Abstract

Gas turbines and particularly their hot path components exhibit relatively high maintenance cost and short in-service inspection cycles, in comparison with most main components of thermal power plants. The overall share of production capacity covered by gas fired combined cycle and CHP plants utilising gas turbines is still increasing, and the short inspection and maintenance cycles suggest corresponding strong growth for NDT services.

NDT of gas turbine components is typically specified for three different purposes: within the shop during different phases of manufacturing or repairs, for user acceptance of new or reconditioned components, and for in-service assessment on the run/repair/replace decisions. The most demanding inspections in many ways are those performed in the field, as the available methods and access for the inspections can be fairly limited. Majority of the field inspections apply visual and surface techniques, but also ET and UT appear to find increasing use as they provide the potential advantage of indicating flaw depth. Examples are shown on in-service and ex-service inspections of hot end components, particularly turbine blades and vanes.

Introduction

Importance of power plants using gas turbines has been increasing worldwide in regions with access to suitable fuels, mainly natural gas. Good efficiency in combined cycle and CHP service, lower emissions than from most other fossil plant, relatively modest investment cost and fast technical development are features behind the increased popularity of land-based gas turbines. However, the maintenance cost of gas turbines remains high because of high operating temperatures, short overhaul cycles and expensive materials required. Frequent overhauls also imply frequent inspections to ensure correct timing and extent of maintenance. Extensive use of NDT is therefore needed for inspecting gas turbine components. NDT is generally used for three different purposes:

• for quality control in the shop during different phases of manufacturing or repair,
• for user acceptance of new or reconditioned components, and
• to support in-service assessment for run/repair/replace decisions or forensic assessment after failures.

Typical faults and defects targeted in NDT of gas turbine components include original defects and deviations from manufacturing or repairs, and defects of coatings and base materials emerging and growing during service. Apart from cracks or other discontinuities, the deviations can also appear for example as wear, corrosion, excessive strain, or blocking or inappropriate positioning of cooling channels.

Much of the maintenance cost as well as risk related to unplanned gas turbine downtime is due to the hot end components of the gas turbines. Here, examples are shown on NDT methods applied for in-service and ex-service inspections of gas turbine blades and vanes.

NDT of land based gas turbines - surface inspections

The in-service inspections on-site are the most demanding ones, because of limited access to many components and limitations in the available (useful and accepted) NDT methods. The majority of the on-site work make use of visual inspections (Figs 1 and 2) and surface techniques. The dominance of surface inspections appears to remains although ET and UT find increasing use where depth indication of the defects is needed.

Fig 1: A gas turbine blade after failure, showing damage in a visual inspection.

Penetrant testing (PT) is efficient and well suited for revealing discontinuities such as cracks, pores and other defects opening to the outer accessible surface. On this surface, PT also indicates the length, shape and density of the defects. However, PT provides no direct indication of the depth of the defect, and this is the most important drawback of all PT methods.

Fig 2: TBC coating in a combustion chamber of a gas turbine after service.

PT involves cleaning of the surface to be inspected, and spreading of the liquid penetrant which then enters open defects due to capillary forces. After removal of excess penetrant from the surface, a developer is added to extract the penetrant from the defects and to spread sufficiently on the surface to visually indicate with enhanced contrast the location, shape and projected size of the defect. PT therefore typically requires the following phases:

1. surface cleaning
2. spreading of the penetrant, with waiting time for entering the defects
3. cleaning of additional liquid from surfaces
4. spreading of the developer
5. surface cleaning under approprite lighting conditions
6. final cleaning of surfaces to avoid corrosion and other damage.

Depending on the way the excess liquid is removed, penetrants are divided into water washable, solvent washable and post emulsifier types. These types show different degrees of sensitivity so that the water washable ones are least sensitive, and post emulsifier types most sensitive ones.

The materials used for PT (penetrant, cleaning liquid, developer) should be inter- changeably compatible. Typically, chemicals from same series and manufacturer fulfil this condition. The range of application temperatures of liquid penetrants is about 15 - 50 °C. Material to be inspected should not be porous, and the defects should not be closed at the surface; for example, shot peening or comparable surface treatment may close defects at the surface. Furthermore, the defects should not be filled by foreign material such as dirt, oil or grease which may prevent the penetrant from entering the defect.

In manual PT, solvent or water washable penetrant liquids are usually applied. Difficult (e.g. rough) surfaces are better suited for water washable liquids. Small, easily cleaned or corrosion-sensitive components are generally tested by using solvent washable liquids. Emulsifiers are most frequently used in partly automatised inspection stations, used particularly for testing of large amounts of relatively similar components.

Fluorescent penetrants are used similarly as colour dye penetrants, but requires using UV-A light (at least 10 W/cm², see EN 571-1) for the inspection. On the other hand, fluorescent PT provides best sensitivity to defects and is widely recommended by manufacturers for gas turbine components.

Fig 3 shows example results from PT of gas turbine blades. Dense field of cracks may induce indications where individual indications cannot be easily discerned. Under such conditions it may be more effective to inspect without developer, although the sensitivity is reduced so that smallest defects are not necessarily indicated.

Fig 3: Thermal cracking indications in MCrAlY coating at the leading edge of a blade, as shown by a) colour dye penetrant testing; b) fluorescent penetrant testing.

Eddy current testing

In eddy current testing (ET), a magnetic field is induced by a coil to the conductive material to be tested. The subsequent material response will create an opposing magnetic field that will in turn induce eddy currents in the material. These eddy currents, and the changes in their phase and amplitude can be measured by using a sensor coil.

The impedance of eddy currents is affected by sensor distance, conductivity and permeability of the material, surface defects and input frequency (Fig 4). In addition, the sensor structure will influence the sensitivity of measurement. The sensors can function either by absolute or differential principle. In the absolute method, only one coil is used with possibly a separate induction coil to compensate for the impedance change from coil heating. Absolute sensors are mainly used to find and map widely distributed e.g. corrosion damage and for measuring materials properties. Differential sensors are used for observing local discontinuities such as cracks or pitting damage. ET requires good calibration but has the potential to indicate defect depth at least when the defect density is not overly high. In addition, ET may provide better sensitivity and therefore earlier indications for life assessment of uncoated components than inspections based only on surface techniques.

a)

b)

Fig. 4. a) ET indication from a 1 mm deep crack of a gas turbine blade; b) corresponding indications from EDM calibration notches in the same material.

Ultrasonic testing

Ultrasonic testing (UT) is traditionally even less common for gas turbine inspections than ET, although UT also has the potential to indicate defect depth and is widely applied elsewhere in engineering. Some of the main obstacles for using UT in gas turbines are related to size, geometry, sensor access and applied multiple material systems. However, progress in the UT techniques has made it easier to apply them also in gas turbines.

For best results in ultrasonic immersion testing with focused sensors, the test echo from the target is optimised (defocused). By shifting the sensor closer to the target, the focal point moves deeper into the material. Simultaneously, leaky Rayleigh waves are induced on the material surface. These waves are very sensitive to discontinuities such as cracks, porosity or delaminations close to the surface, up to a depth of about one wavelength which depends on transducer frequency. By using a wide band transducer and sequential filtering of frequencies, the effective sensing depth can be varied to observe the defect structure in the depth direction. The observed signals include a directly refracted longitudinal wave (Ld), wave form change component (transverse to longitudinal wave, S+L), leaky Rayleigh wave (LR) and surface echo (SE).

Fig 5: Leaky Rayleigh wave ultrasonic probes for immersion and contact testing.

Fig 6: Indications from EDM calibration defects (50 m to 500 m ) as detected with leaky Rayleigh wave from a MCrAlY coated vane.

The distance between the directly refracted longitudinal wave and the wave form change component remains constant in relation to the surface echo, and independent of the degree of defocusing in the depth direction. The acoustic properties of the material will determine the locations of these signals in time-amplitude display. Selecting appropriate degree of defocusing will position the leaky Rayleigh wave so that signal interference from these other signals can be avoided.

When contact technique is used instead of immersion, testing in the field is easier but the sensor construction has a fixed degree of defocusing (Fig 5). This requires that the difference in sound velocity between the transducer wedge and the tested material is maximised, so that a small curvature can be applied in the tranducer crystal. The resulting sensor has been shown to be sufficiently sensitive to indicate calibration defects only 50 m in depth (Fig 6). Calibration to known reference defects in a comparable material is of nearly similar importance for UT as for ET.

Discussion

Common NDT methods have been reviewed in brief for applications in gas turbine inspections. Of the overall volume of inspections, visual and penetrant inspections still dominate, as they can be applied in the field faster and with relative ease in comparison with most other common NDT methods. Visual and surface methods have one major drawback, however, as they do not provide a reliable indication of the defect depth. In modern gas turbines where critical hot path components rely on coatings to protect the underlying load bearing metal, a growing defect that has penetrated this barrier will pose a direct threat to the availability of the gas turbine plant. Because there is limited access to inspect the critical components between the maintenance periods and because damage can proceed relatively quickly, good performance of NDT is of obvious importance for reliable and economical operation of gas turbines.

To also obtain an indication of defect depth, eddy current and ultrasonic testing show immediate potential. Both require careful calibration to comparable defects and materials, but some limitations remain so that these techniques are less extensively used in the field inspections. In particular, accessibility of the sensors to tight and intricate details in a gas turbine is often limited. Also, coated multilayer systems are not easily inspected with ET or UT, especially when modern ceramic thermal barrier coatings (TBC) are applied. Finally, coatings in complex internal cooling passages and arrays of tiny cooling holes of modern turbine blading are not easily inspected with traditional NDT methods. Nevertheless, developments in the ET and UT techniques are promising for metallic coated and uncoated systems. Apart for defect depth indication, ET is also likely to provide a better sensitivity and therefore earlier indications for life assessment of uncoated components than inspections based only on surface techniques. Sensor limitations have delayed the parallel developments in UT, but the obstacles do not appear insurmountable.

The difficulties involved in the field inspections may imply that ET and UT techniques will find wider use in the shop applications, for manufacturing, repairs and quality control. However, increasing applications in the field inspections are likely particularly for ET.

Ideally, much or all off-line inspections would be replaced by on-line monitoring, with the potential benefit of much faster reaction to any deviation from expected or desired component behaviour, and to any opportunity that may occur in the market. However, no known monitoring technology is foreseen to have or develop the capability to replace off-line non-destructive inspections entirely.

Conclusions

Nondestructive inspections of gas turbines are an essential part of condition management for gas turbines, because they exhibit relatively high maintenance cost and short in-service inspection cycles, in comparison with most main components of thermal power plants. Since the production capacity of gas fired combined cycle and CHP plants using gas turbines is still increasing, the short maintenance cycles suggest strong growth for NDT services. The most demanding applications of NDT are in the field inspections, particularly because of limited access to tight and intricate details of gas turbine components. The field inspections mainly apply visual and surface techniques. Also ET and UT find increasing use to indicate the flaw depth, which is not obtained in surface inspections. The technical development especially in ET suggest widening applications for metallic surfaces. Simultaneously, new challenges for NDT have emerged e.g. from ceramic coatings, complex coated cooling channels and wide arrays of very small cooling holes. The technical difficulties will also ensure that off-line NDT will remain a major tool for the life management of gas turbines.

Acknowledgements

This work is based on the NDT part within the project on gas turbine materials in a national program on Materials in the Service of Energy Technology (1998-2002). Financial support by the Finnish gas turbine users group and the National Technology Agency is gratefully acknowledged.

References

1. Pitkänen, J., Hakkarainen, T., Kuusinen, P., Lahdenperä, K., Särkiniemi, P., Kemppainen, M. and Pihkakoski, M. "NDT methods for revealing anomalies and defects in gas turbine blade", Insight 9(2001)43, pp. 601-604.

© NDT.net - info@ndt.net

|Top|