Keywords

1 Introduction

The requirements regarding performance and reliability of modern aircraft engines place high demands on the integrity of severely stressed engine components such as the turbine, compressor and fan blades and their condition assessment within the scope of inspections and maintenance measures. In order to be able to detect and assess the need for repair at an early stage and to consider it in further planning, a fast and flexible initial assessment of engine components is becoming increasingly important.

In particular, the turbine blades of the first stage after the combustion chamber are subject to high thermal, mechanical and chemical loads during operation, which leads to changes in the microstructure of the high-temperature materials used, and thus also in the safety-relevant material properties with increasing operating time (Bräunling 2015). In order to withstand the severe environmental conditions, the turbine blades are equipped with a coating system that reduces the effects of thermal and corrosive loads. Due to the specific structure of such a turbine blade, appropriate testing technology is necessary for valid testing. Within this subproject of the CRC 871, a systematic investigation of the corresponding structural changes was carried out and the causes and effects on the component properties was analyzed. Within the context of the CRC 871, several new repair processes for engine components have been developed. Non-destructive defect testing (e.g. crack detection) and material characterization of the repaired areas were carried out. The focus was on quality assurance of the regeneration measures in the regeneration path. Thus, the material and component condition were evaluated depending on the selected repair processes as the new repair processes also affect the resulting material and component properties in the repair-welded, repair-brazed and coated areas.

The basic structure of a turbine blade shown in Fig. 1 can be divided into three different materials and functions. The substrate, which ultimately has to withstand the mechanical, thermal and chemical stresses is usually made of a nickel (Ni)- or cobalt (Co)-based superalloys. The degradation resulting from chemical attack is counteracted with the help of a corrosion protection layer, which is applied directly onto the base material. The standard are MCrAlY coating systems. The external thermal barrier coating (TBC) consists of a ceramic, often yttrium-stabilised zirconium oxide (YSZ), which significantly reduces the temperature of the base material due to its low thermal conductivity (Kumar et al. 2015; Reed 2006).

Fig. 1
A diagram of the structure of a turbine blade, starts with the base material and superalloy such as nickel or copper. This is covered by a corrosion protection layer, ranging from 20 to 50 mu meters, and ends with a thermal barrier coating.

Structure of a turbine blade consisting of base material, corrosion protection layer and thermal barrier coating; the coating thicknesses are given in μm

Full size image

In addition to the structure of a turbine blade, knowledge of typical types of damage is necessary in order to adapt the testing techniques in a targeted manner and to be able to evaluate the quality of the test results. The damage that occurs in the course of operation can be assigned to geometry, such as foreign object damage (FOD) or burnings, to structure, for example in the form of cracks, or to the material itself in form of corrosion (Carter 2005). Figure 2 illustrates typical damage types of an operated turbine blade. These defects affect the functionality of the blade and can lead to final failure if regeneration measures are not taken. Therefore, early detection of defects and evaluation of their impact on reliability is essential.

Fig. 2
A photograph of a turbine blade with typical damages from operational loads, including F O D and spalling, cracks, oxidation, and burning.

Turbine blade showing typical damages due to operational loads: the individual damage types have been assigned to the categories geometry (purple), structure (red) and material (green)

Full size image

2 Objective

In order to carry out a targeted regeneration of individual components, it is essential to have detailed information about the location and extent of defects. This will be achieved, in particular, with the aim of flexible initial assessment, which represents a core element of this subproject. The special properties of a high-pressure turbine blade require a combination of different non-destructive testing (NDT) methods to fully characterize the condition of the coating system and the base material itself. By combining NDT methods, the inspection resources can be used such that high quality of the inspection results with the lowest possible inspection effort. This adaptive inspection sequence is to be used after the disassembly of an engine to allow a targeted alignment of the regeneration paths. With regard to highly stressed turbine blades, the characterization of thin layer and edge zone conditions plays a particularly important role. However, an intact microstructure of the base material, which can be negatively influenced mainly by high thermal and corrosive loads, is crucial for component integrity. Furthermore, the detected defects must be classified.

The flexible initial assessment, illustrated in Fig. 3, shows one approach how such a system can be designed. In the first process step an evaluation of the condition of the base material is performed. For this purpose, a harmonic analysis of eddy currents (HA-EC) is carried out with a sensor in integral design. This test is intended to evaluate microstructural degradation so that the base material can subsequently be assessed as operational or not. If the condition assessment of the base material leads to a negative result, no further testing resources need to be used, as this criterion leads to a reject. The next step is to check the blade for cracks, spalling and delamination of the coating system using active high-frequency thermography (HF-Thermography). These defects can be regenerated depending on the location and size of the defect. As a final test to ensure a holistic view of the blade condition, a characterization of the coating system needs to be performed. Here, high-frequency eddy current (HF-EC) test is used. The high frequencies ensure that the eddy currents mainly propagate in the subsurface zone, and thus the condition of the coating system can be evaluated. These steps can be carried out without having to strip the coating of the turbine blade. If no relevant microstructural change or damage can be detected in all three steps, the turbine blade can still be used without repair measures.

Fig. 3
A model diagram of the flexible initial assessment process chain begins with material conditions, followed by crack detection, then layer characterization, and concludes with coating.

Schematic illustrating the flexible initial assessment process chain, see main text for details

Full size image

However, the flexible initial assessment developed is not only intended to allow targeted regeneration, but also to evaluate the regeneration measures themselves and to check the condition of the regenerated component. Here, it is important to reliably detect any process faults, such as lack of fusion after a welding process. With this approach, a process chain is available at the end that can also be used in quality assurance (Melchert et al. 2021; Rocha et al. 2018; Wang 2013).

A further working hypothesis is that the non-destructive in situ recording and characterization of the damage development of regenerated specimens in comparison to reference specimens under thermal and cyclic loading provide relevant information about the regeneration-related changes in the component properties, especially under cyclic thermomechanical loading. This allows evaluation of the repair methods used in connection with the functional benefit of the regenerated components, which results in a significant gain in knowledge for the evaluation of the repair measures.

By combining the non-destructive testing techniques with fatigue tests, a statement on the expected remaining service life can also be made. For this purpose, eddy current techniques, thermographic and acoustic methods were used in situ on a fatigue test rig. In addition to a residual life estimation, this combination also generated evaluation criteria for the HA-EC technique.

3 Material and Methods

Harmonic Analysis of Eddy Currents

The flexible initial assessment consists of three different testing techniques. HA-EC was used for material characterization. In the course of preliminary investigations, it was shown that the use of nickel-based materials at high temperatures in a sulphurous atmosphere leads to significant changes in the magnetic property profile due to the formation of ferromagnetic phases. This indicates a local change in the microstructure of the material, which is paramagnetic in its initial state. The eddy current sensors can be designed for integral testing, as shown in Fig. 3, or for local testing in scanning technology. To realize a problem-adapted testing, various sensors were developed and used in the subproject.

Thus, the magnetic inductive test method of harmonic analysis of eddy current signals was used in the present study as it can be applied to actual components. The harmonic analysis of eddy current signals is a non-destructive test method, which is suitable for the sensitive evaluation of magnetic material properties. The measuring principle of this non-destructive testing technique is shown in Fig. 4. A coil carrying an alternating current (AC) generates an alternating electromagnetic field (excitation signal), which induces eddy currents in an electrically conductive test specimen. In a (locally) ferromagnetic specimen, this also causes substantial remagnetizing processes. The resulting eddy currents and the remagnitization in the material generate a secondary magnetic field, which is superimposed on the primary field. These magnetic fields then induce a voltage in a measuring coil. If the sample has ferromagnetic or ferrimagnetic properties, the sinusoidal wave form of the voltage–time profile obtained for paramagnetic substances is now superimposed by so-called higher harmonics, which are due to the distortion of the magnetic field due to the nonlinear relationship between field strength and flux density. Using a Fast-Fourier-Transformation, the measured signal can be spectrally decomposed and the amplitude of the fundamental (1st harmonic) wave as well as the 3rd and 5th harmonic can be obtained. The eddy current test in combination with harmonic analysis is a well-established method for characterizing the properties of steel materials (Maaß et al. 2004, Mercier et al. 2006; Stegemann et al. 1998). In addition to the values for the amplitudes of the harmonics, this method can also be used to determine the Curie temperature of the examined area. The Curie temperature describes the temperature at which a material changes between ferromagnetic and paramagnetic behaviour. If the material is in the paramagnetic state, there is no distortion of the signal, and the higher harmonics vanish. This can be exploited to determine the Curie temperature. The Curie temperature is not a constant but depends on the loading history of the material. If the material is exposed to increased temperatures, more oxidised volume is formed. This is reflected in an increase in the Curie temperature, so that a statement about the material’s oxidation condition can be made with this method (Barton 2022, Barton et al. 2021, Delaunay et al. 2000, Garat et al. 2005, Jia et al. 2014).

Fig. 4
Several graphs depict the relationship between excitation and the measuring signal. The excitation graphs include amplitude versus time, density versus magnetic force, and flux versus density. These result in measuring signal graphs, which depict amplitude versus time and a corresponding bar graph.

Influence of ferromagnetic (top) and paramagnetic (bottom) materials on eddy current signals (Barton et al. 2021)

Full size image

3.1 Induction Thermography

Thermographic test methods can be divided into passive and active methods. In active methods, it is not the natural thermal radiation of the test object that is decisive, but an actively generated temperature change. Active methods are particularly suitable for detecting detection of cracks and other thin and sharp-edged defects in electrically conductive materials. Different from other optical methods that use the visible light spectrum, induction thermography is not based on effects such as reflection and shadowing, but uses heat generation and heat flow resulting from actively introduced electromagnetic energy in an examined area. The strong eddy current field created during inductive excitation causes increased heat release in areas of disturbed flux due to the local increase in current density. The increase in current density and the corresponding temperature rise take place mainly at sharp edges and other narrow geometric shape features of the test object, as well as in areas with defects and inhomogeneities in the material. The defect information resulting from the temperature distribution on the sample surface can be obtained and visualised using a thermal imaging system. The fundamental parameter that determines the applicability of this method and its sensitivity for defect detection is the eddy current penetration depth, which depends on the magnetic permeability and electrical conductivity of the material (Netzelmann et al. 2016, Schloblom et al. 2016). The thin-walled turbine blades made of nickel-based superalloys with low electrical conductivity (σ ≈ 0.6 MS/m) and low magnetic permeability (μr ≈ 1) generally require high-frequency excitation frequencies for damage detection. Assuming that not only the base material of the turbine blade but also the thin protective layer system of the blade is to be tested, excitation frequencies in the megahertz range are necessary to effectively generate strong eddy currents in the subsurface zone of the test object. The usual transistorised high-frequency induction generators operate in the kilohertz range. By contrast, the setup used was equipped with a high-performance tube-type generator that excites at frequencies between 0.1 and 3.5 MHz. An external microcontroller controlled the induction generator in pulsed mode. The locally introduced short-time high-energy pulses improve the dynamics and sharpness in the image of the temperature field and simultaneously reduce the global heating of the component. A series of successive excitation pulses can be used to generate a thermal response in the component, resulting in a specific temperature–time profile for each pixel. The best results were obtained with excitation times between 50 and 100 ms at 50% duty cycle (Schlobohm et al. 2016). When testing metallic components, long excitation times lead to temperature equalisation in the excited area and ultimately in the entire component due to the thermal conductivity of the material. As a result, the differences in the temperature field quickly disappear. Consequently, long excitation times reduce the contrast in the image between irregularities and the background and the measurement sensitivity becomes lower (Reimche et al. 2008, 2009). With pulsed mode excitation, the local relative temperature change was recorded. The thermal response was analysed and filtered with a Fast Fourier Transform (FFT) based algorithm at each pixel of the recorded image sequence. The advantage of this approach is that the local emissivity variations of the sample surface, e.g. due to impurities, can be compensated.

Figure 5 shows the schematic structure of the inductive high-frequency thermography system. The synchronisation of the individual components by the control system in particular is of major importance here. Induction thermography is suitable for real-time detection of cracks and similar defects with high optical resolution at different scales. By using changeable lenses with different focal lengths, it is possible to vary the magnification and the inspection times, i.e. coarse-fast/fine-slow. Furthermore, the inspection can be executed in automatic mode. For the inspection of the specimens, the thermography system was equipped with a macro lens that allows a lateral resolution of 15 μm per pixel and a free working distance of 300 mm. The presented high-frequency induction thermography technique allows distinguishing between cracks in the bulk and non-critical surface scratches, thus avoiding false defect indications. Tests on TBC-coated samples have shown that high-frequency induction thermography can be used not only to detect cracks in the substrate and in the coating material, but also to detect changes in the thickness of the thermal barrier coating and detachments between the coating and the substrate (Schlobohm et al. 2017).

Fig. 5
A circuit diagram of an H F thermographic system includes the following components. Test Object, Thermal Response, Pulsed H F Excitation, Generator, Image Sequence, Thermographic Camera, Controls, and Data Processing.

Schematic diagram of the HF thermographic system; see main text for details

Full size image

3.2 High-Frequency Eddy Current Testing

During the operating life of a turbine blade, the chemical composition of the aluminium-based corrosion protection layer (PtAl, MCrAlY) changes due to oxidation at high temperatures. In particular, aluminium oxides form in the highly stressed areas of the turbine blade and the content of metallic aluminium decreases significantly—the corrosion protection layer loses its protective function and must be regenerated (Bräunling 2015; Schlobohm et al. 2017). Maintenance, repair and overhaul (MRO) companies are permitted to repair a turbine blade coating system only two or three times. Therefore, there is a great interest in non-destructive characterization of the coating condition before the repair process. Since the electrical conductivity of the corrosion protection layer also decreases with the reduction of the metallic aluminium, the electrical conductivity can be used as an indicator of the condition of the corrosion protection layer. As will be shown below, the eddy current technique is a suitable test method for a quick and non-destructive assessment of the coating condition. Modern eddy current test systems use an eddy current sensor to generate an alternating magnetic field with a characteristic frequency in the low single-digit MHz range. Depending on the test frequency, the eddy currents can be induced at different depths of the test object. These eddy currents generate a secondary magnetic field with a characteristic phase shift and field strength and lead to an equally characteristic, sinus-shaped voltage induced in the sensor's measuring coil. The phase shift and amplitude of this measurement signal are mainly based on the electrical conductivity in dia- and para-magnetic materials and can therefore be used to estimate the layer condition (Delauny et al. 2000). However, due to the high penetration depth of the eddy current, conventional eddy current testing systems are not suitable for characterizing the coating state of the very thin corrosion protection layer separately from the base material (Bräunling 2015; Cosack 2009).

Within the framework of the CRC 871 a new HF-EC test system that operates in the frequency range up to 100 MHz with custom-build and miniaturised sensors, as shown in Fig. 6 was developed. With these sensors, it is possible to test in the gas path without having to dismantle the individual blades. Due to the high test frequency, a typical penetration depth of < 100 μm can be realized, allowing a separate and non-destructive characterization of the turbine blade coating condition (Frackowiak et al. 2018; Bruchwald et al. 2016).

Fig. 6
Four photographs display different types of sensors. The first is a 12 m m H F-E C sensor. The second and third are H F-E C sensors measuring 6 m m and 4 m m, respectively. The fourth is an E C sensor.

Miniaturised HF-EC sensors of the 1st and 2nd funding period and application of the sensors in the gas path

Full size image

3.3 Fatigue Tests with Combined NDT

As a result of the application of new repair processes, the mechanical-technological characteristic values of the components in the repair-welded, repair-brazed and coated areas change. In order to determine the characteristic material properties in the area of the repair points and to evaluate the repair methods used, a test rig that allow non-destructive in situ recording and characterization of the damage evolution in the repaired area under thermal and cyclic stress was developed, cf. Fig. 7. The system is capable of applying temperature cycles and mechanical load cycles to a sample. The sample is heated via an induction coil and can be cooled down as required by an active air cooling system.

Fig. 7
A diagram depicts an eddy current sensor in a waiting position. It includes the following components, T M F-subsystem, water chiller, load frame, induction coil, H--E C, an E-sensor, and an I-R-camera.

Fatigue test system with integrated NDT: The detailed view shows the eddy current sensor (HA-EC) in waiting position, the acoustic emission (AE) sensor in the shielding housing including sound conductor with contact to the round specimen and in the background the thermographic camera (IR-Camera)

Full size image

Since the test rig is capable of heating the samples to over 1000 °C by induction, additional measures had to be implemented to protect the test equipment. The sensor for harmonic analysis of eddy currents was installed on a positioning unit so in cooling phases it could be moved until it contacted the specimen surface. Since the sensor only records values ones the temperature is below the Curie temperature, a fixed installation of the sensor does not offer any benefit especially in high-temperature phases. The acoustic emission used for crack detection, on the other hand, depends on a continuous recording of measured values. It was not possible to attach the sensor directly to the specimen surface or the grips of the testing machine, as the electromagnetic compatibility between inductive heating and acoustic emission sensor was not given. Due to the inductive heating in the frequency range of 200 kHz, the AE sensor was massively disturbed without shielding. Another challenge was the high temperatures in the immediate vicinity of the sample. The temperature acting on the sensor could be reduced to tolerable values with the help of a sound conductor. At the same time, this also reduced the negative effect of the electromagnetic field of the induction coil on the sensor. Still, a complex shielding of the sensor against the intensive electromagnetic field was necessary. The thermographic camera was permanently installed. Recordings were performed by internal timer or by an external trigger of the test system.

4 Results

4.1 Harmonic Analysis of Eddy Currents

The results of the fast damage detection with an integral HA-EC sensor are shown in Fig. 8 (right). Nineteen blades of the same type were examined using this technique. Fourteen blades showed a high third harmonic amplitude reading, which correlates with significant changes in the microstructure, caused by corrosion, as the material does not have ferromagnetic properties at room temperature in the non-degraded state. The amplitudes here show large differences, which indicate a different volume of altered material. Five blades show no amplitude (blade number: 4, 6, 7, 9, 11), so that no structural change was indicated at room temperature.

Fig. 8
2 illustrations. Left. The photograph depicts an experimental setup consisting of a sensor, which includes a gripper, blade, and sensor. Right. A bar graph depicts measured amplitudes versus blade numbers.

Fast damage detection. Left: experiment set-up consisting of sensor, turbine blade and gripper at the end effector of an axial robot; right: measured amplitudes of the 3rd harmonic

Full size image

In addition to integral testing, it is also possible to carry out local surface testing using the same method. Only different sensors are needed for this purpose. The inspection times are increased in this case, but with this option, the damaged areas can be displayed in relation to the location. In order to reveal the blade areas damaged by high temperature oxidation, a local HA-EC sensor was used, cf. Fig. 9. The green circle in the left figure corresponds to the measuring spot size of the sensor. Especially in the area of the leading edge, the blade shows an increased amplitude of the 3rd harmonic. The remaining measuring points show clear external damage in the form of spalling. However, the base material is not affected by oxidation at these points (Barton 2022).

Fig. 9
2 illustrations. Left. A photograph illustrates the amplitudes of the third harmonic assigned to individual measuring points. Right. There is a heat map of these amplitudes, which illustrates the measured values as a percentage from 0 to 100.

A local HA-EC sensor was used to determine the blade areas affected by high temperature oxidation. Left: Amplitudes of the 3rd harmonic assigned to individual measuring points. Right. Testing of an entire blade surface and subsequent transfer of the measured values to a 3D mapping

Full size image

The Fig. 9 on the right visualises the test result of a complete inspection to identify oxidised areas in the base material. As expected, the leading edge is most affected by oxidation according to the measurements. The oxidised portion decreases in the direction of the blade root. In order to only determine the oxidised volume, but also about the intensity of the oxidation present, the Curie temperature must also be taken into account in addition to the amplitude of the 3rd harmonic. Figure 10 shows a clear linear correlation between the Curie temperature and the minimum chromium content. As the chromium content decreases, the Curie temperature increases. The chromium content is in itself directly connected to the oxidation progress, as the chromium atoms diffuse out of the subsurface zone and form chromium oxides on the surface. Figure 10b demonstrates the good correlation the 3rd harmonic in relation to the chromium depleted volume (Barton et al. 2021).

Fig. 10
2 line graphs. Graph A depicts the Curie temperature versus minimum chromium content, displaying a single declining trend. Graph B illustrates amplitude versus depth, exhibiting a single inclining trend.

Curie temperature versus minimum chromium content (a) and amplitude of the 3rd harmonic as function of the depth of the chromium-depleted region (b) (Barton et al. 2021)

Full size image

4.2 Induction Thermography

The HF induction thermography is able to detect delamination of the coating system and cracks in the base material below the coating. Figure 11 shows the crack detection using the example of a turbine blade with an artificial crack. The crack can be seen in Fig. 11, left. Prior to inspection, the blade was coated with a thermal insulation layer. After the coating process, the crack is not longer visible. No additional brazing material was applied at this point, i.e. the crack was still present underneath the coating. The blade was then inspected for cracks and delamination in the coated state. The test results are visualised in Fig. 11 on the right with the help of a 3D mapping. The result shows a clear heating at the position of the crack. As expected, the heating is only visible at the crack tip.

Fig. 11
Three photographs of different turbine blade conditions and tests are labeled from a to c. Photograph A depicts a turbine blade with an artificial crack. B depicts a turbine blade coated with a new corrosion protection layer. Photograph C illustrates a test for cracks using a high-frequency process.

Stripped turbine blade with artificial crack (a), which was subsequently coated with a new corrosion protection layer and thermal barrier coating (b) and then tested for cracks using HF thermography (c)

Full size image

Clearly, the artificially generated crack is different in its dimension and shape from real cracks. Therefore, in further tests, blades with cracks formed during service, which are significantly finer and smaller, were also tested. Since the system can also be used for quality assurance, it was also important to ensure that any defects present after the MRO process, such as lack of fusion in the welding process, can be reliably detected. For this purpose, a series of turbine blades was examined, which had real cracks in the tip area. These cracks were closed with a high-temperature brazing process and then a new coating system was applied to the blades.

Figure 12 shows a blade after the regeneration process, the previously clearly visible crack is no longer visible here and seems to be filled by the brazing material. Yet, the superimposed data from the thermographic test clearly revealed that there is still a crack-like defect present. To verify this, the area was metallographically analysed. The blade was cut at the expected position of the crack and then examined under a light microscope. The inset to Fig. 12 (right) shows that the crack present prior to the repair has not been completely closed. At the pressure side, the crack is closed. This defect is therefore classified as an internal defect.

Fig. 12
Two close-up photographs of a turbine blade with a crack in the tip area requiring repair and a metallurgical analysis of the same tip area.

Optical images of a turbine blade with a crack in the tip area that required a regeneration process. Left: superimposed data from HF thermography test revealing the crack; Right: subsequent metallographic analysis of the tip area

Full size image

High-Frequency Eddy Current Testing

The high-frequency eddy current technique with test frequencies in the MHz range allows separate characterization of the thin coatings and testing for defects in the coating system and the edge zone of turbine blades. The eddy current technique has its main advantage in determining the TBC thickness and assessing the corrosion protection potential of the underlying corrosion protection layer (Frackowiak et al. 2018).

The experimental tests were carried out on first stage turbine blades made of the nickel-based superalloy René 142 with a PtAl corrosion protection coating and YSZ as the TBC. Figure 13 shows the test result for the individual measuring points of the turbine blade shown. The measurement result is within the tolerance range, which was defined by measuring two reference blades coated with a ceramic non-conductive TBC with a thickness between 0–200 μm. One of these reference blades had an unstressed PtAl layer, while the other completely lacked such a layer. The phase position of the eddy current signal was chosen in such a way that the relevant geometry and the lift-off effect were mainly represented by the real part of the measured values in the complex plane. In order to be able to specifically evaluate and represent the occurring influence of the lift-off and geometry effect, the sensor was slowly moved towards the sample. This leads to the characteristic profiles of the working points in the complex plane. The magnitude of the lift-off effect corresponds to the thickness of the non-conductive TBC. When the sensor is in full contact, the lift-off effect is minimal and corresponds to the layer thickness at that position. The degradation of the metallic aluminium in the anti- corrosion coating due to oxidation and formation of the non-conductive ceramic Al3O2 during turbine operation leads to changes in the electrical conductivity of the coating over time. These changes are represented by the shift of the measured values mainly along the imaginary axis of the complex plane. The electrical conductivity of the corrosion protection layer is thus a characteristic property that can be used to assess the condition of this part of the coating system using the HF-EC measurement technique (Frackowiak et al. 2018; Reimche et al. 2013).

Fig. 13
A model graph illustrates a coating system using the H F E C measurement technique, depicting high and low stressed areas. Right, a line graph plots 2 shaded downward lines for Rene 142 and Rene 142 + P t A I. A close up view of both lines is depicted in the inset at the bottom left.

Assessment of the coating system using HF-EC measurement technique; in a single process, both corrosion protection and thermal barrier coatings (TBC) can be evaluated (Frackowiak et al. 2018)

Full size image

Fatigue Tests Combined with in Situ NDT

The developed fatigue test rig with combined NDT allows correlation of the NDT data with the fatigue curves. In general, tests were carried out with virgin smooth specimens, as well as tests with notched specimens and specimens regenerated by other subprojects.

The use of acoustic analysis to detect cracks proved to be challenging due to the inductive heating used. In isothermal fatigue tests at room temperature, i.e. without the use of inductive heating, the system can be coupled to the specimen by means of sound conductors to detect crack signatures. In the case of thermo-mechanical loading, the system was at least temporarily disturbed in all tests to such an extent that no reliable in situ crack detection could be realized.

By contrast, thermographic tracking of the surface could be reliably implemented. Still, it is still difficult to monitor the crack growth. On one hand, this is due to the duration of fatigue tests in the low-cycle fatigue (LCF) range, and thus the very large amounts of data that would be generated by a complete thermographic observation. In addition, it is not possible to monitor the entire surface of the sample with a single thermographic camera; this would require at least three cameras, and even then, part of the surface area would still be obscured by the induction coil.

By using a high-temperature HA-EC sensor, in situ microstructure monitoring could be realised. Using an adapted isothermal LCF test with the test sequence shown in Fig. 14c, the sensor was able to record data in recurring cooling phases. Upon cooling, the amplitudes of the 3rd harmonic were recorded, and thus the respective Curie temperature could be determined cyclically. Several tests with varying mechanical strains were carried out such that time effects and changes due to the mechanical stresses could subsequently be evaluated.

Fig. 14
2 photographs and a multiple-line graph are labeled from a to c. Left, the photograph depicts a sensor in a waiting position, B, depicts coils. Right, a frequency graph of mechanical strain and specimen temperature versus time is depicted. Specimen temperature depicts a downward curve at the end during the cooling phase.

Fatigue test cycle with combined HA-EC monitoring: in a the HA-EC sensor is in the waiting position and the measuring spot of the pyrometric temperature detection is indicated by the laser spot; in b the HA-EC sensor is in data acquisition (DAQ) mode during the cooling phase marked on the graph in (c)

Full size image

Figure 15 illustrates data obtained from these different tests. In the left diagram, the amplitude of the 3rd harmonic at a sample temperature of 25 ℃ is plotted versus the absolute number of cycles. It is evident that an increased mechanical stress leads to an increased growth of the amplitude. If the x-axis is now transferred from absolute number of cycles to percentage values for the fatigue progress and the respective values of the Curie temperature are used instead of the 3rd harmonic, an intersection point becomes visible. When a Curie temperature of approximately 90 ℃ is reached, each of the specimens tested has reached about 80% of its fatigue life.

Fig. 15
2 multiple-line graphs. Graph A, amplitude versus testing cycle depicts 3 upward increasing lines for positive and negative 0.3%, 0.5%, and 0.7%. Graph B, Curie temperature versus fatigue progression, depicts 3 fluctuating upward increasing lines for positive and negative 0.3%, 0.5%, and 0.7%.

Left: evolution of the amplitude of the 3rd harmonic with the number of test cycle; right: change in Curie temperature as a function of the normalized fatigue life

Full size image

Apparently, the Curie temperature is a usefull indicator of fatigue damage, as oxidation of the subsurface zone in this case is closely correlated with fatigue crack initiation and early growth. Thus, by continuously monitoring the electromagnetic parameters, useful input for fatigue life prediction can be provided (Barton et al. 2022).

5 Conclusions

In work the development of non-destructive testing techniques for the differentiated evaluation of the coating system of a high-pressure turbine blade has been presented. Low-frequency harmonic analysis of eddy current signals is used to evaluate the condition of the base material. This can sensitively detect ferromagnetic phases that have developed in the paramagnetic base material as a result of high-temperature corrosion processes. In addition to the integral determination of such phases, they can also be detected locally. In this way, damage can be evaluated in terms of its position and distribution. This testing technique is also suitable for determining the Curie temperature in components made of nickel-based materials. The maximum change in the microstructure can be derived from this. These testing techniques make it possible to evaluate the condition of a turbine blade, for example, and thus make a net decision as to whether this blade can continue to be used with or without repair, or whether it must be replaced (Barton 2022).

HF induction thermography was developed to detect cracks and damage in the protective layers. High-frequency, pulsed excitation generates an eddy current distribution in areas close to the surface, which is influenced by defects such as cracks or delaminations. This can be detected by means of a thermographic camera. In this way, these defects can be detected and evaluated by imaging. In order to be able to image the defects in high resolution, a thermographic camera with a macro lens was used and component-matched inductors were also developed. The phase and amplitude information of the individual image pixels was used for evaluation.

With the high-frequency eddy current test with test frequencies in the MHz range, the protective coatings can be evaluated. Sensors adapted to the task were developed for this purpose. In the case of the ceramic thermal barrier coating, degradation manifests itself in a reduction in thickness, which can be detected via a lift-off sensor. The formation of oxide on the corrosion protection layer reduces its electrical conductivity, which can also be detected by the high-frequency eddy current test. In this way, the condition of both protective layers can be recorded in parallel in one measurement and the condition can be evaluated from the result. The combination of these testing techniques provides data for the selection of the appropriate regeneration process depending on the component condition.

Furthermore, the findings from the condition assessment of the turbine blade were used to evaluate the quality of repair processes under thermo-mechanical load. For this purpose, non-destructive testing techniques were integrated into a suitable test rig. This made it possible to detect cracks at an early stage and to characterize the change in microstructure during cyclic loading.