Gearbox condition monitoring procedures

tickbitewryMechanics

Oct 30, 2013 (3 years and 10 months ago)

66 views

Gearbox condition monitoring procedures


Walter Bartelmus
1
,
Radoslaw Zimroz

2

1
,2

Diagnostics and Vibro
-
Acoustics Science Laboratory, Pl Teatralny

2, 50
-
051 Wroclaw,
Wroclaw University of Technology, PL

{walter.bartelmus, radoslaw.zimroz}@pwr.wroc.pl

Abstract

There is a need to treat a gearbox as a subsystem which consist
s

of several el
e-
ments like gears, bearings and shafts incorporated into a box
. The gearbox is i
n-
corporated into a system of a drive, for example an electric motor, and a driven
machine. When a system is in operation the mention elements interacts each other.
When preparing condition
monitoring
method one has for disposal results of

the
research directed to evaluation of isolated faults like a tooth crack, tooth breakage,
pitting, scuffing, misalignment and so on. It is taken in advance that only one of
the mentioned faults occurs in the system. The research is done at the condition
of
constant load or constant rotation speed. There is also done research under cond
i-
tion of different constant levels of the load and rotation speed. The scenario of
degradation process of gearboxes is that only one fault occurs and developed in
the system
. In real gearbox systems many different scenarios of degradation pr
o-
cess may occur. The presented paper will show gearbox condition monitoring pr
o-
cedures, which is equivalent to different scenarios. There are given some steps,
which ought be considered wh
en gearbox condition monitoring procedures are d
e-
veloped. Having in mined these steps one can control the gearbox degradation
process extending a gearbox live and reducing maintenance cost, creating, what is
called, the failure prevention technology. For p
resenting this paper stimulates us
also the papers presented in the M
echanical
S
ystem and
S
ignal
P
rocessing

“Sp
e-
cial Issue” on the “Condition monitoring of machines in non
-
stationary oper
a-
tions” where some authors have tendency of treating the machine as t
he collection
of separate not connected elements in the same way as it was stated before
.

Keywords:
gearbox,
condition monitoring, design factors, operation factors

1

Introduction

Vast literature on condition monitoring for example [1] to [3] do not gives the
diagnostics procedures for gearbox condition monitoring as one think a gearbox as
2


Walter Bartelmus and Radoslaw Zimroz

a system, which consists of many elements like: bearings, gears, shafts and there is
interacti
on between these elements, when the system is in operation. For develo
p-
ing the suitable gearbox condition monitoring procedures there is a need to deve
l-
op some steps:

o

First step, one ought to understand the relation between factors having influence
diagno
stic signals and condition of gearboxes as it is given in [4] and [5].

o

Second step, one ought to know information on relation between the factors,
which have influence on vibration diagnostic signal and diagnostic signal
presentation in the form of spectra

as it is given in papers [6] and [7].

o

Third step, one ought to understand the different form of gearbox degradation
processes as it is given [8] to [16].

On the bases of all the mentioned papers there is possibility do develop the
gearbox condition monito
ring procedures.

2

Relation between factors having influence diagnostic signals
and condition of gearboxes

The factors [5] are divided as it is given in Fig.1. The primary factors are divi
d-
ed as the design and production technology.



Fig.1 General division of factors influencing vibration signals

The secondary factors are divided by operation and change of condition. The
design factors can be determined on a design specification presented in the form
of design drawings where all d
etails of diagnosed object are presented. The pr
o-
duction technology factors are resulted from production that means, parts machi
n-
ing and their assembling. They are imbedded into the product and they reveal i
t-
self during object/ machine operation. Because t
he operation condition change
during operation so they should be taken into consideration. The design, produ
c-
tion technology factors determine the reference machine condition. In this paper
the reference machine condition is determined by the load suscepti
bility characte
r-
istics (LSC) [11] and [12]. It is the load trending of diagnostic features at the cu
r-
rent condition for a given stage of a gearbox. The LSC is the result of regression
analysis, presented later. The operation factors are determined by outer

load vari
a-
tion, which may be connected with excavation process in mining industry or var
y-
Gear box condition monitoring procedures

3

ing wind power in a wind turbine. The change of condition of gearboxes is dete
r-
mined by all forms of faults, which occur during gearbox operation. In the Fig.2
[10] t
here is given the scheme of the system where an interaction between comp
o-
nents in multistage gearbox system and external components to the gearbox such
as: electric motor/engine, damping coupling, external load are presented. The pr
e-
sented system operates
in certain environment (temperature, humidity, dustiness),
which has influence on degradation


change of a gearbox condition. Having in
mined all the mentioned factors one can control the gearbox degradation process
[8], [13, [14] extending a gearbox live

and reducing maintenance cost creating,
what is called, the failure prevention technology.


Fig.2 Interaction between components in multistage gearbox system and external comp
o-
nents to the gearbox such as: electric motor/engine, damping coupling,
external load

3

Diagnostic signal presentation in the form of spectra

Fig.3 shows two driving systems for the bucked wheel of a bucked wheel excav
a-
tor. In both systems there is used a planetary gearbox. In the system given in
Fig.3a) the planetary gearbox i
s characterised by standstill rim (gear z
3
) and rota
t-
ing sun (gear z
1
). The planetary gearbox in system Fig.3b) the rim (gear z
5
) and
the sun (gear z
3
) are rotating. Equivalently their ratios are given in (1) and (2) .





(1)





(2)

These two systems are presented for comparison to show the difference in the
design of the system and the difference in the rat ios of the planetary gearboxes.
Further consideration is directed to the system given in Fig.3b. F
or planetary gea
r-
4


Walter Bartelmus and Radoslaw Zimroz

box condition monitoring there is a need to calculate a meshing frequency, more
about the meshing frequencies calculations look into [6] and [7]. According the
principle given there the meshing frequency for the planetary gearbox incorpora
t-
ing into the system Fig.3b is given by the statement (3). More details of the system
presenting in Fig.3b is given in Fig.3c and Fig.3d, so the meshing frequency for
the planetary gearbox in system Fig.3b is





(3)

where:

n
2j



related speed rotation of a shaft which rotates with the speed rotation
n
2

[RPM]
,

n
2



absolute speed of the second shaft [RPM],

n
j



arm/carrier speed rotation [RPM],

z
3



number of teeth in gear 3.


To use the above statement one ought have more st
atements which are connected
with the gearbox system given in Fig.3b. The complete ratio of the system is



(4)

The bevel stage ratio equals to




(5)

The planetary gearbox rat io equals to as given in (2). The
cylindrical stage gear
ratio equals to




(6)

The arm/carrier speed rotation



(7).

For the gearbox condition monitoring there also needed other frequencies co
n-
nected with gearing local faults ([6] and [7]) and
roller elements bearings faults.

Gear box condition monitoring procedures

5

a) b)


c) d)


Fig.3 Two schemes a) and b) of different designs of driving systems, c) and d) more details
for driving system b), in d) bucked wheel is neglected

6


Walter Bartelmus and Radoslaw Zimroz

4

Different form of gearbox degradation processes

As it is given in [15] when considering a gearbox failure on
e should take into
consideration primary and secondary misalignment of shafts which are in a gea
r-
box system. The presented paper is mainly concentrated on inner gearbox shafts
and gears misalignment (IGSGM). The IGSGM can be evaluated during gearbox
operat
ion on the base of vibration signal and operation load of a gearbox. On the
base of vibration and load characteristic presented in two dimension space the r
e-
gression line can be evaluated. This regression line is called the susceptibility
characteristic (S
C) [11], load trending. The primary cause of misalignment can be
assessed at the beginning of a gearbox operation by SC, load trending. The prim
a-
ry cause of misalignment evaluated by SC is the measure of gearbox quality. The
inner secondary cause of misal
ignment is developed during a gearbox operation
in the process of bearing wear. In the paper [11] it is discussed the issue of the
cause of a gearbox secondary misalignment. It is described the case of planetary
gearbox condition evaluation where is giving

the evidence that the frictional wear
of bearings cause IGSGM. The IGSGM is evaluated on the base of SC and bea
r-
ing frictional wear is measured after the planetary gearbox dis mantling and tooth
gearing faults as surface micro cracking has been noticed.

So
me discussions on gearbox root cause analysis is given in [8]. Here is given a
case describing the root cause analysis (RCA) of a bevel gears, which shows that
the reason of gear condition change is IGSGM. The RCA is based on the
knowledge gained from expe
riment presented in [11]. The bevel gear (Fig.4)
shows developed scuffing ( there is more on surface distress in [16]) and a tooth
gear fat igue development, and heat tints. The heat tints show that the surface bulk
temperature would raise to about 250
o
C.
It was also inferred that the tooth brea
k-
age is the fourth cause a tooth gear degradation process. The primary cause is mi
s-
alignment, second scuffing, third fatigue crack (details in figure 5) , fourth brittle
tooth breakage. Fig.6 shows a scheme of fatigu
e fracture area with beach marks
and brittle fracture area. These two considered cases shows that the main cause of
gearbox failure is IGSGM. At the end ought to be underlined that primary cause of
misalignment, which comes for example from misaligned moto
r and gearbox shaft
position should be eliminated before putting system into operation according the
technologies given in [17].

Gear box condition monitoring procedures

7


Fig.4 Heavily effected/destructed gear teeth, and colours heat tint


Fig.5 View of damaged gear, fatigue
fracture area with beach marks, brittle
fracture area

Fig.6 Scheme of fatigue fractu
re area
with beach marks and brittle fracture
area

To understand the increase of bulk temperature of gear teeth some computer
simulation study is given here, more theory is given in [4]. The reason of i
n-
creased temperature is

the power lost during teeth friction.

The friction power is equal to a product of the slipping velocity and the friction
force. The slipping velocity can be calculated as an absolute value of the diffe
r-
ence between the tangent velocities to the gear teet
h.




(8)

The instantaneous friction lost power is




(9)

Plots of instantaneous friction loss are shown in Fig.7 and 8 for different fri
c-
tion coefficient values: μ = 0.02 and 0.1.

8


Walter Bartelmus and Radoslaw Zimroz


Fig.7 Plot of power of system losses due to friction in meshing for


=0.02:1


4,different
periodsofsstemrun:1accelerationofthesstem,2freerun,3loadingofthesstem,4
rununderconstantload.lotofinstantaneousefficiencvaluesfo



=0.02.

Instantaneous efficiency value is defined as




(10)

where:
N
s



the motor’s power, W;


N
lost



the power of the losses due to friction in the meshing, W.


Fig.8 Plot of power of system losses due to friction in meshing for


=0.02:1


4,different
periodsofsstemrun:1accelerationofthesstem,2freerun,loadingofthesstem,4run
underconstantload.lotofinstantaneousefficiencvaluesfor


=0.1.

Plots of the above parameter when the gear is properly lubricated and the fri
c-
tional resistance is defined by friction coefficient

= 0.02, Fig.7. For compar
i-
son, plots of the power of the losses and efficiency plots for

= 0.1 are shown in
Fig.8. As it follows from Fig.7, efficiency is in a range of 0.098
-
1 and at

= 0.1
it ranges from 0.98 to 1 (Figure.8). A comparison of Figures 7 and 8 shows that
for steady running (period 4) at
m varying from 0.02 to 0.1, i.e. at a fivefold i
n-
crease in the value of the frict ion factor, the maximum instantaneous losses are r
e-
spectively Nl
lost

= 38 W and 195 W and hence the loss ratio is 195/38 = 5. Thus
Gear box condition monitoring procedures

9

the fivefold increase in the friction factor

corresponds to the fivefold increase in
the power of the frictional losses. For now has been given comparison when a gear
run in perfect condition

= 0.02 and at oil dry frict ion

= 0.1. It is dry fri
c-
tion for fine roughness condition. When gear surface is in the scuffing condition
Fig.4 the friction coefficient may arise to

= 0.4 even more. In the case of

=
0.4 using linear relat ion which govern the above c
onsideration an instantaneous
efficiency (10)
= 0.245. This support the idea that surface bulk temperature a
s-
sessed on heat tints Fig.4 can be high, and in considered case is about 250
o
C

5

Load susceptibility characteristics as a meas
ure of gearbox
condition

Load susceptibility characteristics according to [11] and [12] are the measure of
gearbox condition. The load susceptibility characteristics can be expressed by
linier relation [11]

y=ax+b


(11)

where y


the value of the signal
feature,

x


the operating conditions (instantaneous speed in this case) indicator,

a, b


the parameters to be determined.

a) b)


Fig.9 a) Data distribution of measured diagnostic parameters b) Load
susceptibility chara
c-
teristics as diagnostic features for planetary gearbox as a function of rotation speed RPM;
for a gearbox in good (“o” dots) and bad condition (“x” dots)

Here is described the case of planetary gearbox condition evaluation where is
giv
ing the evidence that the frict ional wear of bearings cause IGSGM. The
IGSGM is evaluated on the base of SC and bearing wear is measured after the
planetary gearbox (Fig.3b) dis mantling and tooth gearing faults as surface micro
cracking has been noticed. I
n a paper [11] the term load susceptibility (LS) or SC
is introduced. The load susceptibility is given by the regression characteristics in
Fig.9b) as diagnostic features for planetary gearbox as a function of rotation speed
10


Walter Bartelmus and Radoslaw Zimroz

RPM; for a gearbox in good (“o”

dots) and bad condition (“x” dots). In this case,
presentation of susceptibility characteristics for an electric motor is based on a li
n-
ear relat ionship between the transmitted moment and rotation speed. It means that
one may use the load susceptibility c
haracteristics as the function of a load or
function of a rotation speed RPM as it is given in Fig.9b). In the presented cases of
the load susceptibilities the characteristics are interpreted as follows. The case for
a good condition of gearbox shows that
planetary gearbox behaves as a linear sy
s-
tem under increasing load, that means with increasing load the system elements
deflection increases, causing linear increase of diagnostic feature, which is dyna
m-
ic inter teeth force related. In the case of bad con
dition as result of frictional wear
of bearings the gear mesh under the condition of an IGSGM what gives a linear
increase of the gear cooperation error and linear increase of inter teeth force,
which cause linear increase of a vibration acceleration sign
al as it is presented by
linear regression line Fig.9b in a case of the bad gear condition. Fig.9b) also
shows very good separation of data for the good and bad condition. Better than is
given in Fig.9a) when the data distribution functions overlap each ot
her.

6

Proactive and failure prevention technology and
concussions

In Fig.10 there is given modified from [18] basic scheme of elements of vibr
a-
tion gearbox diagnostic method and elements of failure prevention technology.


Fig.8 Modified basic scheme of ele
ments of vibration gearbox diagnostic method and el
e-
ments of failure prevention technology

Gear box condition monitoring procedures

11

In the scheme Fig.8 there are considered several elements as:

Factors influencing vibration signal, which are divided as is given in the intr
o-
duction into four group
s namely: design, production technology, operation, cond
i-
tion change.

Interaction of gearbox elements as is given in Fig.2 and environment influence
should be considered.

On the basis of knowledge on a gearbox degradation process [8], [13


16],
which c
omes from the factors analysis, interaction of gearbox elements, enviro
n-
ment influence different degradation scenarios can be described.

Taking also into consideration knowledge on a gearbox degradation process
[8], [13


16] choice of proactive measures o
ught be used as like primary and se
c-
ondary misalignment assessment, oil particles and water content assessment. For
the secondary misalignment the load susceptibility characteristics should be used.

In degradation scenarios there one ought to take into co
nsideration order of
possible events, which describe: primary unbalance and misalignment, increase of
rolling element bearing backlash and secondary misalignment. Teeth or tooth pi
t-
ting, scuffing, fracture, breakage as it is described in the chapter 4.

On

the base of primary misalignment evaluation proactive measure should be
under taken eliminating some possible misalignment according to [17].

On the base of secondary misalignment evaluation proactive measure should be
under taken, which goes to rolling element bearings replacement.

On the base of oil particles and water content assessment oil should be purified
or replaced to reduce a gearbox
elements wear.

For signal analysis should be used: vibration spectrum, t ime frequency spectr
o-
gram, envelope analysis, spectral kurtosis, cyclostationary analysis and used the
results of these analysis for

presenting them in the form of

load susceptibility
characteristics.

Inferring on gear condition should be made during a gearbox operation from its
beginning and on the base of different signal analysis make suggestions on use
proactive measures, to reduce cost of maintenance.

6

Acknowledgements

This pape
r was financially supported by Polish State Committee for Scientific
research 2010
-
2013 as research project NN 504147838.

7

References

[1] Neville W, Sachs P E (2007) Practical plant failure analysis. Taylor & Francis

[2] Hatangadi A A (2005) Plant and
machinery failure prevention McGraw
-
Hill

12


Walter Bartelmus and Radoslaw Zimroz

[3] Gridkhor P,. Scheffer C (2004) Practical machinery vibration analysis and predictive
maintenance. Elsevier

[4] Bartelmus W 2006 Condition monitoring of open cast mining machinery. Oficyna
Wydawnicza Politechni
ki Wroc
ł
awskiej Wroc
ł
aw

[5] Bartelmus W (2009) Influence of design, production technology, operation and change
of condition factors on gearbox vibration diagnostic signals. Proceedings, Resent D
e-
velopments in Acoustic, Noise and Vibration. The Sixteenth
International Congress on
Sound and Vibration Kraków, Poland, 5


9 July

[6] Bartelmus W,. Zimroz R (2011) Vibration spectra characteristics frequencies for cond
i-
tion monitoring of mine machinery compound and complex gearboxes. Scientific Papers
of the In
stitute of Mining of the Wroclaw University of Technology, Mining and geol
o-
gy XVI Wroclaw : 17
-

34

[7] Brtelmus W (2011) Fundamentals for condition monitoring and diagnostics for driving
bucket wheel system with overload mechanism of bucket wheel excavat
or Scientific P
a-
pers of the Institute of Mining of the Wroclaw University of Technology, Mining and.
and geology XVI Wroclaw : 5


16

[8] Bartelmus W 2008 Root cause and vibration signal analysis for gearbox condition
mon
i
toring, Insight Vol.50 No4 April

[9] Bartelmus W 1992 Vibration condition monitoring of gearboxes. Machine Vibration
1992 no.1 : 178
-
189

[10] Bartelmus W, Zimroz R (2009) Vibration condition monitoring of planetary gearbox
under varying external load. Mechanical Systems and Signal Proces
sing 23 : 246
-
257.

[11] Bartelmus W, Zimroz R (2009) A new feature for monitoring the condition of gea
r-
boxes in non
-
stationary operation conditions. Mechanical Systems and Signal Pr
o-
cessing 23: 1528
-
1534

[12] Bartelmus W, Chaari F, Zimroz R, Haddar M (2010
) Modelling of gearbox dynamics
under time varying non
-
stationary operation for distributed fault detection and diagnosis

European Journal of Mechanics
-

A/Solids 29 : 637
-
646

[13] Bartelmus W, Zimroz R (2009) Gearbox condition degradation models. Proceedi
ngs of
Conference on Condition Monitoring and Machine Failure Prevention Technology.
Dublin, Ireland, 22
-
25 June

[14] Bartelmus W (2011) Gearbox damage process. Journal of Physics. Conference Series.
[electronic document] 2011, Vol. 305, nr 1, 9p. 9th I
nternational Conference on Da
m-
age Assessment of Structures, DAMAS 2011, Oxford, UK, 11
-
13 July

[15] Bartelmus W 2012 Secondary misalignment as the main cause of gearbox failure I
n-
ternational Conference on Condition Monitoring, 12
-
14 June, London, United
Kingdom

[16]Morales
-
Espejel G E (2012) Understanding and preventing surface distress. Gear Sol
u-
tions, July: 29
-

38

[17] Wowk V (2000) Machine Vibration; Alignment McGraw
-
Hill

[18] Bartelmus W, Zimroz

R (2011) The gearbox vibration load susceptibility as the mea
s-
ure of its condition


approaches and procedures for diagnostic relation discovery. Inte
r-
national Conference on Condition Monitoring, June, Cardiff , United Kingdom

[19] Wocka N. (2003) Rola
i znaczenie sprz
ę
g
ł
a przeci
ąż
eniowego,
Role meaning

of ove
r-
load
coupling,
W
ę
giel Brunatny nr.4 (45) : 20
-
27