134

CHAPTER 7

INVESTIGATION OF BEARING FAILURE

7.1 INTRODUCTION

The rolling element bearings play a crucial role for the proper

functioning of the gearbox used in the WTG. The reliability of bearing in the

special environments like corrosive, high temperature, high power, high speed

and high vacuum zone is very important. Many problems are noticed in the

WTG gearbox during operation even though scheduled maintenance at

stipulated interval is carried out. Oil analysis, drive train alignment and tower

torque are linked to bearing faults. Bearing failure in the wind industries

cannot be tolerated because it leads to catastrophic losses in power production

due to down time, cost of repairing, replacement of parts and so on. A survey

related with research study reveals that the bearing defects and the bearing failure are the main sources of gearbox failure in the wind turbine generator.

Gearboxes are being used to convert high speed and low torque to

high torque and low speed in general. Gearbox in the WTG is used in the

opposite way for converting low speed and high torque from turbine rotor into

low torque and high speed to an electric generator through shrink disc and

fluid coupling. This is accomplished by using large size gears and bearings

than that are being used in a typical machine drives. Generally, gearbox is not

the most likely component to fail in the wind industries rather bearings are the

most common parts susceptible to fail in the rotating machinery including the

wind turbine generator gearbox. In general, machines do not break down or

fail without giving some sign of warning, which is indicated through an

135

increased vibration. It is possible to determine the nature of severity of the

defect thereby predicting the machine’s useful life or failure point by repeated

measurement and analysis of the vibration in the machine, physical inspection

of the gearbox through oil analysis and etc.

7.2 BEARING FAILURE HISTORY IN THE WTG AND ITSCONCERN

A study was carried out to investigate the bearing failure in an

intermediate stage non-drive end of the gearbox used in the WTG. Table 7.1

gives the frequency of bearing failure in the WTG model under investigation.

It is observed from Table 7.1 that the bearing fails within 2 years of usage

against the recommended minimum design life of 10 years. In the event of

such pre-mature failure, it demands immediate replacement of bearing for

getting uninterrupted power supply from the wind turbine. Figure 7.1 shows

the general arrangement of drive train components in the wind turbine

generator. The wind turbine under investigation consist of a speed increasing

gearbox comprising of one planetary stage and two helical stages (Figure 7.2)

to achieve the final speed ratio of 1: 74.917. The position and description of

the bearings at the helical stage of the gearbox is given in Table 7. 2.

Table 7.1 History of bearing failure

WTG commissioning on Date of failure

Down time in days for replacement of an

intermediate bearing 10/12/2004 20/02/2006 1031/03/2006 06/06/2007 731/03/2004 09/01/2009 505/06/2007 05/07/2009 1402/03/2007 07/09/2009 830/03/2006 09/09/2009 1304/04/2006 16/08/2010 9

136

Figure 7.1 Arrangements of drive train components in the wind turbine generator

Figure 7.2 Sectional view of the gearbox

137

Table 7.2 Bearing details

Position Intermediate Stage (IMS) High Speed Stage (HSS)Drive end Non-drive end Drive end Non-drive end

Helicalgear stage

Cylindricalroller bearing

230 NJ EC / C3 D5

Cylindricalroller bearing INA LSL 19 2330 A 0993

72 / V11

Cylindricalroller bearingNU 328 ECJ/

C3

Cylindricalroller bearing

NU 228 ECJ/C3Four point

angular contact ball bearing (QJ328 N2MA/C3)

The overloading of the wind turbine was due to fluctuating wind

force, non-synchronizing of pitching, sudden braking, sudden and frequent

grid drops and all these have induced an undesirable force on the system.

These forces not only damage the bearings, seals and the couplings but also

spoil the gearbox or the generator eventually. Apart from this, the flanks of

both the drive and the non-drive end pinion have pressure marks due to excess

axial movement of the shaft, scuffing wear and pitting wear when wind loads

are high. Besides, any bearing failure in the gearbox while the wind turbine

generator is in operational mode requires huge man hours and machine

stoppage to set right the turbine. Moreover idle of the turbine for service

during high wind season leads to customer dissatisfaction due to loss of

revenue and loss of power generation. Further, bearing fault will harm smooth

functioning of the gear trains. As discussed in section 5.2, if pinion fails at an

intermediate stage then either replacement of that pinion or overhauling the

gearbox itself is very difficult considering the tower height and weight of the

gearbox. The replacement of the gearbox and the lubricant attribute to 38% of

the cost of the turbine parts. Replacing a gearbox in a 1.25mW turbine may

cost more than rupees one crore considering the cost of the gearbox, crane

rental, labour and revenue loss.

138

7.3 PROBLEM PHASE

Filter choke alarm was noticed in the WTG controller on

16th August 2010. On thorough inspection of the 2-stage (50 µm and 10 µm)

filter element (Figure 7.3) enormous metal particles are found in the retainer

plate (Figure 7.4) located at the bottom of the filter. The wind turbine was

stopped immediately and thorough gearbox inspection was done to pin point

the failure location. The study concluded that the failed component was the

intermediate stage non-drive end bearing (Figures 7.5(a) and 7.5(b)). It is also

discovered that the metal particles removed from the bearing got mixed up

with the lubricant and got meshed between the intermediate pinion and the

low speed gear (Figure 7.6). It was observed from site feedback that most of

the wind turbines required significant repair and even complete overhauls

within five to seven years or even before the benchmark time as the wind

turbine downtime is attributed to the gearbox related issues.

The bearing failure occurs very often at the non-drive end of the

intermediate stage. A team comprising of operation, maintenance and product

development engineers has been trying to explore the possibility of designing

heavier bearing and employing new bearing manufacturing process for

trouble free operation of the wind turbine. This chapter is focused to predict

how and why the bearing fails in the wind turbine generator gearbox and to

arrive the remedial measure to minimize the occurrence of failure in future.

139

Figure 7.3 Choked gear oil filter

Figure 7.4 Metal particles at the retainer plate

140

Figure 7.5(a) Failure in intermediate non-drive end bearing

Figure 7.5(b) Failure in intermediate non-drive end bearing

141

Figure 7.6 Metal particles over the gear teeth

7.4 INVESTIGATION ON BEARING FAILURE

7.4.1 Examination of Drive Train Alignment and Oil Analysis

The high speed shaft of the gearbox is coupled with the fluid

coupling with the help of brake disc, an axial damper and the ‘H’ spacer. The

coupling in turn is connected with the generator by means of ‘D’ flange, a key

and a locking screw. The technical team inspected the damaged bearing

(Figures 7.5(a) and 7.5(b)) and predicted that the failures may be due to

over-load by wind force or misalignment between the generator and the

gearbox.

Preliminarily, the current position of the asynchronous generator

shaft relative to the gearbox shaft was examined through laser-optical

alignment technique (Figure 7.7) to conclude the mode of failure, as most of

142

the bearing failures in rotating machinery are attributed to mis-alignment in

the drive train components. A vertical offset misalignment of 0.14mm and

vertical angular misalignment of 0.12 mm /100 mm were found while

inspecting the shaft alignment using easy laser measurement and alignment

system. The generator foot positions were 0.07 mm at the front side and

-0.16mm at the rear end. Similarly, a horizontal offset misalignment of

0.30mm and horizontal angular misalignment of 0.08 mm/100 mm were

noticed. The generator foot positions were 0.22 mm at the front side and

-0.39mm at the rear end. According to the maintenance manager, the

acceptable limit of the horizontal and the vertical offset misalignment is 0.10

mm and the angular misalignment is 0.08mm/100mm and this slight

misalignment will not cause any catastrophic effect on the bearing life and

also on the operation of the gearbox. So it was confirmed from the alignment

study that the failure of the intermediate non-drive end bearing is not because

of the mis-alignment in the system.

Figure 7.7 Drive train alignment in the WTG

143

Further, the gearbox under consideration is lubricated by mineral

oil having a viscosity index of 460 for bearing and gear. The outcome of an

oil analysis of a recent bearing failure location that was conducted in March

2010 is presented in Table 7.3.

Table 7.3 Oil analysis report

Test Protocol Unit Result TAN ASTM D664-01 mg.KOH,/gm. 0.96 Colour ASTM D1500:2007 ASTM 7.5 Flash point (COC) ASTM D92-05a °C. 268Kinematic viscosity (at 40 Deg. C.) ASTM D445-2006 mm2/s. 456

Kinematic viscosity (at 100 Deg. C.) ASTM D445-2006 mm2/s. 29.79

Viscosity index ASTM D2270:2004 _ 97Pentane insolubles ASTM D893-05a % wt. < 0.01 Water content ASTM D6304-04a-Procedure A % wt. 0.016Particle count ISO 4406-1999(E) _ 21/19/15P Q index _ _ 3Elemental analysis ASTM D5185-05 _ _Aluminium, Barium, Boron, Cadmium, Calcium, Chromium, Lead, Magnesium, Manganese, Molybdenum, Nickel, Tin, Titanium and Vanadium

ppm < 1

Copper _ ppm 3Iron _ ppm 12Phosphorous _ ppm 192Silicon _ ppm 13Sodium _ ppm 8Zinc _ ppm 74

The objective of this study is to know the maintained cleanliness

level of the gearbox oil before the bearing failure and also to know whether

there is any abnormality in the oil. During the examination it was found that

the levels of elements Sodium 8 ppm, Silicon 13 ppm and Aluminum less

144

than one ppm which are the agents for formation of oxides were found to be

within acceptable level however it may lead to failure in future. It is obvious

from the Table 7.3 that the obtained cleanliness level as per ISO 4406-1999 e

(1999) indicated by the particle count 4µm, 6 µm and 14 µm was 21/19/15

which denote that the oil is about to reach the contamination stage in the near

future.

7.4.2 Material and Geometry

The preliminary step in failure analysis is material identification.

The bearing under consideration is made up of EN31 steel that has

metallurgical and processing characteristics. The material properties of EN31

are given below:

Modulus of elasticity : 203396 N/mm2

Tensile strength : 2240 N/mm2

Yield strength : 2034 N/mm2

Density : 7833 kg/m3

Hardness : 58 – 63 HRC

The hardness of the inner race, outer race and roller of the bearing

were measured in Rockwell Hardness Tester in 'C' scale and found to be in

the range of 55-57 HRC which is the recommended hardness range for the

bearing materials.

7.4.3 Compositional Analysis

The chemical analysis of the roller and races was carried out using

ARL Spark Analyzer as per OES method to confirm the material specification.

The result obtained from the laboratory is presented in Table 7.4. It is

145

observed from Table 7.4 that the composition of failed bearing confirms to

the specification of EN31material. So it is confirmed from the compositional

analysis that the bearing failure was not because of the problem with the

material.

Table 7.4 Chemical analysis of the bearing material

Element Specification of EN31 material

(Wt%)

Composition of failed bearing

Carbon (C) 0.9 – 1.1 1.02Silicon (Si) 0.15 – 0.30 0.31Manganese (Mn) 0.25 – 0.45 0.43Phosphorus (P) <= 0.025 0.027Sulphur (S) <= 0.025 0.008Nickel (Ni) 0.25 max 0.15Chromium (Cr) 1.3 – 1.6 1.60Molybdenum (Mo) 0.10 max 0.09Vanadium (V) --- 0.012Copper (Cu) 0.25 max 0.18Lead (Pb) 0.002 max 0.016Iron (Fe) 96.5 – 97.32 96.54

7.4.4 Visual Inspection

The visual inspection on the roller revealed that scoring line

representation for contact wear was observed at the middle portion and drive

end portion of the roller. In addition to this, surface and sub-surface fatigue

wear i.e. contact fatigue (spalling) at non-drive end of the roller (Figure 7.8)

on several rollers (Figure 7.9) due to overloading was also found. The same

could be seen on both inner races (Figure 7.10). The damaged outer race of

the bearing is shown in Figure 7.11.

146

Figure 7.8 Wear on roller

Figure 7.9 Damaged rollers

147

Figure 7.10 Wear on bearing inner race

Figure 7.11 Damaged outer race of the bearing

148

7.4.5 Analysis of Temperature and Power

To identify the reason for excessive material removal on the rollers

and races, the peak temperature and power at which the wind turbine (latest

bearing failure location) was operated before the bearing failure were

downloaded from the turbine controller for evaluation. It is obvious from the

temperature and power curve (Figures 7.12 and 7.13) and Table 7.5 that the

instant peak power generation of the turbine has gone up to 2024 kW in the

month of July-2010 and 1853 kW in the month of August-2010 which is

higher than that of the rated power of the gearbox (1250kW). This excess

load on the gearbox and frequent change in wind speed caused overloading on

the bearing resulting in severe damage to the bearings and its elements.

Figure 7.12 Produced power and observed temperature in the month of July 2010

149

Figure 7.13 Produced power and observed temperature in the month of August 2010

The maximum peak power of the turbine and the temperature

observed at various bearing locations in the gearbox for the last two month is

presented in Table 7.5. It is evident from Table 7.5 that the temperature of the

gear oil sump, both drive and non-drive of the high speed bearing and an

intermediate drive end bearing were found on higher side but within the

maximum limit of 80° C and 90°C. Further, it is evident from Table 7.5 that

all the temperature in the month of August 2010 were observed to be higher

by 2 to 3 degree than the temperature in the month of July 2010 even though

the active power is less by 171 kW. This rise in temperature is an indication

of the initiation of the bearing problem in the gearbox.

150

Table 7.5 Month wise peak temperatures and power

Month Activepower (max)

Gear oil sump temp

(max)

HSS drive end

bearing (max)

HSS non- drive end bearing (max)

IMS drive end bearing

(max)

July-10 2024 kW 69° C 76° C 77° C 66° C

August-10 1853 kW 71° C 78° C 79° C 69° C

7.4.6 Gluing and Bearing Clearance Study

Gluing is the process of introducing Loctite 640 or Loctite 641 an

adhesive compound to the surface of the bearing outer race based on the

clearance to avoid relative motion between the bearing and the housing

(Figure 7.14) resulting in the avoiding axial movement of the helical pinion

during operation. Besides, if more axial movement of an intermediate pinion

is noticed then it will either damage the gear tooth or it will lead to loading to

the bearing and its element. In this case, the recommended axial play at an

intermediate stage bearing is 0.07 to 0.10mm. In wind industries, the gluing

is introduced only to the cylindrical roller bearing which comes at an

intermediate (idler) stage as the IMS gearing is to be linked with both the low

speed and the high speed stage gears. Further, during removal of an

intermediate non-drive end bearing for replacement, it was noticed that this

particular gearbox has not undergone the gluing process. It is obvious from

the above study that the lack of gluing may be one of the reasons for bearing

failure in the gearbox.

151

Figure 7.14 Gluing process to avoid relative motion between the bearing and the housing

7.4.7 Metallography Study

The specimen of the failed bearing was prepared by fine polishing

preceded by rough and intermediate polishing and then etched with 2% Nital

solution for examination. The microstructure of the damaged bearing was

analyzed under optical microscope as well as SEM. Tempered martensite

structure with fine carbide (Figure 7.15) is seen on the bearing surface during

micro-structure examination. The specification of the material includes

tempering treatment which was clearly evident from the micro-structure

examination.

152

Figure 7.15 Micro structure of bearing material

7.4.8 Fractography Study

The fractured bearing was cleaned ultrasonically using acetone, and

examined under SEM. The outcome of the investigation is shown in Figures.

7.16 to 7.19. As described in section 6.3.4, no surface preparation and coating

were done on the surface of the bearing components (ASM Handbook 1987)

and an accelerating voltage of 25 kV was applied during the SEM analysis.

Wear debris and corrosion products (Figure 7.16) were seen on the surface of

the rollers as well as on the races of the bearing due to the presence of some

kind of corrosive and wear elements in the gearbox oil. The mode of failure of

the bearing is fatigue and cleavage fracture (Figures 7.17 and 7.18).

Figure 7.19 shows the SEM image of the bearing with high cyclic fatigue and

brittle fracture. Further, there was no evidence of cracks and micro-cracks on

the bearing surfaces during the study.

153

Figure 7.16 SEM image of the bearing showing the wear debris and corrosion products at X50 magnification factor

Figure 7.17 SEM view of the bearing showing fatigue and cleavage fracture at X27 magnification factor

154

Figure 7.18 SEM view of the bearing showing cleavage fracture observed at X50 magnification factor

Figure 7.19 SEM view of the bearing indicating high cycle fatigue and brittle fracture observed at X50 magnification factor

155

7.4.9 Energy Dispersive X-ray Spectroscopy Analysis

The damaged bearing was examined under Energy Dispersive

X-ray Spectrometer (EDS) and the outcome of the investigation is illustrated

in Figures 7.20 and 7.21. An accelerating voltage of 28 kV and probe current

of 1.00000 nA were applied during the examination. In the EDS report, the

horizontal and vertical axes represent the energy range (0 to 20 keV) and the

counting rate (max 3308 cps) respectively. The bauxite elements observed in

the first sample (Figure 7.20) were Sodium (Na), Aluminium (Al) and Silicon

(Si) . The percentage of mass of Silica was prominent (16.02 %) at 1.739 keV

followed by Alumina (8.17 %) at 1.486 keV and Sodium 2.22 % at

1.041 keV.

Similarly, there were three bauxite elements viz, Sodium,

Aluminium and Silicon were observed in the second sampling (Figure 7.21).

In this case, the percentage of mass of the Silicon was 3.20 % at 1.739 keV

energy range. It was evident from the EDS study that the reason for the

bearing failure is due to the presence of bauxite particles in the gearbox oil.

The source for the Silicon (bauxite element) in the gearbox could be either

due to falling down of Silica gels from the breather which was mounted over

the gearbox for breathing or due to the damage of the breather. The bauxite

element in the gearbox oil resulted in excessive material removal on the

rollers as well as on the races of the bearing during peak power generation.

156

Figure 7.20 EDS report of the bearing for first sample

157

Figure 7.21 EDS report of the bearing for second sample

158

7.5 RESULTS AND DISCUSSION

The visual observation and the microstructure examination

confirmed that an excessive damage of the roller as well as the surface of the

race is primarily due to the contact wear resulting in scoring line followed by

spalling because of surface and sub-surface fatigue wear. The results of the

chemical analysis (Table 7.4) revealed that the bearing material is in

accordance with EN31 specifications. It is learnt from the research study that

the bearing failure had happened due to high cyclic fatigue and brittle fracture

(Figure 7.19). The Gluing study predicts that the failure of the bearing may be

due to relative motion between the bearing and the housing. Further, the

source for high percentage Silicon Oxide in the gearbox oil is due to either

from the damage of the breather or drop down of the Silica Gels from the

breather which is normally mounted over the gearbox for breathing.

7.6 CONCLUSION

The major conclusions drawn from the above analysis are given

below:

i) The debris collected at the filter unit (Sections 7.2 and 7.3)

was resulted from the contact wear (Scoring) followed by

surface and sub-surface wear (Spalling) on rollers and races of

an intermediate non-drive end bearing which have lead to

breakdown of the WTG.

ii) The chemical and metallographic examination has revealed

that the root cause of failure of bearing is not because of the

defect in the material and the heat treatment process.

iii) The metal particles of the damaged bearing got mixed up with

the lubricant and entered in to the filter unit which caused the

159

filter choke alarm in the wind turbine controller. Apart from

this, tiny metal particles might have either escaped from the

filter unit or directly from the failure bearing area which

entered the gearbox and got meshed between the gearing.

iv) The drive train alignment study does not prove that the failure

of bearing is due to the misalignment in the system.

v) The oil analysis shows that the oil was about to contaminate

and hence the failure was because of contamination of the

gearbox oil.

vi) The experimental investigation has revealed that the failure of

the bearing is due to the presence of bauxite elements in the

gearbox oil which had triggered an intermediate stage bearing

and the pinion failure.

vii) The study on temperature and power proved that the pre-

mature failure of the bearing is due to overloading caused by

the continuous peak power generation of the wind turbine

during high wind season.

viii) The investigation on clearance and gluing showed that the

failure of the bearing is due to excess clearance between the

bearing (outer race) and the housing bore due to lack of

gluing.

ix) Visual inspection and the SEM study have confirmed that the

nature of the bearing failure is surface fatigue due to high

cyclic fatigue phenomenon.