Failure Analysis of High-Pressure Turbine Blades in Steam Power Plants

Introduction

In a power facility, steam turbine blades transform the linear motion of high-temperature, high-pressure steam into the rotary motion of the turbine shaft [1,2] . If the blades of the turbine fail, the power facility will perish. This has the potential to cause economic losses if the disruption lasts for an extended period of time. To improve the reliability of the turbine system, it is necessary to analyse the failure of the turbine blades. Previous research has demonstrated that creep damage, fatigue, corrosion, erosion, and environmental attack (oxidation, heat corrosion, erosion, and foreign matter damage) are the most prevalent failure mechanisms for turbine blades [3]. According to Cahyadi, M., the most common causes of injury to steam turbine blades are stress corrosion, deposits, water erosion, and fatigue [4]. While Kubiak Sz, J. et al. noted that turbine blade root failure typically occurs in high-stress concentration zones as a result of structural or material geometric discontinuities [5], this is not universally the case.

This electric steam power facility with a capacity of 2 x 135 MW has been operational since 2015. From the start of operation of the electric steam power plant until 2021, several maintenance activities have been performed, including Serious Inspection (SI) maintenance. During the Unit 1 Serious Inspection, it was discovered that the ninth-stage high-pressure turbine blade had failed. Visual inspection and nondestructive testing are performed for this project. To determine the failure mechanism of the blades, mechanical properties, metallographic examination, microhardness scanning, and energy spectrum testing were also performed on fractures.

Experimental Work

In this paper, the damaged 9th stage turbine blade (GB/T 8732) from a 135 MW gas turbine functioning at 535 °C is investigated. Figure 1 depicts a damaged turbine blade and analysis sampling. Beginning with samples in acceptable condition (samples 1 and 2), damaged turbine blade samples (samples 3 to 16) are then collected. Table 1 provides the chemical composition of the used material and the standard material. The ninth-stage turbine blade was utilized to investigate and evaluate the spectrum of chemical composition, metallographic analysis, micro-hardness test, and microscopic analysis of failure. Using an optical microscope (Olympus-GX51, Japan), metallic structures and inclusions were observed. The microhardness of the sample was measured using a microhardness instrument calibrated to the ASTM E384-11 standard. In the meantime, a Scanning Electron Microscope – Energy Dispersive X-Ray Spectroscopy (SEM-EDS, Hitachi FlexSEM 100, Japan) and local chemical analysis of the turbine material were used to investigate damage to turbine blade fractures.

Figure 1.Figure 1. The general view of damage in the gas turbine blades (a) damage to the ninth stage turbine blade (b) sampling position from the root of the gas turbine blade (c) deformation of high-pressure steam-turbine diaphragm (d) sampling positions 1 and 2 (in good condition)

Results and Analysis

Operation Data

From the trend of turbine operating parameter data, including the correlation of power, pressure, flow main steam, and vibration during the month preceding the discovery of turbine blade failure during "operational hours" The vibration is still within the normal range, with a value of 98.62 m, based on the results of observations of the tendency of the turbine operating parameters (Figure 2). In addition, one day after event number 2, the seismic trend increased to 107.72 m. At the third event (after four days), the vibration level exceeded the alarm threshold of 181.03 m dan shaft vibration was in the Trip category. The shaft vibration alarm limit for Steam Power Plants is 127m per ISO 10816-2 (Mechanical vibration) for steam turbines and generators, while the trip limit is 250m. In addition, a CSI 2600 vibration analyzer was used to measure vibrations and determine the dominant spectrum.

Figure 2.Figure 2. Trending operations of turbine parameters.

Based on data collected using a CSI 2600 vibration analyzer with the shaft measurement method using a displacement sensor in the X and Y dimensions, it can be seen that the alarm vibration limit for the ISO 10816-2 steam turbine and generator is 127m and the Trip limit is 250m (Figure 3(a)). In the meantime, based on shaft vibration data with an elliptical orbit frequency order and complete spectrum for 1X, the displacement value reaches nearly 250 microns, and for 1Y, it reaches nearly 200 microns. Figure 3 (b) demonstrates that the dominant 1X order spectrum on bearings indicates unbalanced, bending, or lengthy rotors.

Figure 3.Figure 3. (a) Shaft vibration measurement results using a displacement sensor in the X and Y directions, (b) shaft vibration data order frequency elliptical orbit shape and full spectrum

Macroscopic inspection

According to the results of macro-observations conducted with a digital camera, the fractures that occur in various high-pressure turbine blade specimens are quite diverse. In general, faults are distinguished by the morphology of the fracture, beginning with 1) crack initiation in the fillet rood land area, 2) crack propagation, and 3) final fracture [6]. These three stages of failure indicate that the blade has fractured due to fatigue loading. This is evident in Figure 4 for samples 8 and 12, where multiple initial cracks can be observed. Sample 12 demonstrates that the ultimate fracture area varies between samples, the second significant phenomenon. Therefore, it is suspected that there are at least two types of applied tension that cause blade fracture. According to macro-observations, pitting corrosion was also detected on the root zone depicted in Figure 5.

Figure 4.Figure 4. Macroscopic appearance of the damaged blades.

Figure 5.Figure 5. Macroscopic appearance of the blade showing pitting corrosion

Chemical Composition

The chemical composition of the blades is analyzed using optical emission spectroscopy (OES). Chemical analysis reveals that the blade material is steam turbine blade steel. Table 1 summarises the quantitative chemical composition of the specimens relative to the standard Chinese steam turbine blade material (GB/T8732-2004). Based on Table 1's data, it is evident that the chemical composition of the specimens and the material composition of the Chinese standard steam turbine blades differ.

Microstructural Evaluation

The data from microstructure observations are tabulated to facilitate identification and facilitate data collection. Multiple samples, sampling points, and locations were observed for their microstructure. According to the microscopic observations presented in Table 2, the microstructure is tempered martensite in which periodization has begun to develop. The microstructure of martensitic stainless steel is martensite that has been quenched and tempered during the blade production process [7]. While the spheroid can be accelerated by operating at high temperatures, as depicted in Figure 2 of Table 2, the microstructure label 1A.3 has a carbide dimension that is, on average, smaller than the microstructure of the sample blade, 15.

Fractographic analysis

Observations of turbine blade fractures using SEM and EDS were conducted on uncleaned and cleaned samples. A root blade fracturing sample (Figure 4 sample 14) was subjected to spectral analysis, which identified oxygen, silicon, sodium, chromium, and calcium as deposit products in the turbine system (Figure 6). It appears that this turbine blade sample was fractured prior to its collection. If confirmed by macro observation, it is likely that the blades are not broken at the same moment. On the cleansed samples (Figure 4 sample 8), further spectrum analysis revealed that the oxide layer and deposits could be eliminated. In addition, optical emission spectroscopy (OES) testing confirmed that the sample's Cr content was close to the value of the initial material composition of the blade (Figure 7).

Figure 6.Figure 6. The spectrum of the EDS test results on sample 14 root blades without the cleaning process

Figure 7.Figure 7. The spectrum of EDS test results on sample 8 root blade after the cleaning process.

In addition, SEM observations were performed to ascertain fracture patterns after the sample was cleaned. Observations made prior to cleaning the sample were obscured by oxide, so they could not be seen clearly. This is evident in the image of sample 14's root blade fracture, which is covered in oxide (Figure 8 (a)), whereas sample 8's cleaned root blade fracture sample (Figure 8 (b)) demonstrates fissure initiation.

Figure 8.Figure 8. (a) SEM image of the root blade fracture of sample 14 covered with oxide (b) SEM image of the root blade fracture of sample 8 with crack initiation.

Increasing the microstructure image's magnification will provide more specific information. This was performed on sample 16 in which fracture initiation was discovered (Figure 9(a)). After increasing the magnification of the root blade fracture in sample 16, a transgranular fracture was observed (Figure 9(b)). Continued with a higher magnification on the root blade fracture of sample number 16 (Figure 9(c)). According to the obtained SEM images, fractures occur at grain boundaries, and there is precipitate between grains.

Figure 9.Figure 9. (a) Image of the root blade fracture of sample 16 with crack initiation (b) SEM image of the root blade fracture of sample 16 which indicates a transgranular fracture (c) SEM image of the fracture at the grain boundaries and the presence of precipitates between the grains of sample 16.

Figure 10.Figure 10. Directions and points of hardness testing on samples 1A, 14 and 15

Hard test results data

Figure 10 depicts the directions and test sites for the hardness tests conducted on samples 1A (refers to table 2), 14, and 15, after which the microhardness test results are tabulated.

insert figure

Based on Table 3 data, the average hardness ranges from 243 to 268 HV, which falls within the range of Chinese National Standards for steel blades for steam turbines (225-260 HV) [8]. Figure 11 depicts the distribution of hardness values for samples 1A, 14, and 15 based on the test sites and the direction of the X-Y axis. The location of the posterior root blade has the lowest hardness value, as depicted in Figure 11. This is consistent with the root blade fault pattern in which the fracture propagation fault occurs in the area of the blade that is the weakest.

Figure 11.Figure 11. Results of hardness measurements (a) Y-axis hardness distribution for samples 1A, 14, and 15 (b) X-axis hardness distribution for samples 14 and 15.

Discussion

Fishbone diagram analysis is used to analyze the fundamental causes of problems by identifying potential causes based on results and measurement data, discussions, field findings, and references. Root Cause Failure Analysis [9] is a series of logical steps that guide observations through a process that separates facts that include an activity or failure.

Figure 12.Figure 12. Fishbone diagram causes of high-pressure turbine blade fracture failure

According to the study's findings, six main groups are responsible for high-pressure turbine blade fractures;

a) Variations in turbine steam flow and pressure that are always accompanied by a rise in shaft vibration fluctuations.

b) The vibration spectrum analyzer CSI 2600 shaft data reveals an unbalanced bearing and warping.

c) The maintenance findings are not followed up on.

d) The chemical composition of high-pressure turbine blades differs from the China National Standard GB 8732 material standard for steel blades for steam turbines.

e) Based on metallographic and fractographic observations, it was determined:

· Observations of macro samples indicate that the fracture is of the "fatigue failure" type, in which failure begins with the initiation of a crack in the fillet section, followed by gradual crack propagation, and ultimately the final fracture.

· The discovery of crack initiation on multiple blades led to the conclusion that there was an indication and role of corrosion in the fillet region (stress corrosion in the fillet region).

· This is one method of microstructural degradation of this type of steel based on the microscopic observation that the sample material is of the modified martensite type where carbide spherization is beginning to appear.

· The EDS test reveals that the fracture surface is covered with oxides and deposits (one of which is silica), indicating that root blade fractures do not occur randomly but rather over time.

· The SEM examination confirmed the presence of multiple fracture initiations in the broken samples. Increasing magnification reveals that the sample is transgranular fractured [10] , indicating that the mechanical and material factors outweigh the corrosion factor.

f) The turbine blade hardness test yielded an average value of 243-268 HV, which falls within the standard range for the material, which is 225-260 HV. However, there is a soft area on the root blade, which corresponds to the location where fracture propagation occurs.

Conclusions and recommendations

The conclusions from the analysis of turbine blade fracture failure analysis are as follows:

• According to turbine operating data, fluctuations in turbine steam flow and pressure were always accompanied by a rise in shaft vibration fluctuations.

• During maintenance, friction with the turbine blade caused deformation of the fixed blade diaphragm/stator HP turbine blade stage 9/steam flow guide and damage to the stage 9 diaphragm gland sealing ring.

• Macro observation reveals that the fracture is a result of fatigue, and on some blades, corrosion may have played a role (stress corrosion in the fillet area).

• SEM-EDS results indicate that the surface of the fracture is loaded with oxides and deposits, indicating that root blade fractures do not occur randomly but rather over time.

• Fatigue is the failure mechanism of the high-pressure turbine blade fracture, with the degradation of the blade material facilitating the initiation and propagation of fissures.

The suggestions/recommendations that can be given include:

• Required further testing of corrosion because there is a crack initiation on the fault side.

• Deposits in the steam turbine require additional testing to ascertain the quality of the steam entering the turbine.

• Each time a turbine is taken out of service, an NDT examination is performed in response to the observation of fracture initiation on the root blade.

Acknowwledgements

This research is supported by PT. PJB Services, Operation & Maintenance

References

1. Ziegler D., Puccinelli M., Bergallo B., Picasso A.. Investigation of turbine blade failure in a thermal power plant. Case Studies in Engineering Failure Analysis. 2013; 1(3)DOI
2. Saxena Sanjeev, Pandey J.P., Solanki Ranjit Singh, Gupta Gaurav K., Modi O.P.. Coupled mechanical, metallurgical and FEM based failure investigation of steam turbine blade. Engineering Failure Analysis. 2015; 52DOI
3. Kolagar A.M., Tabrizi N., Cheraghzadeh M., Shahriari M.S.. Failure analysis of gas turbine first stage blade made of nickel-based superalloy. Case Studies in Engineering Failure Analysis. 2017; 8DOI
4. Cahyadi Muhamad. ANALISA KERUSAKAN SUDU TURBIN UAP MATERIAL MARTENSITIC STAINLESS STEEL 410 PADA STAGE TERAKHIR. Jurnal Teknik Mesin Cakram. 2020; 2(2)DOI
5. Kubiak Sz J., Segura J.A., Gonzalez R G., García J.C., Sierra E F., Nebradt G J., Rodriguez J.A.. Failure analysis of the 350MW steam turbine blade root. Engineering Failure Analysis. 2009; 16(4)DOI
6. Fatigue and Fracture: Understanding the Basics. ASM International; 2012.
7. ChineseStandard. GB/T 8732-2004. 2004.
8. Manurung Vuko A T , Wibowo Y.T.J. , Baskoro S. Y. Panduan Metalografi,. 2020.
9. Keith Mobley. Root Cause Failure Analysis. Elsevier; 1999. DOI
10. William D. Callister Jr., David G. Rethwisch. Materials Science and Engineering: An Introduction. Wiley; 2018.