Aeroengine Safety

Fig. "Diffusion brazing" (Lit. 21.2.2-1 up to Lit. 21.2.2-5): Especially the replacement of the blades/vanes from stator and rotor with a comparatively short lifetime mean a cost factor which is not to understimate. This motivates repairs and its development. It has shown that repaired parts (HP- turbine stator segments) can be made again serviceable at one forth of the new parts price.
Proved have high temperature brazings, which can close also cracks which can gape in the millimeter range, build up wear surfaces and can replace whole areas (Lit. 21.2.2-3). These processes are designated as „diffusion brazing with wide gap” (vertical sketch sequence right), „high temperature wide gap brazing“ or „activated diffusion healing” (ADH according to GE).
They differ from the high temperature brazings, used at new parts. They can be only applied for a defined narrow gap (vertical sketch sequence left, Lit. 21.2.2-5, volume 4 Ill. 16.2.1.4-9). For wide gaps they are inapplicable because they form brittle phases.

Both brazing processes use from repair welds (Fig. "Expection of fatigue strength from repair weldings") differing brazing metals which are widely similar to the base material. Wide gap brazing metal contains more coarse powder grains, similar to the base material. These are mixed to a paste with a binder and an addition which lowers the melting temperature. This addition consists out of elements like boron and/or silicon. After the brazing process itself is finished under the high process temperature the diffusion equalizes the concentration of the additions. With this the melting point of the braze joint invreases and approaches this of the base material. Therefore such a braze joint can afterwards not easy separated by heating, for example if a failure exists.

High temperature brazing repairs must solve a special problem. After operation times af many 1000 hours in modern aeroengines the hot parts are heavily oxidised as well at the surface as in the cracks (frame below). These oxides prevent the bonding/wetting with the braze metal and so a satisfying brazing temperature. This means oxides must before sufficiently removed. This happens with a combination of abrasive blasting (oxide blasting/vapour blasting), etch processes and heat treatment. Oxides are removed out of the cracks by annealing in hydrogen or an aggressive atmosphere like fluorine. Anyway the cleanliness of the surfaces for brazing stays problematic.

Because of the braze characteristics (brittleness and low creep strength, compared with the base material) and possible remainders of oxides, high temperature repair brazings can be similar problematic as repair welds (Fig. "Expection of fatigue strength from repair weldings"). It can be assumed, that the strength properties of the undeteriorated base material can not be fully reached. If necessary the OEM limits respectively specifies a combination of new parts and repair parts in the aeroengine (Fig. "Limited inserted number of repaired parts" and Fig. "Limits of welding and brazing repair"). There must be kept closely attention at such limitations.

Figure "Brazing distortion problems" (Lit. 21.2.2-7): Brazing and high temperature brazing (diffusion brazing, Fig. "Diffusion brazing") can trigger distortion under unfavourable temperature control and uneven temperature distribution. Are the temperature differences sufficient high, thermal stresses can exceed the lowered warm yield strength respectively creep strength (volume 4, Ill. 16.2.2.4-14 and Ill. 16.2.2.4-15.1). Then plastic deformation will take place (upper diagram). Especially complex, integral, casted components like turbine stators of small aeroengines (sketch right) are endangered by distortion. This is especially true for a restrained thermal expansion between the stiff shroud rings and the relatively thin cross sections of the radial orientated vanes (volume 3, Ill. 12.5-15). A deformation of the vane trailing edges influences the extremely important and exactly to meet free flow section(Bild 21.2.7-3).

A further problem arises, if parts which consist of different materials, are brazed. Here a differing thermal expansion behaviour can trigger distortion (bimetal effect).

An exception are seal/rub coatings of elder aeroengint types from brazing metal. For this application used silver brazes have a markedly higher thermal expansion than steels and Ni alloys (diagram below, Lit. 21.2.2-9). Therefore this coating inside of a steel ring tends to tensile stresses during cooling after the brazing. This promotes the lifting from the bearing ring (volume 2, Ill. 7.2.2-31) at operation temperature. This can also occur when the adherence of the coating is not optimal (volume 4, Ill. 16.2.1.4-22). The consequence is a catastrophic labyrinth failure with the failing of the aeroengine (volume 2, Ill. 7.2.2-4).

Figure "Deteriorations by the brazing prozess" (Lit. 21.2.2-7): Surface contaminations at used parts by hot gas corrosion (sulfidation) and oxidation can also dangerously contaminate the brazing atmosphere in the furnace. This danger does not exist in the production of new parts (volume 4, chapter 16.2.1.4). Therefore in the following will be pointed at such potential deteriorations during brazing.

Nitrogen diffusion (nitriding, Lit. 21.2.2-10): In most cases existing oxygen forms a protecting oxide layer from the remained oxygen. This avoids the invasion of nitrogen from the brazing atmosphere and so a deterioration.

Therefore only seldom a uncontrolled nitrogen diffusion in a nitrogen brazing atmosphere occurs. This can be regarded as an inner corrosion process (inner mitriding). Requirement for an inner nitriding is an oxide free surface in atomic nitrogen. In Cr containing steels form needele-shaped structures of Cr nitrides and degrade the material properties. Does the nitriding prevent a FeCrAl alloys the formation of a protecting Al₂O₃ layer, the sensibility for hot gas corrosion of the brazed parts increases markedly (similar „green rot“).
Ni alloy are vastly durable for nitriding.

Embrittlement by sulfur (sulfidation, Lit. 21.2.2-10, volume 1, Ill. 5.4.5.1-4) can develop at nickel alloys, if sulfur containing media act at high temperatures. This deterioration propagates preferential along the grain boundaries and acts intercrystalline embrittling. A cause for the sulfur can be not sufficient removed sulfidation on the part during repair. Thinkable is sulfidation in the inside of hollow blades (volume 1, Ill. 5.4.5.2-2.1). Sulfidation can progress in Ni alloys by an order of magnitude faster than `simple' oxidation.

Phosphorous embrittlement can develop at iron and nickel alloys in connection with phosphor containing Cu braze metals.

Carburization/decarburization (Lit. 21.2.2-10) arises undesirable, if the brazing atmosphere contains a sufficient concentration of free carbon.

Carbonisation: The necessary carbon develops by the temperature caused decomposition of (organic) carbon compounds (e.g., oil, grease, paints, plastics). A cause can be the sooting of the binder from the brazing paste. The carbonisation of austenitic steels, less of Ni alloys, happens in an atmosphere which prevents the formation of a protecting oxidation layer. A sort of inner corrosion process with carbide formation develops. It preferential propagates along Cr carbides at the grain boundaries and twin boundaries. The bonding of the Cr may markedly increase the proneness of the grain boundaries for corrosion and oxidation during operation (see also sensitising, volume 1, Ill. 5.4.1.1-3).

Decarbonisation is caused by a diffusion of the carbon out of the surface. It occurs at hardenable respectively case hardened steels. This carbon oxidises. A soft surface with lowered hardness and fatigue strength will develop.

Hydrogen embrittlement (volume 1, chapter 5.4.4) can occur with the diffusion of hydrogen into metals. Especially prone are high strength steels but also Cu alloys. For the diffusion necessary atomic hydrogen develops at brazing temperature from dissociated hydrocarbons of humidity. For an embrittlement already seemingly harmless small quantities are sufficient.

Crack formation at stresses which rise above the heat resistance/creep strength during the brazing process (volume 4, Ill. 15.1-19 and Ill. 15.2-13). Those stresses can arise from temperature differences/gradients in the part:

• During heating or cooling.
• Differences of thermal expansion (volume 4, Ill. 16.2.1.4-11) against a (brazing) device or between different materials.

As internal stresses (e.g., from a strain hardening or machining) they may already existed. In complex geometries with notches respectively steep changes in stiffness (Fig. "Brazing distortion problems") from experience repair brazings are just there, where are also the highest thermal loads during operation (e.g., transition from the blade to the shroud). In such a case it must be reckoned as well with an increased danger of crack development during the brazing process as also later during operation. The lifetime of such a braze will be „ill fated”.

Oxide formation (Lit. 21.2.2-10): For a sufficient lifetime of a hot part, stable, dense oxides are essential. Especially an Al₂O₃ layer can be extremely aggravating for the brazing prozess. Such oxides prevent the distribution of the braze metal and the wettability, necessary for the bonding. This is the most frequently reason for the lack of fusion and flaws in the braze seam. Such during the brazing process unwanted oxidation layers can form of oxygen from different sources:

• Brazing atmosphere/ insufficient vacuum (e.g., leakage).
  
• chemical reactions, too little „reducing“ acting atmosphere (hydrogen, vacuum).
• Not suitable/humid flux.

The lower the oxigen partial pressure and the higher the temperature, the better is the reducing effect. From the so called Ellingham diagrams it can be seen, under which conditions a certain oxide at least theoretically can be removed in a brazing atmosphere. Because there also other influences play a role, such estimations must be be secured by tests.

Carbide formation (carbide precipitation) can occur in the temperature region above 425 °C (volume 1, Ill. 5.4.1.1-3) at unstabilised CrNi steels and some Ni forgins/wrought alloys (e.g., Nimonic 80A). Thereby Chromium at the grain boundaries forms carbides Karbide. So it is no more available for corrosion protection. So sensitized materials are prone for intercrystalline corrosion (intergranular attac = IGA) and stress corrosion cracking (SCC) in process baths (e.g., etching before penetrant inspection) and aggressive media during operation.

Initial/incipient fusing (alloying) of the base material: High brazing temperatures can lead to a solution of the braze metal in the base material (Fig. "Diffusion brazing"). Does in such a case the braze metal drain along the surfaces, cross sections can get markedly thinner (brazing erosion). Attention must be payed at this effect during brazing typical thin walled honey combs (volume 4, Ill. 16.2.1.4-7 and Ill. 16.2.1.4-10). The processing temperature of high temperature brazing metals often lies close under the solidus temperature (softening temperature) of the base material. Thereby local initial fusings can occur (Ill. 16.2.1.4-13).

A further failure mode are local initial fusings by incumbent low melting particles like coating powder (e.g., aluminium for diffusion coating), machining chips of other alloys and littered brazing powder. Even such microscopic small damages (volume 4, Ill. 16.2.1.4-14) can inaccaptable influence the operation properties.

Do in the concerned cross section sufficient high tensile stresses prevail, the danger of liquid metal embrittlement (LME, volume 4, Ill. 16.2.2.3-11) exists.

Figure "Risk potential of brazing repairs" (Lit. 21.2.2-8): In this case at a repair brazed HP turbine vane/nozzele it came to a break out at the airfoil (sketch below, description of the event in Fig. "Borescope findings at turbine blades and vanes"). After the so called `partitioned alloy component healing' (=PACH) penetrating cracks have been high temperature brazed (sketch above left, Fig. "Diffusion brazing"). For the proving a set of nozzles was assembled and went into operation, monitored by borescope inspections. Misunderstandings prevented the identification of the fauilure in time.

The beak out of the nozzle airfoil in the region of the repair brazed cracks was triggered by thermal fatigue, lead to an intense disturbance of the gasflow. This provoked resonance vibrations of the HP turbine blades running behind (sketch above right). It came to a fatigue crack and the destruction of the whole blading.