White Paper Monitoring Failure - icareweb.com

I-care Reliability Group www.icareweb.com | [email protected]

Copyright© I-care SPRL or official supplier of I-care Reliability Group - All rights reserved. While every effort is made to assure the information in this document is accurate, I-care SPRL

does not accept liability for any errors or mistakes that may arise. Specifications are subject to change without notice.

Monitoring Tools – Predictive Analytics for Maintenance – Reliabil ity Solutions – Training & Coaching

White Paper

P a g e | 1

Monitoring Failure Mechanisms By Ray Garvey, R&D Engineer

Why do components fail? What can we do about it? This article explains eight common

failure mechanisms, types of equipment to which each applies, and recommends non-

intrusive monitoring techniques to discover why components are in various stages of

progressive failure.

This article builds on earlier publications on the failure mechanismsi, wear rates ii, stress wave

analysis iii, the role shear plays in failure mechanismsiv, and wear out failure mechanismsv.

Sonic and ultrasonic stress wave analysis using microphone and radio wave sensors are

featured in this article. It will be shown how these novel techniques supplement,

complement, and advance the state of the art regarding condition monitoring for several

failure mechanisms.

Eight Failure Mechanisms Abrasion, corrosion, fatigue, and adhesion, cavitation, erosion, electrical discharge and

deposition are failure mechanism from the referenced literature. Characteristics of each one

are stated below and in Table 1.

1. Abrasion

Silica dust particles are transported by lubricant to a narrow clearance between moving

surfaces. The hard particles too large to pass through embed into one surface and cut the

other. Shear force between the lubricated hard particles and the moving surface cut a v-

notch into the moving metal surface. This cutting process emits a spectrum of mechanical

vibration from the point of abrasion and generates abrasive wear debris which is carried

away by the lubricant. This mechanism is generally not self-propagating and easily offset

by particulate contamination control. It affects nearly all mechanical systems. This

mechanism can be triggered by a surge in circulating system or by a defective breather.

https://www.machinerylubrication.com/Read/29230/condition-monitoring-machinery





White Paper

P a g e | 2

Illustrations : abrasive wear particles

2. Corrosion.

A corrosive substance attacks metal and convert surfaces from strong, thermally and

electrically conductive metal into soft, electrically and thermally resistive oxide. The

resulting oxide is easily rubbed off by shear which exposes fresh metal for sustaining

oxidation. This mild rubbing emits stress waves from vicinity and wipes soft metal oxides

into the lubricant, exposing metal to the oxidation process. This mechanism is offset by

moisture contamination control. It can be triggered by process contamination, coolant

leak or defective desiccating breather. Corrosion affects nearly all electrical and

mechanical systems and is synergistic with all of the other mechanisms.

Illustration : corrosive wear particles

3. Fatigue

Roller bearings and gears often fail due the process of rolling contact which eventually

results in material fatigue cracks and spall. Compression between rollers and races and

between gear teeth produces sub-surface Hertzian contact shear that eventually work

hardens the metal until microcracks originate, grow, interconnect, and then release metal

debris typically in forms of chunks, platelets, and needles. This emits stress waves from

impacts and releases the metal debris into the lubricant. This mechanism is offset by

minimizing dynamic loading from imbalance, misalignment and resonance, by static load

reduction, and by other good maintenance practices. It can be triggered by improper fit

or thermal growth. It affects mechanical systems with loaded bearings and gears. As





White Paper

P a g e | 3

described below, the mechanism of cavitation also causes cyclic sub-surface shear

resulting to material fatigue cracks and spall.

Illustrations : fatigue wear particles

4. Adhesion (or boundary wear)

A metal-to-metal contact occurs when the lubricant film designed to eliminate friction

and separate roller from race or journal from shaft fails due to inadequate lubrication.

The increase in friction and shear causes mixed mode and boundary mode lubrication

regimes. The contact emits stress waves. Compression with mixed mode and boundary

lubrication results in shear and friction that causes intense heating, melting, and

discoloration. It releases metal debris and metal oxides into the lubricant and emits a

spectrum of vibration. This mechanism is offset by maintaining correct lubricant at the

correct level and operating at design speed and load. This mechanism may be triggered

by too slow speed, too high load, too low viscosity, and inadequate lubricant delivery. This

mechanism affects nearly all mechanical systems with loaded components. Adhesive wear

and other boundary wear damage is progressive, self-propagating, and accelerates

corrosion.

Illustrations : severe sliding wear particles

5. Cavitation

Liquid cavitation leading to solid surface damage is stimulated by cyclic fluid flow dynamic

pressure variation in vicinity of the surface. In a slow part of the pressure cycle suction

enables evacuated micelle nucleation originating from solid surface irregularities. Highly





White Paper

P a g e | 4

saturated dissolved gas from surrounding liquid may diffuse into expanding bubbles.

Later in the pressure cycle, suction is released, bubbles implode toward the nucleation

surface cycle. The implosion causes a shape-charge supersonic surface impulse, analogous

to the pop from the end of a bull whip, transfers compression and shear stress waves

following the collapse. Subsurface shear from the fluid-structure stress wave dislocates

sub-surface material morphology. Eventually the dislocations lead to fatigue cracks and

then spall. Note that when the bubbles contain partial pressure gases diffused from near-

saturation surrounding liquid, then there is also localized intense heating from the

compressed gases. The cavitation impulses from cavitation events emit stress waves

dislodge particulate debris. This mechanism typically occurs on impellers, pumps, valves,

other flow devices supporting the described cavitation damage process. Cavitation

damage is offset by fluid flow design, control, speed and surface treatment. It triggered

by pressure, flow, and speed variation. Cavitation damage is normally progressive, self-

propagating, and often leads to fatigue cracking and stress corrosion cracking.

6. Erosion

High velocity particulate liquid or solid matter impacts a solid surface causing intense

points of compression resulting in deformation and shear that emits stress waves from

the points of impact and dislodges debris from the damaged surface. This mechanism is

offset by protecting the surfaces of interest with energy absorbing coatings. This

mechanism affects valves, pipes, baffles, impellers, and many other electrical and

mechanical components exposed to streaming particulate matter.

7. Electrical discharge

Electrons transported as parks, partial discharges, and arcs blast target surfaces with

intense local compression causing deformation and shear that generates a wide spectrum

of mechanical and electrical energy. Electrons pass through gaps at supersonic speeds (say

30 m/s) emitting radio waves and sonic booms, generating intense local heat damage to

surfaces and producing various gaseous substances such as hydrocarbons and ozone. This

progressive mechanism ionizes proximate matter to form a discharge or plasma track.

The mechanism may be offset by maintenance of clean, dry, fit for use materials and

compartments. It is triggered by moisture, deteriorated insulation, ground faults,

looseness, and corroded contacts. This mechanism affects all statically charged and

electrically powered equipment including electrical switches, circuits, wiring harnesses,

connections, breakers, transformers, compartments, controllers, motors, DC and variable





White Paper

P a g e | 5

frequency drives, generators, filters, shaft bearings, and housings requiring an electrical

earth ground.

8. Deposition

This mechanism results from a dysfunctional and progressive accumulation of foreign

material on a critical component. Two examples of deposition are precipitated varnish

formation and accumulation on a control valve and fibrous material accumulation on a

fan. Varnish accumulation on a control valve may lead to plugging and sticking. Fibrous

material accumulations on a fan may lead to imbalance and potential fire risk. The

deposition mechanism is offset by detecting, interpreting, and addressing the specific

progressive accumulation process. Each corrective action plan is specific to its

characteristic process.





White Paper

P a g e | 6

Table 1. : Eight common failure mechanisms with equipment, contributing factors, proactive measures and

condition monitoring.





White Paper

P a g e | 7

Failure mechanism monitoring techniques

• Elemental spectroscopy

X-ray fluorescence (XRF) elemental spectroscopy of filter patch specimens is preferred for

large particle failure mechanisms including abrasion, fatigue, and severe adhesion. Optical

emission spectroscopy and XRF are both suitable for corrosion and mild adhesion

mechanisms.

• Particle count and particle shape classification

Particle counts at >4, >6, and >14 m size ranges enable condition monitoring for

contamination control. Direct imaging automatic particle shape classification or

microscopic wear particle analysis enable distinguishing of failure mechanism.

• Radio wave arc/spark detection

This new stress wave analysis technique non intrusively detects arcing, sparking, and

partial discharge events in electrical and electromechanical systems. See FIGs. 6 to 8.

• Inspect and special test

Electrical discharge for oil filled compartments may benefit from dissolved gas analysis

(DGA) looking for evidence of turn-to-turn arcing. Deposition and accumulation of matter

on flow controls, filters, screens, valves, fans, and oil compartments, is a failure

mechanism resulting from a variety of operational conditions. Inspection and testing

protocol depends on these things. For example, membrane patch colorimetry (MPC) is a

preferred testing technique to identify evidence of varnish precursors.

• Stress wave analysis

Wear mechanisms of abrasion, rubbing (associated with corrosion), fatigue, adhesion

(boundary), cavitation, erosion, and electrical discharge can be non-intrusively sensed

using a suitable analog sensor. Accelerometers, microphones, radio wave sensors, current

probes, and magnetic flux sensors are other examples of analog sensory input devices

used in stress wave analysis. Analog to digital data is oversampled and selectively

decimated to derive simultaneous sonic and ultrasonic peak-hold waveforms. A peak-hold

stress wave analysis waveform may be either maximum rectified peak or maximum peak-





White Paper

P a g e | 8

to-peak. The data shown in FIGs. 1-8 are the peak-to-peak type, and the sensors used for

these measurements were microphone or radio wave sensors as stated.

• Thermal imaging

Electrical resistance and electrical arcing and mechanical friction produce hot spots

detectable with thermal imaging.

• Total ferrous

A magnetometer is preferred for determining total ferrous concentration (PPM) for all

ferrous oxide and ferrous metal particles from molecular to abrasive wear particle size

range. This tool is very useful for quantifying wear and severity for ferrous debris in

lubricating fluids.

• Vibration analysis

Mechanical vibrations below a maximum frequency of interest (FMAX) are monitored to

characterize proactive root causes of the failure mechanisms such as imbalance,

misalignment, looseness, resonance, and soft foot. They are also monitored in

combination with stress wave analysis techniques to characterize predictive incipient to

catastrophic stages of failure.

• Viscosity

Lubricant misapplication, e.g., wrong oil, is frequently identified by verifying correct

viscosity for in service lubricants. This directly relates to monitoring for inadequate

lubrication associated with adhesion.

• Water, coolant, and neutralization number

A convenient on-site method for monitoring water and coolant and total acid or total base

number is transmission infrared spectroscopy of in service lubricant. Laboratory titration

methods (Karl Fischer, TAN, and TBN) are also effective. Water, coolant, and acid are all

related to corrosive wear mechanisms.





White Paper

P a g e | 9

Tension, Compression and Shear “Nothing, well almost nothing, fails in compression,” is the answer from a DOE Y-12 laboratory

chief scientist, John Googin, when I asked about failure mechanisms.

John suggested that at times when I think compression is a primary cause of failure, closer

study of evidence would certainly expose either tensile or shear mechanism initiating a

progressive failure sequence. Three decades later I have not found an exception, and shear

force is nearly always a primary contributing factor from incipient to catastrophic failure

mechanisms.

Lubricated load bearing surfaces allow machines to do work by way of compression through

a lubrication film. Work is application of force through a distance or pressure through a

volume. Mechanical systems are designed to perform work and to have very long life by doing

that work through applied tension and compression. The designer intended long life gets cut

short by shear. Failure mechanisms of abrasion, corrosion, fatigue, adhesion (or boundary

lubrication regime), cavitation, erosion, and electrical discharge each have a common failure

element of shear.

The following eight figures demonstrate how compression and shear may be revealed by

simultaneous collection of sonic and ultrasonic stress waves from at least 80 kHz sampling

rate stream of digital data. These techniques apply to measurements from various sensors

including piezoelectric accelerometer, electret condenser microphone, and radio wave

antenna. An appropriate sensor is selected for the non-intrusive measurement of failure

mechanisms. Preferred stress wave analysis methods simultaneously perform peak-hold

sonic analysis of 500 Hz high pass signals and peak-hold ultrasonic analysis of 20 kHz high pass

signals from one oversampled data stream.

All graphs in Figures 1 to 8 include an orange and a blue plot. The abscissa (Y-axis) is signal

strength millivolts (mV) and the ordinate (X-axis) is time in seconds (s). The area under the

orange line is the total ultrasonic peak energy above 20 kHz. The area between the blue line

and the orange line is the total sonic peak energy between 500 Hz and 20 kHz. The ultrasonic

energy under the orange line is related to shear energy transfer mechanisms of friction and

turbulence. The sonic energy between the blue and orange lines is related to forceful

compression energy transfer reflecting work done by force through distance or pressure

through volume.





White Paper

P a g e | 10

Figure 1 displays typical adhesion and fatigue impact events. These 12 examples show forceful

high amplitude compression impact event patterns which are characteristic for these metal-

to-metal collisions. Notice in every case the ultrasonic signal (orange) remains small

compared with sonic signal (blue).

FIG. 1 : Microphone sensor collected stress waves from typical impacts involving a variety of adhesion events

and fatigue defects.

Figure 2 shows how a reciprocating compression mechanism is performing work through

compression with negligible shear. In this case the lubricant provides full hydrodynamic fluid

film minimizing shear throughout continuous reciprocating sliding contact.

FIG. 2. : Airborne stress waves from reciprocating compression with hydrodynamic fluid film separation.





White Paper

P a g e | 11

Figure 3 shows how stick-slip friction during mild adhesive rubbing wear is largely ultrasonic

without much sonic energy. The ultrasonic high pass above 20 kHz orange lines are “in front”

of the sonic high pass above 500 Hz blue lines. However, by definition, the high pass peak-

hold above 500 Hz is always equal to or greater than high pass peak-hold above 20 kHz. A

characteristic sonic wave pattern is evident in the right graph. Overall, Figure 3 represents

shear due to friction during mild adhesive rubbing wear.

FIG. 3. Airborne stress waves from mild adhesive rubbing wear.

Figure 4 is similar to Figure 3 with abrasive cutting wear doing physical work evidence by the

separation between ultrasonic orange and blue lines.

FIG. 4. Airborne stress waves from abrasive wear.





White Paper

P a g e | 12

Figure 5 shows the process of erosion with fluid and particulate impacts delivering

compression and shear events in these time domain plots. These typical plots were recorded

at a waterfall. Coincidentally, water action shows up in the blue color, particle impacts are

sand color.

FIG. 5. : Airborne stress waves from erosion.

Figure 6 shows radio wave energy during partial discharge events. Note that electrical

discharge events tend to be very, very fast ultrasonic events. However, in partial discharge

the radio wave signals appear slow and attenuated as seen in this figure. Notice sonic energy

is several times stronger than ultrasonic radio wave energy. Perhaps this is evidence that the

electrons departing from conductors are not completely escaping the surrounding insulating

matter.

FIG. 6. Radio stress waves from electrical partial discharge.





White Paper

P a g e | 13

Figure 7 shows plots from microphone and radio wave sensors producing sonic and ultrasonic

stress waves tracking events occurring during continuous plasma arcing electrical discharge

with 10 kV arcing over a ~10 mm gap. Typically arcing produces fast ultrasonic events.

Therefore, the orange line tends to overlay and obscure the blue line in the plots. Arcing

events are more common than sparking events for electrical equipment 2 kV and above..

FIG. 7. Sonic and ultrasonic event stress waves using microphone and radio wave sensor to monitor a

continuous plasma and arcing electrical discharge with 10 kV

Figure 8 shows plots from microphone and radio wave sensors producing sonic and ultrasonic

stress waves tracking events occurring during spark events from 120 V source. Sparking

events are more common than arcing events for electrical equipment 480 V and below.

FIG. 8. Microphone and radio wave sensors detect stress waves from sparking electrical discharge originating

from a 120 V source.





White Paper

P a g e | 14

Conclusion This article attempts to describe how equipment fails and what may be done to improve

overall reliability. The article identifies common mechanisms of failure: abrasion, corrosion,

fatigue, adhesion, erosion, cavitation, electrical discharge, and deposition.

Each mechanism has contributing factors, proactive measures, and affects various kinds of

equipment. Several preferred non-intrusive monitoring techniques are identified to enable

proactive and preventive efforts for improved reliability.

Simultaneous sonic and ultrasonic stress wave analysis using microphone and radio wave

sensors which advance state of the art are featured from the broad list of valuable condition

monitoring techniques.

Source of this article :

Ray Garvey - R&D Engineer, I-care Reliability Inc. Ray is an engineer and inventor named on two dozen US patents associated

with oil analyzers, infrared thermography, machine monitoring, and

composite structures. Ray is known for his participation in developing CSI

5100 and CSI 5200 minilabs. Ray received his BS Degree from West Point

and MS degree from the University of Tennessee. His professional

certifications have included Professional Engineer (PE), Certified

Lubrication Specialist (CLS), and US Army Engineer (LTC retired). Ray

worked for the US Army Corps of Engineers, US DOE Uranium Gas

Centrifuge Program, CSI Emerson Process Management, I-care Reliability

Inc., and Spectro Scientific

i “Identifying Root Causes of Failure with Condition Monitoring”, Ray Garvey and Pat Henning, Machinery Lubrication Magazine, December 2012 ii “How Machinery Wear Rates Impact Maintenance Priorities”, Ray Garvey, Machinery Lubrication Magazine, March 2003 iii “Intelligent Decimation: Closing the Gaps Between Vibration and Oil Analyses”, Ray Garvey, Machinery Lubrication Magazine, April 2019 iv “Composite Hull for Full-Ocean Depth”, R. E. Garvey, Marine Technology Journal, Volume 24, Number 2, June 1990 v “Converting Tribology Based Condition Monitoring into Measurable Maintenance Results”, by Ray Garvey and Grahame Fogel, Computational Systems Inc., 1998

White Paper Monitoring Failure - icareweb.com

Documents