CA1210081A - Method for diffusion welding - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A method is disclosed for joining metal parts by diffusion welding, which comprises initially the establish-ing of a joint between opposed bounding surfaces of the respective metal parts, the joint having generally increas-ing height from the periphery towards the middle thereof.
The joint is provided with a connection to an external source for flushing gas and to a vacuum source. During preliminary heating and pressing together of the metal parts, the joint is supplied with an inert or reducing flushing gas, e.g. hydrogen, until the joint is closed by diffusion welding along its periphery. Thereafter, the joint may be evacuated, whereupon further heating, quick compression and a certain holding time provide com-plete diffusion welding of the metal parts. It is also suggested to reduce the cross section of the metal parts in the joint area and press these out to their original cross section during the diffusion welding.

Description

~l A METHOD FOR DIFFUSION WELDII~G.

The present invention relate3 to a method or joining parts of metal by means of diffusion welding, comprising the following steps: establishing a joint between opposed bounding surfaces on the respective parts to be joined, heating of the parts to a predetermined temperature in at least those areas bordering on the joint, joining the pa~ts by pressing the parts together, and cooling the parts.

Advantages and disadvantages of diffusion welding are known i.a. from Welding Handbook, 7. edition, volume III, page~
312 et sec.

Among the advantages can be mentioned that joints can be formed having properties and microstructures which are very similar to those of the base material and joining is possible where the form of the work p:ieces makes it difficult to use other methods. Furthermore, join-ts can be made having mini~al 2Q deformations and without con,ecutive machining or forming.

~mong the disadvantages can be mentioned that the cost of the equipment is usually verv high, a fact -that limit the size of the parts that may be diffusion welded in an economic manner. Furthermore, the necessity of supplying heat and a high compressive force in an environment such as vacuum or ~rotective atmosphere will g ve rise to grave problems as regards the necessary equipment and t:herefore constitute a substantial limitation in the applicability of the method.
In addition, one has assumed that the method requires very careful and thorough surface preparation and, besides, that it requires more time than conventional me~hods.

Nevertheless, it is an object of the present invention to adapt a method of the above mentioned type for making a pipe-line for offshore transportat:ion of gas and oil. Making of such pipelines today is done by welding together pipe sections manually or by CO2 automatic equipment on board a laying vessel. Usually the welding of the pipes occurs at several 8~

stations concurrently in order to incxease the laying speed.
Due to practical reasons the stations are placed on a hori-zontal line ana several welders can weld on each station concurrently. The finished pipe is let ou-t behind the laying vessel over a so-called "stinger" which prevents the pipe from breaking directly behind the laying vessel and runs in a S-shaped curve down towards the sea floor. A
certain tension is maintained in the laid pipeline, and for this purpose the laying vessel must have anchors by means of which it is pulling itself forward during the laying of the pipeline. These anchors must from time to time be moved forward in front of the vessel and for this purpose it is necessary to use auxiliary vessels~

For various reasons it would be desirable to be able to run the pipeline generally vertically downwards from the laying vessel, so that the pipeline would run in a single hend down towards the sea floor. This would i.a. entail better control of the stresses in the laid pipeline, and one could avoid anchors and auxiliary vessels and instead make do with dynamic positioning of the laying vessel. With the pipeline vertically oriented on board the laying vessel, however, it would be very difficult to use more than one welding station.
With the present slow welding procedures this would therefore not give sufficient laying speed to be of practical interest.

The present invention enables two pipe sections to be joined so quickly that sufficient laying speed can be obtained in order to use vertical orientation of the pipeline on board a layi~g vessel, thus enjoying the advantages this will entail.

Thus, the invention provides a method for joining elongate metal parts of generally constant cross-section by diffusion welding, comprising the steps ofo (a) establishing a joint between opposed bounding surfaces on the respective parts to be joined, which parts in the joint area initially are reduced in cross section, so that the joint initially is formed with a cavity with generally ~2~81 increasing height from the periphery to~ards the middle and is provided with at least one connecting conduit, (b) heating the parts to a predetermined temperature at least in those areas bordering on said joint, the cavity of the joint during the heating being supplied with reducing flushing gas until the joint is tightly closed due to diffusion welding along the periphery, (c) joining the parts by pressing them together with a pre-determined speed for the further diffusion welding and for restoring said reduced cross-section to substantially said generally constant cross-section of the parts, and (d) thereafter coolinq the parts.

By means of the stated form of the joint one obtains high surface pressure along the periphery of the joint at the beginning of the method. Therefore, diffusion can start along the periphery at such low temperatures that oxidation of the surfaces are avoided. Flushing with inert gas pre-vents access of oxygen so that oxidation is prevented. If a reducing gas is used, one can remove oxides which previously may have formed on the surfaces of the joint~ After closing of the joint along the periphery in this way, which will manifest itself in that flushing gas no longer escapes, one may connect the connecting conduit to a vacuum source so that the joint is put under vacuum. This will remove resi-dual gases and make oxides and other impurities dissociate to a certain degree upon increasing temperature. By adapt-ing temperature and pressure the joint is concurrently closed ~Z~

- 3a -quickly, whereupon complete diffusion welding can take place in a short time. Quick compression is advantageous because it produces higher pressure in the joint area, which is due -to the creep stress of the material increasing wi-th the speed 5 of deformation.

According to an advantageous feature of the invention, at least one of the bounding surfaces of the joint is provided with serrations along the periphery. This will ensure 10 sufficient escape of flushing gas during the early phases of the joining.

Furthermore, it is very advantageous to provide at least one of the bounding sur~aces of the joint with a concave form.
15 Thereby an advantageous stress distribution is obtained during the final compression of -the parts, with resulting assurance of complete fusion along the entire joint.
I

~2~

A further advantageous form~of the joint is obtained if ajt least one, preferably both bounding surfaces are given conical ~orm. The conical form is easier to machine in a sufficiently exact manner, and if both surfaces are conical, one convex and the other concave, one will obtain a self-allgning effec~
between the parts when they are pressed together. Such a form will also provide good assurance of complete fusion.

In order to increase the con act pressure in the joint area during compression, it is suqgested according to the invention to reduce the cross section of the parts in the joint area.
Thereby a pressure in the joint area is obtained which is several times higher than the uniaxial creep stress of the material at the prevailing temperature, the result being quicker and better diffusion.

Several different inert flushing gases may be used, but helium is preferred because ~1uring its manufacture it is cooled down to such a temperature that it is virtually guaranteed to be oxygen free. Besides commercially available and reliable helium detectors can be used to co~trol that the periphery of the joint is absolutely closed before the vacuum is applied.

If one instead uses a reducing gas for flushing, hydrogen will be very advantageous. I~lot only will it remove any oxide layers but it will also make it possible to observe visually if the joint is clo!;ed because any leaking hydrogen will burn on the surface. Hydrogen will not have any harmful effect on the relatively soft structural steels to be joined by the method.

According to the invention it: is also proposed to provide the joint with an activating alloy, i.a. 60/40 Palladium/Nickel.
This alloy may be introduced as a thin strip or possibly be applied on one or both bounding surfaces by means of electro-plating. This will i.a. result in reduced diffusion time~

The connecting conduit for supplying flushing gas or removing ~2~Q~

gas for pulling a vacuum, may advantageously be arranged in one of the parts ~o be joined, preferably near its periphery.
Thereby the conduit may be plugged immediately after the diffusion welding by means of a deep welding electrode ox Plasma-TIG while the temperature oE the parts still is sufficiently highl for instance 400 ~ 600C.

The invention shall be furthPr explained wi~h reference to the exemplifying embodiments shown in the accompanying drawings.
Figures l and 2 show a first embodiment of two parts at beginning and completed diEfusion welding, respectively.

Figures 3 and 4 show a second embodiment of the two parts at beginning and completed difusion welding, respectlvely.

Figures S and 6 show a further exemplifying embodiment of two parts at beginning and completed diffusion welding, respeotively.
Figure 7 shows the parts of Figure 6 upon further compression.

Corresponding parts are given the same reference numerals in all the figures.

Figure l shows a section through a por-tion of two parts l and

2, which may be bolts of structural steel such as ST 52-3.
Figure l may also be visualized as showing an axial section through a thick-walled pipe, the axis of the pipe being at the left of the figure. Between the parts l and 2 a joint 3 is formed, which is bounded by a concavely curved bounding surface 4 on the part l and a plane bounding surface 5 on the part 2. A connec~ing conduit 6 is arranged in the part l, said conduit opening into the cavity 3 of the joint and with its other end being connectably alternately with a source of flushing gas and a vacuum source ( not shown)O The arrows F indicate a variable compressing force, while the bell-shaped curve to the left of the parts l and ~ indicates the axial temperature distribution in the parts.

When the parts are to be joined, they are first brought in the position shown in Figure l and subjected to a suitable com-pressing force F. This force may be created by means o~ asimple jack system (not shown), which for instance may comprise a clamping ring on each of the parts l, 2, the clamping rings being interconnected by means of hydraulic cylinders. O-ther forms may of course be envisioned, depending upon the form and cross section of the parts. Thereupon the heating is started, for instance by means of an induction coil (not shown). Concurrently,flushing gas is supplied through the connecting conduit 6. The flushing gas will first leak out along the periphery 7 of the joint due to minor `15 irregularities or serrating on the bounding surfaces of the joint. The purpose of the flushing is to keep oxygen away from the joint surfa~es in order to prevent oxidation of these during the heating, possibly also to remove oxides that may already be present. Regardless of how well the bounding surfaces 4, 5 have been cleaned before the welding, even a short exposure to the oxygen of the air will cause an o~ide layer of a thickness in the ~rder o~ 350 - lO00 ~, depending on the air temperature and humidity.

When the temperature of the material near the periphery 7 of the joint has reached 600 - 800C, for instance with a temperature distribution as shown schematicly to the lef~ in Figure l, diffusion bet~een the parts l, 2 will occur under a suitable compressing force F so that the joint 3 is closed along the periphPry 7. This may be registered in several ways, or instance in that g~s no longer leaks out of the joint, or in that a pressure imposed on the cavit~ 3 o~ the joint no longer decreases.

When the joint thus is closed along its periphery 7, the connecting conduit 6 is connQcted to a vacuum source, which thereupon reduces the pressure in the cavity 3 to about torr. Concurrently,the temperature of the parts l, 2 is increased as suggested schematicly by the curve to the ~Zl~Q8~

left in Figure 2, to a maximum temperature of about 1350C.
With a suitable compressing iorce F the joint 3 will be closed in a ma-tter of seconds. This result is schematicly shown in Figure 2. Colnplete diffusion welding will take place in the course of 15 - 30 minu-tes. ~However, one has obtained complete welding in diffusion times as short as 8 minutes at about 1350C.) Thereafter the parts are cooled in calm air down to about 60~)C, which will take about ~
minutes for a material thickness of ~0 mm. ~t this temperature the connecting conduit may be plugged,i.a. by means of a deep weldiny electrode or Plasma-TI~ welding.

Figure 3 shows two parts l and 2 with generally the same outer form as in Figure l. ]lowever, the welding joint 3 here has a different form, its bounding surfaces 4 and 5 both being conical, one concave and the other convex. The top angles are different so that the joint will have increasing thickness towards the middle. The distance between the apices of the joint surfaces can amount to about 10% of the thickness f the parts 1, 2~ The conical form of the bounding surfaces 4, 5 makes the parts l, 2 self-aligning when they are pressed together. The conical form also helps to avoid insufficient diffusion in the middle of the joint, as easily could happen if the lower bounding surface S was plane and the upper bounding surface ~ conicc~lly concave.

Otherwise, joining of the parts l, 2 takes place in the same manner as indicated with respect to Figures `l and 2. The final result is schematicly shown in Figure 4.
The pressure during the diffusion welding is a very important parameter which is difficult to control. According to the Welding Handbook previously mentioned it is supposed that the pressure cannot be brouyht higher than the uniaxial creep stress of the material at the temperature prevailing at any time without using forms to prevent floatingO The use of such forms is cumbersome and entails increased cost, and in some applications such forms cannot be used due to lack of access. Since the diffusion speed generally increases ~z~

with the square of the pressure, by increasing the pressure to the double or triple of the creep stress, one could bring the diffusion time down to about one four-th or one ninth, respectively, of the corresponding diffusion -time for uniaxial pressure. Alternatively one could permit considerably more of oxides on the joint surfaces and thereby possibly delete flushing of the joint with a reducing gas.

In order to obtain such high pressures, the invention aims a-t establishing a triaxial stress condition in the joint area.
It will be known that in tensile testing of thicker test rods a contraction will OCCUL- by shear deformation, a cleavage fracture finally taking place in the middle ~one. This cleavage fracture is caused ~y the axial tension in the middle being very much higher than 1he yield stress of the material, the material on both sides o the contraction holding back by means of radial stresses. For this reason the axial stress increases, the yieldin~ subs:idin~ when the difference between the axial stress and the rad:ial stress becomes less than the yield stress (Trescas principle).

According to the invention one aims at utilizing the triaxial phenomenon by providing the parts l, 2 with a substantial constriction in the joint area as compaxed to the adjacent material. An example is shown in Figure 5 In this example hydrogen is used as flushing gas, as is evidenced by the flames indicated along the periphexy of the ~oint.

When the flames disappear one knows that the joint is closed, i.e. one does not have to use any form of detection equipment in this connection. When the joint has ~een reduced and closed along the edge, application of the full compressive force will make the joint close under a very complicated stress picture of dynamic character. The triaxial stress condition exactly at the closing moment ~ill, due to the constriction and the relatively cold surrounding material, give stresses transversely of the joint which are 5-6 times the creep stress for equally thick parts. This situation is illustrated in Figure 6. Next, the parts are pressed ~l2~8~
. g together wlth a relatively mc,dest force until they have obtained generally the same cross section over the entire length, as is suggested in Figure 7.

Thus, one has obtained the correct shape in the welded parts without the use of an external form, a very high contact pressure in the deciding phase, and a distribution of oxides tnat might have been present to double the area or more, with a corresponding thinning of the oxide layer a~d increasing globular conversion, including accelerated recrystallization.

When -the desired thickness has been reached, the compressive force may be removed and the diffusion time and temperature be held until the residual pores have been closed by re-crystallization. If a much quicker joining is desired, this can be obtained by electroplating the joint surfaces or introducing in the joint a st:rip of activating alloy, for instance of the type Pd/Ni 60/qO.
Due to the very high stresses occuring in the joint area when it is cons-tricted as shown in Figure 5l it is supposecl that the original form of the joint and the pre-finishing of the join-t surfaces are less critical.
Tests performed ~ith the method according to the invention by using bolts of structural steel ST 52-3 having 40 ~n diameter have given excellent: results with respect to for instance ductility and tensile strength. More than 60%
elongation has been measured in tha heat effected zone for bolts which have been bent to hair needle shape. The rupture surface showed essentially shear fracture.

It will be understood that the invention may also be advantageously utilized for other purposes than fabrication of large pipelines. For inst:ance,it may be used in welding joints in trusses for bridges and offshore structures.

It is mentioned above that serration may be used on one or ~l2~
both of the bounding surface, of the joint along their periphery in order to ascertain discharge of flushing gas during the preliminary phases of the welding. In certain cases it may be advantageous to use a slight serration on -the entire bounding surface. Such serration will give an echo during later ultrasonic testing if the joint is not completely fusioned.

Claims (7)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method for joining elongate metal parts of generally constant cross-section by diffusion welding, com-prising the steps of:
(a) establishing a joint between opposed bounding surfaces on the respective parts to be joined, which parts in the joint area initially are reduced in cross-section, so that the joint initially is formed with a cavity with generally increasing height from the periphery towards the middle and is provided with at least one connecting conduit, (b) heating the parts to a predetermined temperature at least in those areas bordering on said joint, the cavity of the joint during the heating being supplied with reducing flushing gas until the joint is tightly closed due to diffusion welding along the periphery, (c) joining the parts by pressing them together with a pre-determined speed for the further diffusion welding and for restoring said reduced cross-section to substantially said generally constant cross section of the parts, and (d) thereafter cooling the parts.

2. A method according to claim 1, wherein at least one of the bounding surfaces of the joint is of concave form.

3. A method according to claim 1, wherein at least one of the bounding surfaces of the joint is of conical form.

4. A method according to claim 3, wherein the bounding surfaces of the joint are both of conical form, one concave and the other convex.

5. A method according to claim 1, wherein hydrogen is used as reducing gas.

6. A method according to claim 1, wherein the connect-ing conduit is disposed in one of said parts.

7. A method according to claim 6, wherein the connect-ing conduit is disposed near the periphery of said part.