·Table of Contents
·Industrial Plants and Structures
NDT methods for revealing anomalies and defects in gas turbine bladesAuthor: J. Pitkänen
VTT Manufacturing Technology
Co-authors : T. Hakkarainen, H. Jeskanen, P. Kuusinen,K. Lahdenperä & P. Särkiniemi
VTT Manufacturing Technology
Technical University of Helsinki
Dye penetrant testing is the NDT method most frequently used for inspecting gas turbine blades and vanes. It is often recommended to use fluorescent dye penetrants for inspection. After some time in service the blade surface is often corroded. In these cases fluorescent dye penetrant is not recommended and normal dye penetrant is more usable. The drawback of dye penetrant testing is its suitability only for surface opening cracks. Subsurface defects are not detected. When a crack is found with dye penetrant testing, it is not possible to estimate its depth, and here other methods are needed.
There are many methods which can be used for detection of cracks, ageing and degradation of a coating, and for thickness measurement. Thermal methods can be used for measurement of wallthickness and detection of near surface anomalies like delaminations (Aladin 1996, Carl et al 1998a), problems in cooling channels (Carl 1998b).
It has been observed that for instance in MCrAlY-coatings the permeability in base material as well as the in coating are changing during the lifetime of a component. In MCrAlY-coating the decrease of Cr-content and /or b -phase change the magnetic properties to more ferromagnetic. Even though the effect is not large, it can be detected in some cases (Czech et al 1998). Czech et al measured the magnetic permeability with a coil in conjuction with permanent magnets. So they could estimate the lifetime of coating. This is also possible with the eddy current method developed by ENEL (Antonelli 1998 a, b). The method developed by ENEL can also tbe used for the determination of the coating thickness of MCrAlY-coatings.
|a) 3 surface breaking cracks on a turbine blade||
b) Eddy current signals from cracks.
Fig 1: Eddy current signals from the three cracks shown in the blade .
|Fig 2: EC thickness measurement of the coating from a 1. stage vane, material X45 with ceramic coating.|
During the lifetime of a material properties can vary drastically as shown in figure 3. In these cases the strong effect from the permeabitility cannot be corrected easily. In this case the coating from IN738 blade is damaged. The damaged area is gives a similar signal to ferritic steel. While in the undamaged area the measured signal was more like one from stainless steel (AISI 316). The changes of the material properties on the surface of a blade make the inspection more difficult both in the case of cracks and in measuring the thickness of the coating.
|Fig 4: Thickness measurement principle The time of flight of leaky Rayleigh wave varies according to the thickness of the coating|
|Fig 5: Thickness measurement with the time of flight of leaky Rayleigh wave from a MCrAlY- coated turbine blade|
The developed probe is also suitable for detection of cracks , if the leaky Rayleigh wave is penetrating into the base material. In case where leaky Rayleigh wave doesn't penetrate into the base material, it is still possible to use normal Rayleigh wave. The leaky Rayleigh wave decays rapidly when crack like defect are present. In the figure 6. we see that with leaky wave mode it is easy to detect even small cracks down to 50m m. In depth the resolution depends on the frequency (wavelength) of the probe.
|Fig 6: (a) MCrAl-coated GT-blade, in which the EDM- notch depths vary between 50mm - 500mm (b) ultrasonic B-scan image in pitch-catch mode. (c) Ultrasonic B-scan image in pulse-echo mode from the blade with EDM-notches|
|Fig 7: Material X45 is measured with contact wide aperture probe is shown in fig. 7. The center frequency of the probe was 12 MHz. The plate is shown in upper picture and C-scan result in lower picture|
Same technique is used for thermal fatique crack detection in fig 7. As it is clearly seen two cracks in this plate can be detected with this contact probe. The plate is about 3 mm thick and one of the cracks is extending throughout the wall (in the middle of the plate).
Other application possibilities for developed probe type are detection of delaminations, measurement of material wave velocity and elasticity.
In X-ray measurement the radiation decays differently in various materials and discontiniuties. The changes are affected by density variation, thickness variation, variation in composition of the material and from lack of material (corrosion, cracks). In figure 8 the realtime radioscopy equipment is shown. The inspected object is located between the source and the image intensifier. The object is radiated and in the image intensifier measured radiation is digitized through video card. This information can be saved on the hard disc or CD-rom. These measurement has been carried out with X-ray tube of 160 kV.
|Fig 8: The PC- real time radioscopy system in evaluating radiographic results|
Fig 9: X-ray picture from the inside structures of the gas turbine blades
|Fig 10: Cooling channel crossing, which could cause from the edm-manufacturing of r the channels.|
In pictures cracks, pores, geometrical thicknesses, material density variations can be detected, figure 9. Especially cooling holes are clearly seen in the pictures, figure 10. Pores have been detected mostly in radiography measuments. The sizes of those pores are about same as the dimensions of the cooling channels. The blocks in the cooling channels can cause extra stress to turbine blade. The cooling of the blade is changed drastically and the temperature distribution on the blades changed can rapidly damage the blade.
Residual stresses were measured from a turbine blade material In 939. The measuring instrumet was XSTRESS 3000 (Stresstech Oy) X-ray diffraction equipment. The radiation was focussed to the object with help of a collimator, the diameter of which was 3 mm. Exposure time was either 20 s or 40 s and y -oscillation ±5°. In the measurement point was the 2 j -rotation ±20° with 10 angles. With this arrangement exposure time for one point was 400s or 2000s. With this measurement the variation in half width of X-ray diffraction spectrum according to the stress state were measured. In figure 11 the measured X-ray diffraction spectrum is shown from the area 2 in table 1, which was thermally loaded. In table 1 are shown the measured residual stresses from 3 areas in the same turbine blade. Additional X-ray diffraction spectrum can give some information about the degradation of the coating and also from the lifetime of the blade.
|Area n:o||sx (MPa)||sy (MPa)|
|Table 1: Residual stress measurements from 3 areas of a turbine blade|
Fig 11: The half width value of the X-ray diffraction spectrum from residual stress measurement
Most of the methods are available in laboratory stage only. But there are some techniques, like eddy current and ultrasonic techniques, which can be utilized in practical inspections on site.
The driving force of NDT-evaluation is of course the utilities' need to optimize the plant operation and to minimise the risks of service damage.
|© AIPnD , created by NDT.net|||Home| |Top||