Modeling multiple risk factors and working age

1OMDECOptimal Maintenance Decisions Inc.Optimal Maintenance Decisions Inc.

Proportional Hazards Modeling(“Extended” Weibull analysis)

Now, what would the optimal policy be if we have the benefit of extra data - namely, condition data?

Parameters Condition data values at time t

th ...2321 rpmaxialVibipsPbppmFeppme

Work orders

CBM data

VA+OA+Other monitored data

A Watchdog Agent

Event information

ect io

DecisionModel

Decisio

Statisical models (EXAKT), Expert Systems, trained neural nets, Case Based Reasoning …

From Data to Decision

Example 1

Example 1 (slide 2)

Example 1 (slide 3)

Example 1 (slide 4)

Example 1 (slide 5)

Example 1 (slide 6)

Example 2 (slide 1)

Example 2 (slide 2)

Building the Model - Graphical Investigation

Data missing in this region

More data investigations

Corrected Silicon

Oil changes

BFeppm

Working ageDOil change interval

Missing oil change events

Missing ‘OC’ events?

Strange history

Strange History?

Investigating the strangeness

No events to support this jump in values

A major challenge in CBM optimization

The Definition of Failure

Initially a failure was declared when wear metals were high.

This was like forcing the model to “chase its tail”.

Needed a physical definition of failure based on the observable condition of the wheel motor at overhaul.

Based on the new definition (gear damage), the number of histories ending in failure doubled.

Model “fit” improved dramatically.

The Wheelmotor Optimal CBM Model

Failed at WorkingAge = 11660 hr

Inspection atWorkingAge = 11653 hr

Inspection atWorkingAge = 11384 hr

Had we replaced at 11384 hr…or 11653 hr…!!!!

Profitability Impact of Optimized CBM

CR = 325% $1M

CR = 544% $1.7M

CR = 650% $2M

Sensitivity analysis

Sensitivity of model to cost

Example 3 – Complex items

Inspections_MA

IdentDateWorkingAgeCovariate1NameCovariate2Name…Comment

Events_MA

IdentDateWorkingAgeEventComment

EventsDescription_MA

EventNamePComment

VarDescription_MA

VariableNameMeasureUnitWarnLimit1WarnLimit2…Comment

CovariatesOnEvent_MA

EventStartingDateEndingDateCovariate1NameCovariate2Name…Comment

IdentToModel

ModelNameIdentNameDate

EventToModel

ModelNameInputEventNameOutputEventNameInputPOutputP

VarToModel

ModelNameInputVariableNameOutputVariableNameVariableDataTypeMeasureUnitWarnLimit1WarnLimit2…

Example 3 (2)

Example 3

The onion skins of CBMCBM (least intrusive)

CBM phys. inspection

CBM overhaul

$Model A

Model A

Model B

Truck Engine

1. Copper appearing in oil analysis

2. Checked filter – copper particles found

3. Drained engine oil – copper particles found on drain plugs

4. Removed the engine sump and discovered that the main oil pump drive gear-retaining bolt had come loose.

5. Failure mode: excessive wear on brass bushings due to improperly torqued retaining bolt

layers of discovery

Reference:Derek Wilcock, Sishen Iron Ore Mine, "Engine Failure Results in Proactive Solutions". Practicing Oil Analysis Magazine. September 2005

Increment the knowledge base!

Item Function Failure Cause Effects

Engine To lubricate oil whetted components

Pumps insufficient lube

Main oil pump drive-retaining bolt comes loose due to incorrect torqueing of the bolt

Excessive play between the drive shaft and the front housing causes excessive wear on the brass bushing located in the pump housing and the brass thrust washers on either side of the gear. Particles of copper are generated and will appear in the oil, the sump, the filter, and drain plugs. If allowed to continue, the engine will be badly damaged, to the point of requiring a complete rebuild. The cost to Sishen Iron Ore Mine would be $181,388 to replace a K2000 series engine.

Machine (several test points)

Vibration spectral data

Screening MatrixColumns:1. 10 preselected orders2. 2 highest non-synch peaks in hi

and lo ranges3. Noise floorRows:1. from previous inspection2. Deviation from Av + 1

Normalization Cepstrum4 highest peaks1x sidebands

Demodulation

PossibleBearingdefects Signal

processing

Decision making

Diagnostic templates

For a major comp. group

Pass or fail each diagnosis

Faultdiagnosis

Relativeseverity

Component Specific Diagnostic Matrices (CSDM)

Unique component specific frequenciesAdjacent componentsDifferences from previous inspection.Deviations from Av + 1

4 brg overhung rotor

Close-coupled (no brgs in pump)

vertical

Component specific diagnostic matrix and diagnostic templates

Peaks in the spectrum with 3.61x spacings

Peaks in the spectrum with 1x spacings

Cepstrum

10 2 3 7654 8 109

1x 1x 1x 1x

Spectrum

Inner race defect travels in and out of the load zone

softer impact

harder impact

Repetitive change in impact forces causing amplitude modulation appearing as sidebands

Harmonic energy related to inner race defect

Expert rule using results of Demodulation

Expert rule using 1x vibrations1. Consider (for simplicity only the

1x vibration levels of) the vertical motor and centrifugal pump (with coupling),

2. Excessive (7-10 Vdb over baseline) 1x vibrations.

3. Could mean motor imbalance, pump imbalance, angular misalignment, foundation horizontal flexibility, a radial or thrust bearing clearance problem, or motor cooling fan blade damage. Which?

4. Axial and radial data at both locations angular misalignment?. 5. Radial is higher than axial motor imbalance or pump imbalance?6. Axial motion is characteristic (due to rocking) of unbalance in a vertical pump.

Which component is unbalanced?7. In a vertical pump one direction, the direction of external structural support, is

always stiffer than the other directions. The radial axis in this case is the direction of structural flexibility.

8. Low 1x levels at the pump in the tangential direction because tangential axis is the direction of high structural stiffness and therefore the tangential component of the vibration due to motor imbalance does not transmit to the pump.

Hybrid decision agent

Optimal decision

Modeling multiple risk factors and working age

Documents

MULTIPLE VIEWPOINTS MODELING OF TABLA SEQUENCES...

Multiple Factors for Organophosphorus Resistance in the ...

A Multiple Regression Analysis of Factors Concerning ...

Factors Greatest Common Factors Multiples Lowest Common.....

Advanced Threat Modelling Knowledge Session · Threat...

THE PSYCHOMETRIC MODELING OF ORDERED MULTIPLE …

THERMO-HYDRO-MECHANICAL COUPLED MODELING: MULTIPLE ...

Regression-based Modeling of Environmental Factors ...

MODELING MULTIPLE OBJECT SCENARIOS FOR FEATURE …

Cognitive Modeling in Human Factors (Generally) and in...

E-Training Integration In Organisation: Modeling Factors ...

Modeling multiple category brand preferenc

Computational Modeling using Multiple Resource...

Factors Influencing Drug Delivery to the Brain: Multiple...

Modeling multiple regimes in business cycles.pdf

Modeling the factors limiting the distribution and ...