods was achieved
7. Gough 8 applied an advanced control of steam superheat
temperature on a utility boiler. The paper described an adaptive
predictive model to the control of the superheater steam temperature. It was proposed to use a variable speed water pump and a
control system for the pump speed to account for the varying
pressure difference between water feed and drum pressure.
Some research papers accommodated the working stresses of
boilers in the control design and analysis of boilers and steam
generators. A new control solution aiming at improving the
start-up procedures of boilers and explicitly taking the thermal
stress values of critical components into account has been presented by Kruger et al. 6,9. The presented boiler model has
shown that it is possible to obtain boiler models using only state
variables, which can be measured or only a few additional nonmeasurable states. An on-line optimization of drum boiler start-up
was investigated by Franke et al. 3. The cost function of the
optimization model was taken to be in terms of the deviations of
the steam pressure and steam flow rates from their reference values, while the thermal stress was taken as an output constraint.
Examples of an improved main steam temperature control and
operation of a drum boiler turbine bypass system were presented
by Peet and Leung 10 to demonstrate the value of dynamic
capability to maintain the steam temperature deviation over wideload ranges or during rapid load changes, which can increase the
permissible rate of load change due to thermal stress limitations.
Taler et al. 11 developed a new method for determining thermal
stresses in boilers during start-up. They presented a new method
of transient temperature field identification in the drum on the
basis of measured temperature changes at locations on the outer
insulated surface of the drum. The methods accuracy was demonstrated by a comparison of thermal distributions obtained by
FEM. An analysis of transient temperature and stress distribution
in boiler components during start-up and shutdown operations has
been presented by WJglowski et al. 12. The thermal stresses
have been determined indirectly on the basis of measured temperature values taken at selected points on the outer surface of a
construction element. Monitoring of boilers thermal and strength
operation conditions allowed finding the dangerous loads on
power boilers pressure elements during the boiler start-up and
shutdown operations.
From the above literature review, it can be seen that none of the
authors addressed the specific problem of thermal stress variation
due to swing rate in the boiler riser tubes. Most of the research
works that dealt with thermal stress variations in boilers and heaters have done it for start-up and shutdown operations only. The
problem of the effect of different swing rates on thermal- and
pressure-caused stresses has not been explicitly addressed, nor has
this stress fluctuation effect on the tubes life studied. The objective of this research paper is to analyze the combined thermal- and
pressure-induced stress response of the drum boiler riser tubes due
to rates of heating load that vary under controlled conditions with
the changed rates of the steam flow rate. The nonlinear dynamic
governing equations are integrated simultaneously using MATLAB
for solving the water volume, the pressure, and the steam quality
at the exit of the riser tubes. The inner and outer temperatures of
the riser tubes are calculated after solving for the heat transfer
coefficient at the inner wall of the riser tubes at different time
steps. For different swing rates of steam flow, the combined
stresses are computed. A FEM model has been used for calculat061301-2 / Vol. 132, DECEMBER 2010
4996.6
263.9
66.5
208.2
40
20
140,000
37
0.0672
0.0762
160,000
11
0.5
ing the combined effective stresses in the riser tubes. The possibility of damage induced by resulting temperature and stress
variations is investigated. An optimized control approach is introduced in the study.
Modeling
Fig. 1
loop
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
a11 a12 0
a21 a22 0
a31 0
a33
dP/dt
fw m
s
m
+m
fwh fw m
sh s
dVwt/dt = Q
dx/dt
dc
Q xh fgm
w
s
+ Vst
P
P
a12 = w s
a21 = Vwt hw
1.136
0.9
667.2
0.7
C5 0.3
0.6683
0.2
1058
0.7
0.3 C5 = 0 for vertical tubes and =0 for horizontal tubes
if Fr is greater than 0.04
w
hw
s
hs
tm
+ w
+ s
+ Vst hs
Vt + mmtC p
p
p
p
p
p
C1
C2
C3
C4
a31 = w
CO 0.65 CO 0.65
Re =
hw
w
s
hs
Vr
xh fg
+ s
1 Vr + 1 xh fg
p
p
p
p
+ s + xw sh fgVr
tm
Vr + mmtC p
p
p
To Ti
Rpipe
and
qDo =
To Ts
Rconv
Rpipe =
lnDo/Di
2Kpipe
and Rconv =
1
D ih
13
r
m
and Di is the pipe diameter
Ar
CO = convection number =
q
Gh fg
15
1x
x
0.8
s
w
14
0.5
16
G2
l2gDi
2.2 Thermal Model. The following equations are used to calculate the inside and outside wall temperatures of the riser tube:
qDo =
G = mass flux =
where is the average value of the void fraction along the riser
tube, Vr is the volume of the riser tubes, mmt is the total mass of
the system, and tm is the metal temperature. w, s and hw, hs are
the density and specific enthalpy of water and steam; h fg is the
specific enthalpy of evaporation h fg = hs h f .
12
G1 xDi
E pt = Psp Pdt
Qt = Qi + kqp E pt + kqi
E ptdt + kqd
17
dE pt
dt
18
where Psp is the drum pressure set point; Pdt is the measured
drum pressure; Qi is the initial heat flow; and kqp, kqi, and kqd are
the proportional, integral, and derivative control parameters,
respectively.
To maximize the swing rate, within the operation constraints,
the problem is formulated as a constrained optimization problem,
where an objective function is minimized with respect to the control parameters and the steam demand. The objective function is
the weighted sum of several performance objectives. The swing
rate is taken to be the percent change in steam load over a period
of 1 min. The performance objectives include the following:
11
where Re is the Reynolds number of the flow inside the tube and
Pr is the Prandtl number for water. Pr is kept constant throughout
the time domain of the solution.
10
where hL is the single phase heat transfer coefficient and is calculated from the DittusBoelter equation 14
Nu = 0.023Re0.8Pr0.4
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
effective = 2 + l2 + r2 l + l + lr1/2
19
= T + P
20
r = rT + rP
21
l = lT + lP
22
where
r2 + ri2
E
2
1 r r2o ri2
r =
ro
T rdr +
ri
T rdr T r2
ri
ro
r2 ri2
E
2
1 r r2o ri2
z =
E
1
23
T rdr +
ri
T rdr
24
ri
ro
r2o ri2
T rdr T
25
ri
rP =
P =
ri2 P
r2o ri2
ri2 P
r2o
ri2
r2o
r2
26
1+
r2o
r2
27
lP = 0
28
The transient thermal stresses part of the total stresses of the riser
tube are calculated based on the identified inner and outer wall
temperatures of the tube using Eqs. 2325. The temperature
061301-4 / Vol. 132, DECEMBER 2010
75
70
level and drum pressure, and maximum rate of feed water. The
optimization problem is then solved using a genetic algorithm
GA. A genetic algorithm is a method for solving both constrained and unconstrained optimization problems that is based on
natural selection, i.e., the process that drives biological evolution.
The genetic algorithm repeatedly modifies a population of individual solutions. At each step, the genetic algorithm selects individuals at random from the current population to be parents and
uses them to produce the children for the next generation. Over
successive generations, the population evolves toward an optimal
solution 15.
65
60
55
50
45
40
0
3000
6000
9000
12000
15000
Time (s)
Solution Procedure
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
65
5000
4900
Calculated Pressure
Experimental data
4800
Pressure, kPa
4700
4600
4500
4400
4300
4200
4100
60
step rise = 5%
step rise = 10%
step rise = 20%
step rise = 40%
55
50
45
40
4000
0
5000
10000
Time, s
15000
20000
300
900
1200
1500
Time (s)
600
the outer wall temperature oscillates within 4.8 K before stabilizing at the steady state value but also reflect the variations in the
steam quality as it affects the heat transfer coefficient. The increase in temperature on the inner wall occurs at a rate of about
110 K/h, and on the outer wall, at an approximate rate of 170 K/h.
The rate of change on the outer wall is comparable to that of a
typical start-up operation, which was estimated at 167 K/h for the
drum inlet 19. The present temperature change occurred at the
steady state temperature of 545 K, while the typical start-up occurred at 360 K. Furthermore, the temperature rise in this case is
very low to cause any kind of thermal shock. The next question to
be considered is how the different swing rates affect the stress
distribution in the riser tube and the service life of this component.
Figure 8 presents the effective stresses at the inner and outer
surfaces of the riser wall tube. These stresses are the combined
effects of the variations in the pressure and temperature across the
walls. The highest values of maximum stress 53.5 MPa and
stress range 6 MPa occur at the inner wall and are due to 40%
step rise. The stresses due to pressure are dominant in this transition zone about 80%. Even though the amplitude of the stress
cycle is low, the cycling of thermal and mechanical stresses even
with low amplitudes is always feared because they may cause
premature fatigue failure. The effect of the present stress reversals
on the fatigue life of the riser tube is examined in the following:
The stress amplitude is estimated at 3 MPa from the following:
a =
53.5 47.5
max min
=
2
2
The mean stress m is taken here as the steady state stress of 51.5
MPa. To include the mean stress effect for asymmetric cycles,
4600
5000
4550
4900
4800
Calculated Pressure
Experimental data
4500
P (kPa)
Pressure, kPa
4700
4600
4500
4400
4400
4300
4350
4200
4300
4100
4000
11000
step rise = 5%
step rise = 10%
step rise = 20%
step rise = 40%
4450
11500
12000
12500
Time, s
13000
13500
14000
Fig. 4 Calculated and experimental results close view of response of drum pressure to variations in steam flow rate
300
600
900
1200
1500
Time (s)
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
538
54
53
Effective stress (MPa)
T (K)
537
536
step rise = 5%
step rise = 10%
step rise = 20%
step rise = 40%
535
534
52
51
step rise = 5%
step rise = 10%
step rise = 20%
step rise = 40%
50
49
48
533
47
300
600
900
1200
1500
300
600
Time (s)
a- Inner wall
1200
1500
a- Inner wall
547
30
546
Effective stress (MPa)
29.5
545
T (K)
900
Time (s)
544
step rise= 5%
step rise = 10%
step rise = 20%
step rise = 40%
543
29
step rise = 5%
step rise = 10%
step rise = 20%
step rise = 40%
28.5
28
542
541
27.5
300
600
900
1200
1500
Time (s)
300
600
900
1200
1500
Time (s)
b- Outer wall
b- Outer wall
Fig. 8 Temporal combined pressure and thermal stress profiles at the inner and outer walls in response to variations in
steam flow rate
results show that the value of swing rate does not significantly
affect the stress level, especially in the region before the midwall
thickness.
aeq =
3
a
= 3.53 MPa
=
51.5
m
1
1
345
Sut
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
55
5%
10%
20%
40%
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
T (K)
542
541
540
539
538
537
536
535
534
45
40
35
30
25
Radius, (r-ri)/(r-ro)
a- at10 seconds.
5%
10%
20%
40%
T (K)
544
542
540
538
536
a- at 10 seconds
546
5%
10%
20%
40%
50
55
45
40
35
30
25
534
Radius, (r-ri)/(ro-ri)
Radius, (r-ri)/(r-ro)
b- at 60 seconds
548
546
544
542
540
538
536
534
b- at 60 seconds
T (K)
5%
10%
20%
40%
50
5%
10%
20%
40%
55
5%
10%
20%
40%
50
45
40
35
30
25
Radius, (r-ri)/(ro-ri)
Radius, (r-ri)/(r-ro)
c- at 120 seconds
c- at120 seconds
Fig. 10 Wall stress distributions at 10 s, 60 s, and 120 s
Nomenclature
heat transfer coefficient at the inner wall is evaluated using a
thermal model that includes the effects of convection evaporation
and nucleate boiling, and the pipe wall resistance is calculated for
estimating the temporal inner and outer wall temperatures. Both
the steam pressure and the riser tube inner and outer wall temperatures were introduced in a finite element model for finding the
effective combined stresses. The temperature distribution across
the wall of the pipe due to the swing rate values is similar at
different time stages, and the levels between the thermal stresses
at the inner and outer walls increase as the time progresses with
higher values of swing rate. The combined temperature- and
pressure-induced stress fluctuations in the riser tube due to swing
rate were found to be too low to result in a riser tube damage,
provided that the boiler and tube were initially at steady state and
that the start-up and shutdown cycles were taken care of in the
design of the boiler system. The proposed control strategy for
minimizing the overshoot temperature and stress responses was
found to be successful in reducing both the peak values and the
rise time.
Acknowledgment
The authors wish to acknowledge the support received from
King Fahd University of Petroleum and Minerals and Saudi
Aramco during this study.
Journal of Pressure Vessel Technology
Ar
CP
D
E
G
h
g
h fw
hf
hg
hs
hw
h fg
k
K
Q
dc
m
fw
m
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
54
Stress (MPa)
53
52
51
Current Control
Optimized Control
50
49
48
47
0
300
600
900
1200
1500
Time (s)
a- Inner wall
30
Stress (MPa)
29.5
29
Greek Symbols
volume fraction of steam in the riser thermal
expansion coefficient
average volume fraction of steam in the riser
Poissons ratio
dynamic viscosity kg/ms
w saturated water density kg/ m3
s saturated steam density kg/ m3
l liquid phase density kg/ m3
v vapor phase density kg/ m3
l longitudinal stress kN/ m2
tangential stress kN/ m2
r radial stress kN/ m2
References
28.5
Current Control
Optimized Control
28
27.5
27
0
300
600
900
1200
1500
Time (s)
b- Outer wall
Fig. 11 Current and optimized controlled combined thermaland pressure-induced stresses for the 40% increase in steam
flow rate
Downloaded 20 Jan 2011 to 144.177.100.5. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm