GB2453945A - Apparatus for Additive Manufacture Welding - Google Patents

Abstract

Additive manufacture apparatus comprising a heat source 14 for forming a melt pool in a substrate 18, a media supply 16 for supplying media 24 to the melt pool and at least one further heat source 26 for applying an adjustable heat flux power or distribution to the substrate such that layers of media 24 are added to the substrate 18. In an embodiment, the further heat source 26 can be array of fibre optic cables connected to a laser and arranged around the heat source 14 (figure 7).

Description

METHOD AND APPARATUS FOR WELDING

This invention relates to methods and apparatus for welding and in particular methods and apparatus for welding or cladding when used as an additive manufacturing process.

Welding may be used as a process beneficial to join components or as an additive process, which may be used to build structures on a substrate.

It is important when welding, and particularly when welding to build a structure on a substrate that the microstructure is uniform with the absence of voids and cracks which may affect the integrity of the structure. It is also desirable to make more efficient use of material through a more consistent material performance.

It is an object of the present invention to seek to provide improved welding methods and apparatus.

According to a first aspect of the invention there is provided additive manufacture apparatus comprising: a heat source for forming a melt pooi in a substrate, a media supply for supplying media to the melt pool, and at least one further heat source for applying an adjustable heat flux power or distribution to the substrate.

Preferably the at least one further heat source is adapted to affect at least one parameter influencing the solidification metallurgy of media supplied by the media supply and melted in the melt pool.

Preferably the at least one further heat source comprises a plurality of fibre optic cables. Each cable may be functionally connected to a laser. Preferably each cable is functionally connected to a respective laser. The or each laser may be a Nd:YAG laser.

Preferably each laser is individually controllable.

The further heat source may be arranged around the heat source in a rose.

The apparatus may further comprise sensing means for detecting the at least one parameter and a control system for continuously controlling the heat flux power or distribution whilst media is applied to the substrate.

Embodiments of the invention will now be described by way of example only, with reference to the accompanying drawings, in which: Fig. 1 is a diagram of a direct laser deposition apparatus.

Figs. 2a to 2c depict a weld bead deposited to a substrate in a series of geometrical arrangements Fig. 3 depicts one possible arrangement for controlling heat flux distribution.

Fig. 4 depicts an alternative arrangement for controlling heat flux distribution.

Fig. 5 depicts a modified flux distribution required to control cooling rates for the geometrical arrangements of Fig. 2 Fig. 6 depicts a path deposition order which minimises uneven geometrical arrangements.

Fig. 7 depicts a preferred head arrangement used to control flux distribution.

Fig. 8 diagrammatically shows a preferred fibre optic arrangement within the laser deposition head of Figure 7.

Fig. 9 diagrammatically shows an alternative fibre optic arrangement within the laser deposition head of Figure 7.

Referring to figure 1, the prior art apparatus

comprises a substrate 10 mounted to a table 12, moveable relative to a laser 14 and a powder delivery nozzle 16. The method of forming a structure 18 comprises directing a beam from the laser 14 onto the substrate 10 or later the forming structure 18, to create a pool of molten metal 22 into which a powder 24 is directed as a jet. Once sufficient powder has been deposited a relatively thin layer of metal remains. The substrate 10 and forming structure 18 are translated so that the structure is formed in layer-wise manner. The process allows a near net material direct manufacture of structures.

By controlling the amount of powder and the location of the base material simple and complex structures may be formed. For gas turbine engine blades and the like, one advantage of this process is that complex aerofoil shapes can be manufactured directly from a computer aided design model without the need for traditional core, wax and shell process steps. It is an essential part of this process that the laser, delivery of powder jet and location of the deposit are computer 32 controlled.

In EP 1637274, where the deposit is controlled to form a single crystal, a relatively crude temperature control may be provided e.g. by forming the deposit within an oven or directing a laser 26 or other heat source at the surrounding substrate. This, along with the provision of cooling jets 28, 30, controls the cooling rate of the deposit such that the single crystal can grow into the substrate.

This crude thermal control does not consider geometrical effects that affect the freezing rate of the deposit. Heat from the deposit flows down thermal gradients and the available cross-section of the deposit determines the rate of transfer.

For example, Figures 2a, 2b and 2c are simplified diagrams of a first second and third deposition scenario. In the first deposition scenario of Fig. 2a the weld bead 22 is deposited directly onto the substrate 12. Since the melt pool extends a small distance, typically 50 to 100pm, there is a large area surrounding the bead 22 for heat transfer with the area being relatively cool. The temperature difference and large area allows for a fast freezing rate.

In the second deposition scenario of Fig. 2b, the weld bead is laid directly upon an earlier and cooled weld bead. The available area immediately surrounding the freshly deposited bead 22 is significantly reduced limiting the area for the conduction of heat from the weld and setting a slower freezing rate.

In the third deposition scenario of Fig. 2c, the weld bead 22 is laid adjacent to an earlier pass. The earlier pass still maintains a residual temperature that limits the flow of heat from the fresh bead 22 and further reduces the rate of cooling. Additionally, the flow from the bead is asymmetric.

Other scenarios will also have different freezing rates. For example a bead deposited at the edge of the substrate.

During deposition, unless the rate of deposition is particularly slow or there is an extremely effective heat sink, there is a progressive build up of heat in the substrate or emerging component. Higher temperatures promote wetting between beads and inter-bead fusion. The disparity caused between deposition at areas of high temperature and areas of low temperature affects the overall quality of the formed component.

The result of a variation in freezing rate affects the solidification microstructure of the weld and manifests itself as fine variations in the degrees of inter dendritic segregation or differences in the microstructural scale or feature size across a component amongst other things.

Inter dendritic segregation primarily affects nickel based superalloys and is detrimental as the segregation effect is linked to loss of hot ductility, giving rise to a susceptibility of cracking at high temperatures, and planes of weakness in the interdendritic precipitate planes. The planes of weakness are detrimental from the perspective of the consistency of material properties and are particularly significant in larger deposition.

To use a material in more arduous service conditions it is necessary to be confident in its properties and achieve a close consistency across the structure. The microstructure is primarily determined by the input consumable, the deposition processing regime, the freezing rate and any subsequent processing, particularly heat treatment.

The pool cross-sectional profile can create variation in the freezing rate of the deposit, the pool wetting angle, melt viscosity and mixing by marangoni flow, which is internal convection driven by temperature or density differences within the pool.

By adjusting the heat flux distribution generated by a heat source it is possible to control the solidification or freezing metallurgy of the deposit and the temperature distribution of the pool.

In one embodiment as shown in Figure 3, the beam from the laser 14 used to form the melt pool is manipulated by tilting it off-axis relative to the component, such as to position 14' . This is not a preferred embodiment as it creates additional difficulties as the tilting adjusts the focal length of the beam and can adjust the formation of the melt pool ahead of the deposition head depending on the area of the melt pool which requires an additional heat input.

As an alternative, the beam typically approximates to a Gaussian intensity profile. By skewing the intensity profile through optical manipulation 34 the temperature flux distribution may be altered.

In the preferred embodiments a secondary heat source is provided that has a heat distribution that is capable of being modified to achieve a desired heat flux to the deposition and deposition area.

Such an embodiment is depicted in Figure 4 the secondary heat source is provided by a laser which may be moved about the melt pool to alter the heat flux distribution. The secondary laser 26 does not need to have a flux density so high as to form a melt pool and is consequently of lower power or defocused when compared with the main beam.

The laser may be directed to the substrate using one or more mirrors (not shown) whose profiles or position may be altered using techniques known in the art.

The secondary heat source may also be provided by a plurality of lasers which have independently controllable direction or power. An alternative method is to use optical fibres directed to the substrate via a mirror and focussed at a diffractive optic which tailors the beam profile.

To achieve the best operation it is desirable to thermally model the deposition process in advance. A control system can then adjust the thermal distribution based on timing or position rather than requiring a thermal sensor close to the melt pool.

Figures 5a to 5c show exemplary temperature flux distributions, in accordance with the invention, for the deposition scenarios of Figure 2a to 2c.

For the example shown in Figure 5a, which is a bead or spot 22 deposited directly onto the substrate 12, the temperature flux distribution (indicated by arrows 40) supplied to the bead 22 is desirably of a pattern to maintain a cooling rate across the bead within a controlled range during the freezing period.

The actual intensity of the flux depends on the overall thermal requirement when the deposition of the structure is modelled. For example, it may be necessary to add more heat to ensure that there is a uniform freezing rate for the whole component, not just where complex geometry interactions affect freezing.

For the example shown in Figure 5b, where a bead or spot is deposited onto a previously formed spot or bead. In this embodiment the heat transfer through to the substrate is significantly less primarily due to the reduced contact area, though the thermal conductivity of the deposit 18 may also be less than that of the substrate 12.

Since the thermal transfer away from the spot or bead 22 is lower the freezing rate is also slower and less heat input 40 is required to achieve a constant freeze rate when compared with the embodiment of Fig. 6a. This is exemplified by the smaller arrows 40 The heat flux distribution should be even and uniform about the heat transfer paths that lead away from the deposited bead 22 to ensure even cooling across the whole bead. There is some heat transfer to the surrounding atmosphere through convection and radiation, but this is insignificant when compared with the heat transfer through conduction.

In the third embodiment of Figure 5c, there is a non-uniform heat transfer from the deposited bead or weld. The earlier deposited bead 22' affects the heat transfer from the deposit.

As seen in Figure 5c, there is greater transfer of heat from the side of the bead 22 adjacent to the colder bead 22' because of the greater surface area than from the opposing side where there is less surface area in contact with the thermally conductive substrate or structure. The side of the bead 22 adjacent the colder bead 22' therefore freezes faster than the other side of the bead leading to variation in the microstructure of bead 22.

To maintain a uniform freeze rate across the bead 22 heat is supplied to the bead in an uneven distribution. More heat (as exemplified by the size of the arrows 40) is supplied to the side adjacent the colder bead 22' than the opposing side to reduce the rate at which it cools relative to the opposing side to ensure a matched freeze rate for both sides and hence the microstructure across the bead.

It is desirable that the freeze rate of bead 22 in all these scenarios are closely matched as it is likely that in building a complex structure all scenarios will be required. It is accordingly desirable to thermally model the whole deposition process to determine the actual quantity of heat to be added such that the freeze rate of bead 22 is similar throughout the process no matter what geometry or other parameter may affect the solidification metallurgy of the bead.

To minimise the overall amount of additional heat which must be added to control the average freeze rate it is desirable to optimise the tool path. Typically, this avoids, where possible, the scenarios of Figs 5c and instead aims to deposit beads or spots that have symmetrical geometric features on their opposing sides.

An exemplary tool path giving the optimum symmetry is shown in Figure 6. Beads 1, 2 and 3 are deposited directly to the substrate 12 and are separated by one bead width. Bead width in this context means the radius of the centroid at which melting takes place under the processing conditions.

Each bead is deposited to a symmetrical geometry and will have a similar freeze rate. Because the substrate has a high mass relative to that of the bead any heat rise to the substrate from the beads is negligible.

The subsequent passes of the deposition head interleaves the first beads 1, 2, and 3 with the subsequent beads 4 and 5. These beads are also deposited to symmetrical geometry, albeit different to that for passes 1, 2 and 3, and the freeze rate from each of these beads is symmetrical. As there is a greater surface area contacting these beads which will improve the freeze rate it may be desirable to supply additional heat to control the rate to match that of the first deposited beads 1, 2 and 3.

Once the first layer is complete the second layer may be deposited (starting with bead 6) by following a similar route. The layer preferably starts above the coldest bead i.e. that deposited first and given the greatest time to cool.

One particularly preferred laser head for providing a controllable flux distribution is depicted in Figure 7. In the embodiment shown the laser head is provided within a concentrically arranged powder feed arrangement 16. Power is supplied from the laser generators 50-54 to the head through a series of fibre optical cables. One laser 50 is provided to generate the energy which melts the substrate to form the melt pool into which the powder is supplied. The remaining lasers 51-54 provide the secondary heat source energy for controlling the heat flux distribution.

The lasers are infra-red fibre lines which produce a wavelength capable of being transmitted by fibre optic. Rather than splice fibre optics, which can be difficult and risks beam quality, it is desirable to provide one laser per fibre optic. This arrangement also provides maximum flexibility. Five lasers are shown in Figure 8 though it will be appreciated that a larger or smaller number may be used. Typically the largest number is limited by the size of the laser head 14 and the number of fibre optic cables which will fit within the head.

Each of the lasers delivers power of the order of tens of Watts for a head having a larger number of optical cables to power up to one kW for a head having fewer optical cables. The preferred combined power delivered is in the range of 100W to 2kw.

The preferred fibre optic cable arrangement is shown in Figure 8. 41 (Forty One) cables are provided as a rosette arrangement, confocal or closely focussed, with each cable having a diameter of 600pm and being surrounded by a sheath. The total diameter of the head is of the order 7mm.

Where the deposition process uses a powder to form the structure the powder head 16 extends coaxially around the cable arrangement.

The central cable 50 supplies the power to the substrate which forms the melt-pool therein. This fibre optic typically connects to the laser with the highest power rating and is most focused on the substrate. The cables surrounding the central optical cable typically connect to lower power lasers and are used to control the heat flux distribution around the weld head. The outer lasers may be less focused than the primary central laser, or arranged in a skewed orientation.

The fibre optics focus the laser energy to the substrate as an array of spots. Each of the spots may be distinct or overlap with one or more of the other spots. It is desirable that the focus of the fibres is set during manufacture of the head. A screw mechanism or other appropriate method may be used to move the focus from each fibre to individually align any spot relative to any one or more of the other spots.

The array arrangement along with the controllable lasers permit a switchable and configurable delivery of power around the perimeter of the melt pool, to allow control of the temperature distribution and isotherm spacing in response to the pseudo steady state heat flow in the region ahead, below and to the sides of the melt pool. The controllable temperature distribution or heat flux permits compensation for geometrical effects to allow for the optimum pool size and shape and cooling for microstructure and solidification control, and the precision of build characteristics.

The laser head is connected to a control system which takes outputs from thermal and solidification models for the component being manufactured and provides a varying heat input intensity distribution to create an optimum processing control band for a feed forward control system to operate within. A thermal camera, or other means, may be used to monitor the actual processing and compare it with that of the model. The model may be recalculated based on any deviation from the empirical data.

The lasers other than the primary laser are used to pre-heat or post heat the deposition area in accordance with the preferred method of operation. By controlling the spacing and distribution of the heat flux for a given material, together with the bead size and travel speed, the cross-sectional profile of the thermal gradient beneath the bead can be influenced.

Pre-heating particularly influences the wetting angle of the pool as well as the thermo-mechanical stress

field locally for the deposition. Post-heating

influences the rate and direction of pool solidification.

Beneficially, the arrangement of the fibre optics in the preferred embodiment enables the flux distribution to be adjusted even for non-linear translations of the weld head i.e. as the weld head moves around a corner, or follows a curve. The control system can adjust the power supplied to each of the laser outputs as the head turns.

Where the head is required to follow a linear path a simpler arrangement of optical fibres may be used as depicted in Figure 9. The distribution of fibres is non-symmetrical around the primary laser. With head travel in the direction of the arrow 60 a first array of lasers provide selective heating to the substrate in advance of the melt pool. A larger array of lasers downstream of the melt pool allow for more sensitive control of the solidification microstructure.

It will be appreciated that the present invention provides an optimum cooling condition that generates uniformity with respect to microstructure, stress, segregation and porosity. By controlling melt pool solidification, more complex and crack sensitive alloy systems can be deposited. Additionally, process stability is improved along with capability and predictability of the component.

Claims (11)

1. Additive manufacture apparatus comprising: a heat source for forming a melt pool in a substrate, a media supply for supplying media to the melt pool, and at least one further heat source for applying an adjustable heat flux power or distribution to the substrate.

2. Additive manufacture apparatus according to claim 1, wherein the at least one further heat source is adapted to affect at least one parameter influencing the solidification metallurgy of media supplied by the media supply and melted in the melt pool.

3. Additive manufacture apparatus according to claim 1 or claim 2, wherein the at least one further heat source comprises a plurality of fibre optic cables.

4. Additive manufacture apparatus according to claim 3, wherein each cable is functionally connected to a laser.

5. Additive manufacture apparatus according to claim 4, wherein each cable is functionally connected to a respective laser.

6. Additive manufacture apparatus according to claim 4 or claim 5, wherein the or each laser is a Nd:YAG laser.

7. Additive manufacture apparatus according to claim 6, wherein each laser is individually controllable.

8. Additive manufacture apparatus according to any preceding claim, wherein the further heat source is arranged around the heat source in a rose.

9. Additive manufacture apparatus according to claim 2, further comprising sensing means for detecting the at least one parameter.

10. Additive manufacture apparatus according to any preceding claim, further comprising a control system for continuously controlling the heat flux power or distribution whilst media is applied to the substrate.

11. Additive manufacture apparatus substantially as hereinbefore described with reference to Figures 7, 8 and 9.