Note: Descriptions are shown in the official language in which they were submitted.
CA 02288367 1999-11-02
Case 6075
HYDRAULIC EXPANSION PRE-STRAINING OF
HEAT EXCHANGER TUBING
FIELD AND BACKGROUND OF THE II~VENTI~C~N
The invention relates generally to heat exchangers and more
1 o particularly to the pre-treatment of tubes for such heat exchangers.
DESCRIPTION OF THE PRIOR ART
The once-through steam generators or heat exchangers,
associated with nuclear power stations and which transfer the reactor-
produced heat from the primary coolant to the secondary coolant that
drives the plant turbines may be as long as 75 feet and have an outside
diameter of about 12 feet. Vlfithin one of these heat exchangers, tubes
through which the primary coolant flows may be no more than 5/8 inch in
outside diameter, but have an effective length of as long as 52 feet
between the tube-end mountings and the imposing faces of the
tubesheets Typically, there may be a bundle of more than 15,000 tubes
in one of these heat exchangers.
In the construction of a once-through steam generator, a plurality of
these small diameter long length tubes are configured in a square array
where they are welded at their top and bottom ends to a tubesheet to
mat«tain this array in the once-through steam generator.
The original once-through steam generators were fabricated using
a sequence where tubes, prior to welding to both tubesheets, were
individually electrically heated such that cooling of the hot tubes after
welding to the tubesheet resulted in tensile strains. This fabrication
1
CA 02288367 1999-11-02
Case 6075
method is not recommended since in the thermal method of tube
prestraining used on the original once-through steam generator the tubes
were heated individually until the desired thermal strain was achieved and
then seal-welded in place- Thus, for the first seal welded tubes, the
desired prestrain was achieved exactly. As the procedure progressed, the
previously welded tubes cooled and started to load the secondary shell
and the tubesheets. In response, these components deflected in the
direction of the load and effectively decreased the length of the
subsequently welded tubes. This mechanism introduced an unwanted,
1 o unconuolled and undefined tensile strain in these tubes. Excessive
tensile stress was dettimental to the tube life. In addition, thermal
prestraining is an expensive and time consuming process.
Since both the tubes and the shell of the once-through steam
generator are restrained by the tubesheets at both ends, interaction
stresses develop during operation due to the relative deformation of the
steam generator shell and the tubes- These interaction stresses come
from several sources. (1) Both the primary and secondary pressures
elongate the secondary shell of the vessel between the two tubesheets;
(2) the combined action of the primary and secondary pressures changes
the tube radius which, in turn. causes a length change of the tube
("Poisson effect'), or a stress from resisting that change; (3) The tube
temperature varies along its length and is different from the lengthwise
temperature distribution of the secondary shell. This causes differential
expansion of the two; (4) the tubes have a higher coefficient of thermal
expansion than the secondary shell which causes a differential expansion;
(5) tubesheet bowing, created by primary and secondary pressures
combined with induced shell and head deflection loads; and (6) the tube
preload introduced during manufacturing.
In the case of the once-through steam generator, tube buckling is
caused by deformation controlled loads and Thus Is not a catastrophic
primary stress failure mode. However, analysis of a tube shows that tube
touching would occur very soon after the tube assumed a bowed shape.
2
CA 02288367 1999-11-02
Case 6075
Therefore, the load Which causes tubes to touch is considered as the limit
load on the tube in compression.
A sight manufacturing tube prestrain of about 1l8 inch over the
length of the tube is considered beneficial to reduce compressive loads on
the tubes under all operating conditions. This has the added benefit of
preventing stress softEning and the resultant reduction in tube natural
frequency for flow induced vibration considerations.
1o In view of the foregoing it is seen that an improved method of
prestraining the tubes Qf a Qnce-thro4gh steam generator was needed
which would not sNbject the tubes to interaction stresses during welding.
~ 5 BRIEF SUMMARY OF THE IN~JENTION
The present invention solves the prior art once-through steam
generator or heat exchanger assembly problems and other problems by
providing a method of prestressing the once-through steam generator
2 o tubes in which the tubes of the once=through steam generator are
prestrained to the desired level using the hydraulic expansion of the tubes
in the tubesheet. Prior to performing the tube joint hydraulic expansion,
path ends of the tubes are welded to their respective tubesheet_ The
subsequent tube radial expansion within the hydraulic expansion zone
2 s creates the desired axial preload_ It has been demonstrated both
analytically and experimentally that tensile stresses are developed during
the hydraulic expansion of tubes which are restrained at both ends_ The
obtained stresses are of the desired magnitude to increase margin to
buciding and increase tube natural frequency to thus increase the margin
3o to detrimental flow induced vibration.
Optimized selection of the final total tube stress is controlled by
controlling the length of hydraulic expansion in the upper tubesheet. The
main advantage is that the developed prestrain is independent of the tube
3
CA 02288367 1999-11-02
case sops
load prior to the expansion. Therefore, the achieved pre-set of a given
tube will be independent of the state of the other tubes resulting in the
desired uniform foreshortening of each tube.
A tensile prestrain of 1l8 inch over the tube length will assure that
all cubing stress limits will be met and the tubes will be at a very low
tensile stress of approximately 3 ksi during full power operation. This
tensile presirain is achieved by controlling the hydraulic expansion
process.
~o
1n view of the foregoing it will be seen that one aspect of the
present invention is to provide a prestraining method for onc~through
steam ganerxater tube assemblies which will be constant for all the
individua~ tubes.
Another aspect ~s to provide a prestraining method for once-
through steam generator tubes once they are assembled into a once-
through steam generator tube array.
z o These and other aspects of the present invention will be more fully
understood from the following description of the invention when
considered along with the accompanying drawings.
BRIEF DESCRiPT~~N OF TIDE iDRAIMII~GS
Figure 1 is a vertical elevation view in full section of a once-through
steam generator embodying the principles of the invention;
3o Figure 2 is a perspective view of a test stand used to develop once
through steam generator tubing hydraulic stress verification data.
Figure 3 is an end view of the Figure 2 test stand.
4
CA 02288367 1999-11-02
Case 8075
Figure 4 is an expanded view of strain gauge location on two tubes
of the Figure 2 test stand
Figure 5 depicts test results of a full tube length expansion strain.
Figure 6 depicts test results of half tube length expansion strain.
1 o DETAtI-Ep DESCRIPTION OF THE PRE~E~iREp EMBODIMENT
The present invention is described in connection with a once-
thro~rgh atGam generator for a nuclear power plant, although these
principles are generally applicable to shell and tube heat exchangers in
any number of diverse fields of activities. Thus, as shown in Figure 1 for
the purpose of illustration, a once-through steam generator unit 10
comprising a vertically elongated cylindrical pressure vessel or shell 11
closed at its opposite ends by an upper head member 12 and a lower
head member 13.
The upper head include$ an upper tubesheet 14, a primary coolant
inlet 15, a manway 16 and a handhole 17. The manway 16 and the
handhale 17 are used for inspection and repair during times when the
once-through steam generator unit 10 is not in operation. The lower head
13 includes a drain 18, a coolant outlet 20, a handhole 21, a manway 22
and a lower tubesheet 23.
The once-through steam generator 10 is supported on a conical or
cylindrical skirt 24 which engages the outer surface of the lower head 13
3 o in order to support the generator unit 10 above structural flooring 25.
As hereinbefore mentioned, the overall length of a typical once-
through generator unit of the sort under consideration is about 75 feet
between the flooring 25 and the upper extreme end of the primary coolant
5
CA 02288367 1999-11-02
Case 6075
inlet 15. The overall diameter of the unit 10 moreover, is in excess of 12
feet.
Within the pressure vessel ~ 1, a lower cylindrical tube shroud
wrapper or baffle 26 encloses a bundle of heat exchanger tubes 27, a
portion of which is shown illustratively in Figure 1 In a once-through
steam generator unit of the type under consideration moreover, the
number of tubes 27 enclosed within the baffle 26 is in excess of 15,000,
each of the cubes 27 having an outside diameter (OD) of 5/8 inch. It has
1 o been found that Alloy 690 is a preferred tube material for use in once-
through steam generators of the type described. The individual tubes 27
in the bundle each are anchored in respective holes formed in the upper
and lower tubesheets 14 and 23 through seal welding the tube ends at the
tubesheets. To support the tubes 27 in their proper positions, and array
of drilled and broached substantially flat support plates 45 is positioned
transverse to the longitudinal axes of the tubes 27 and the axes of the
pressure vessel 11.
The lower baffle or wrapper 26 is aligned within the pressure vessel
2 0 11 by means of pins (not shown). The lower baffle 26 is secured by bolts
(not shown) to the tower tubesheet 23 or by welding to lugs (not shown)
projecting from the lower end of the pressure vessel 11. The lower edge
of the baffle 26 has a group of rectangular water ports 30 or, alternatively,
a single full circumferential opening (not shown) to accommodate the inlet
2 5 feedwater flow to the riser chamber 19. The upper end of the baffle 26
also establishes fluid communication between the riser chamber 19 within
the baffle 26 and annular downcorner space 31 that is formed between
the outer surface of the lower baffle 26 and the inner surface of the
cylindrical pressure vessel 11 through a gap or steam bleed port 32.
A support rod system 28 is secured at the uppermost support plate
458, and consists of threaded segments spanning between the lower
tubesheet 23 and the lowest support plate 45A and thereafter between all
support plates 45 up to the uppemnost support plate 458.
6
CA 02288367 1999-11-02
Case 6075
A hollow torpid shaped secondary coolant feedwater inlet header
34 circumscribes the outer surface of the pressure vessel 11. The header
34 is in fluid communication with the annular downcomer space 31
through an array of radially disposed feedwater inlet nozzles 35. As
shown by the direction of the FIG. 1 arrows, feedwater flows from the
header 34 into the once-.through steam generating unit 10 by way of the
nozzles 35 and 36. The feedwater is discharged from the nozzles
downwardly through the annular downcomer 31 and through the water
1 o ports 30 into the riser chamber 19. Within the riser chamber 19, the
secondary coolant feedwater flows upwardly within the baffle 26 in a
direction that is counter to the downward flaw of the primary coolant within
the tubes 27. An annular plate 37, wedded between the inner s4~face of
the pressure vessel 11 and the outer surface of the bottom edge of an
upper cylindrical baffle or wrapper 33 insures that feedwater entering the
downcomet 31 will flow downwardly toward The water ports 3p in the
direction indicated by the arrows. The secondary fluid absorbs heat from
the primary fluid through the tubes 27 in the bundle and rises to steam
within the chamber 19 that is defined by the baffles 26 and 33.
The 4pper baffle 33, also aligned with the pressure vessel 11 by
means of alignment pins (not shown), is fixed in an appropriate position
because it ~s welded to the pressure vessel 11 through the plate 37,
immediately below steam outlet nozzles 40. The upper baffle 33,
furthermore, enshrouds about one third of the tube bundle.
An auxiliary feedwater header 41 is in fluid communication with the
~Jpper portion of the tube bundle through one or more nozzles 42 that
penetrate The pressure vessel 11 and the upper baffle 33. This auxiliary
3 o feedwater system is used, for example, to fill the once-through steam
generator 10 in the unlikely event that there is an interruption in the
feedwater flow from the header 34. As hereinbefore mentioned, the
feedwater, or secondary coolant that flows upwardly through the tube
bank 27 in the direction shown by the arrows rises into steam. In the
7
CA 02288367 1999-11-02
Case 6075
illustrative embodiment, moreover, this steam is superheated before it
reaches the top edge of the upper baffle 33. This superheated steam
flows in the direction shown by the arrow, over the top of the baffle 33
and downwardly through an annular outlet passageway 43 that is formed
between the outer surface of the upper cylindrical baffle 33 and the inner
surface of the pressure vessel 17 .
The steam in the passageway 43 leaves the generating unit 10 through
steam outlet nozzles 40 which are in communication with the passageway
43. in this foregoing manner, the secondary coolant is raised from the
feed water inlet temperature through to a superheated steam temperature
at the outlet nozzles 40. The annular piste 37 prevents the steam from
mixing with the incoming feedwater in the downcomer 31 _ The primary
coolant, in giving up this heat to the secondary coolant, flows from a
nuclear reactor (not shown) to the primary coolant inlet 15 in the upper
head 12, through individual tubes 27 in the heat exchanger tube bundle,
into the lower head 13 and is discharged through the outlet 20 to
complete a loop back to the nuclear reactor which generates the heat
from which useful work is ultimately extracted.
Referring now to the drawings generally and Figures 2 and 3 in
particular, it will be noted that a test stand or rig 50 has been designed to
investigate two areas of special interest. (1) Explore the effects of tube
insertion into a once-through steam generator configuration and (2)
quantify residual strain during axially constrained hydraulic expansion of
the tubes 27 alone.
The capability to accurately and analytically predict the hydraulic
expansion mechanics was confirmed using finite element modeling as will
3 o be discussed later.
Generally, for once-through steam generator tube applications, the
expansion takes place at each end of the straight tube which has both
ends fastened by seal welds at the tubesheet. An expansion in a U-bend
8
CA 02288367 1999-11-02
Case 8075
or free ended straight tupe results in contraction of the free end of the
tube. This contraction is in proportion to the length of the expansion in
accordance with the Poisson effect. For a 26 inch expansion length, this
axial movement has been observed to be approximately 1/8 inch.
expansion in a fixed ended tube induces strain to the tube instead. What
is not known is the influence of the expansion zone plastic defvmlation in
the distribution of the strain i_e. the strain in the tube could be evenly
distributed throughout the expanded and unexpanded tube or could
accumulate m the plastically flowing region. An even distribution is
1 o expected based on theoretical material mechanics; however, the
magnitude must be verified t4 be analytically predictable so that it may be
considered in the residua) stress, flow induced vibration, and tubelshell
interaction analyses of the once-through steam generator t4be design.
1 s The once-through steam generator proposed design consists of 5/8
inch OD, 0.038 inch wall Sumitomo Alloy 690TT on a 7/8 inch pitch-
Tubesheets are sized at 22 inches thick each- Fifteen tube support
probes exist over the bundle length.
2 o The test rig 50 shown in Figures 2 and 3 has tubelhole and pitch
geometries Selected based on availability of equipment. The stand 50
includes two tubesheet blocks 51 and 52 gun drilled to a 0.93 inch
triangular tube pitch, and two broached plates 53 and 54. The tubes 55,
56 and 57 are 11/16 inch OD, 0.040 inch wall Sumitomo Alloy 690 TT. The
2 5 broached plates 53 and 54 are of similar material and tube-to-hole
clearance as the once-through steam generator broached plate 45, and
holes 58 drilled to a 0.95 inch triangular tube pitch. The edge condition of
the hole 58 is much rougher than the broached plate 45 hole to provide a
conservative condition for tube aprasion assessment. The pitch of the
3 o drilled broached plates 53 and 54 for the test arrangement was larger than
the pitch of the holes 59 of the tube sheet blocks 12 and 14. As such one
central hole 59 was used to align the tube passage by typical production
techniques with the surrounding holes 59 being progressively further out
of alignment- The test holes were the aligned hole and an adjacent hole
9
CA 02288367 1999-11-02
Case 6075
that represents a 0.020 inch offset of the tube passage. A third hole in the
periphery of the pattern was used to assess tubeability and entry abrasion
for a conservative out-of tolerance offset . I_e., '0.050 inch misalignrnent_
The assembly was mounted between heavy structural beams 60
and 61 to approximate the rigidity of the tubesheet/pressure boundary
assembly, and offer stiffness in excess of the tubes 55, 56 and 57 being
investigated.
io
The tubes 55, 56 and 57 were eddy current inspected for
manufacturing burnish mark in full accordance with accepted testing
procedures before insertion into the test rig 50.
The tube 55 was then inserted by normal practice into
corresponding ideally aligned holes 58 and 59 while tube 56 was inserted
into the holes 58 and 59 having a 0.020 inch displacement between the
corresponding broached plate holes 5B and the tubesheet block holes 59.
A third tube 57 was inserted into a peripheral hole 59 wish ~-0.050 inch
2 0 offset relative to the corresponding Groached plate holes 58. The ends of
the inserted tubes 55, 56 and 57 were tack expanded at each end in
preparation for welding.
Ten strain gauges 62 were mounted and equally spaced across the
2 5 free span of tube 55 and tube 56 as shown at Figure 4. The tubes 55 and
56 were welded at their respective ends to t4besheet blocks 51 and 52.
Each of the gauges B2 was used to measure tube expansion in an axial or
transverse direction at its respective location, and strain data was then
recorded to assess any imparted strains from the welding operation.
3D Thereafter, Tube 55 was fuh depth hydraulically expanded, i.e_, 26 9/16
inch length, while the digital data acquisition system recorded the resulting
strain development in the tube 55. After a cursory data review, the partial
depth hydraulic expansion, i_e_, 13.25 inch length at the second tubesheet
block 52 was performed with strain data recorded. The process was then
CA 02288367 1999-11-02
Case 6D75
repeated for tube 55 at the full and partial depth hydraulic expansion as
shown at Figures 5 and 6, respectively.
Eddy current evaluation was performed on the two completely
assembledlinstrumented tubes 55 and 56 and the tube 57 which had
been inserted into the 0.050 inch offset tube holes 58 and 59.
After two weeks, tube 55 was cut at a point ~8 inch from the first,
i.e.. full depth hydraulically expanded, tubesheet block 51 and the strain
1 o relaxation was measured with a dial gauge.
A linear elastic plastic finite element model of the experimental test
stand 50 was developed to provide comparative analytical predictions of
the strain development with hydraulic expansion. The model was an
7.5 adaptation of the 3-D axisymmetric hydraulic expansion model developed
for in-house Tube to Tubesheet Joint Qualification Programs_
Tubes 55 and 56 were inserted through exactly aligned and .020
inch offset corresponding 'holes 58 and 59 and passed through both
2 o tut~esheet holes 59 and broached plate holes 58 with no substantial
resistance relative to U-tube steam generator tubing experience. Mild
resistant 'stop and starts' were encountered by the tubes 55 and 56 as
they were passed through the first tubeshset hole 59 which is a typical
response to 'eyeball' estimation of a perpendicular entry alignment of the
2 5 tube at the peginning of its insertion. F-~cperience has shown that once
the
tube is sufficiently inserted. it guides itself and resistance is virtually
eliminated.
Slight resistance was sensed on insertion of the t4be 57 through
3 o the corresponding 0.050 inch offset holes 58 and 59. However, this was
well within the range of experience with pressurized once-through steam
generator tube insertion. Manual thumb pressure on the tube end was
sufficient to smoothly move the tube 57 into position.
11
CA 02288367 1999-11-02
Case 6075
10
The eddy current evaluation showed no manufacturing burnish
mark in any of the tubes before or after insertion and expansion, when
subjected to eddy current test processes and criterion suitable to typical
Baseline evaluations.
Visual evaluation indicated superficial discontinuity in the tube
Surface finish on tubes 55 and 56. A mild burnish was visible on tube 57
which had no raised metal or discernible depth relative to calibrated
scratch standards i.e_, <0.0005 inch.
NQ strains in the tube free span were detectable from the welding
operation. In response to hydraulic expansion, the strain gauges showed
strain desvelopment in the free span of both tubes SS and 56 that was
uniform and consistent throughout. Finite element prediction of the
expansion strain was shown as very close to that experimentally
measured on tube 56 as shown in Figure 5. The strain levels as
experimentally measured on tube 56 increased again by an expected 50%
in response to the second tubesheet expansion of half the length of the
original as seen in Figure fi.
2S
After two weeks, tube 55 was cut and the strain relaxation was
measured with a dial gaNge. This was intended to investigate any
unknown relaxation effects. The tube relaxed over 0.130 inch which is
comparable to prediction by the finite element model-
The development of the strains to predicted levels shows that the
plastic expansion regions have not absorbed a disproportional share of
the strain due to unexpected non-linearities. The strain is uniformly
distributed throughout the expanded and free span regions. It is
3o repeatable, analytically predictable, controllable (by setting expansion
length) and permanent under the conditions tested in this experiment.
No manufacturing burnish marks were detected by eddy current
testing in spite of ahempts to create a worst case tube passage.
12
CA 02288367 1999-11-02
Case 6075
From the foregoing test results it is seen that hyd~a4lic pre-
stressing of once-through steam generator tubes is possible when done
according to the developed empirical data.
It will be understood that certain modifications and improvements
obvious to people of ordinary skill in this art area were deleted herein for
the sake of conciseness and readability. It is intended, however that all
such be included in the scope of the following claims-
to
13
