Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Improving the diagnostic
capabilities with Big Data
Analytics and Prognostics
JORGE ALARCON | JESUS TERRADILLOSIK4-TEKNIKER
SU
MM
AR
Y IK4-TEKNIKERDIAGNOSTIC BACKGROUNDWHY TO IMPROVE THE DIAGNOSTIC?WAYS TO IMPROVE ITCASE STUDY
INDUSTRIAL MILESTONES
Industry 1.0Steam-driven
production
Industry 2.0Mass production
supported by specialization and electricity
Industry 3.0Production
automation, electronics and
IT
Industry 4.0 Production bycyber-physical
systems
Connected and Smart
devices
1800 1900 1970 2016
EVOLUTION OF
MANUFACTURING
NEW DIGITAL
AGE
NEW MARKET
EXPECTATIONS
CHANGES IN
CUSTOMER
BEHAVIOR
WE ARE MOVING TOWARDS 4TH MAINTENANCE GENERATION
1ª Generation: Mtto Corrective
2ª Generation: Mtto Preventive
3ª Generation: Mtto Predictive
MAINTENANCE MILESTONES
REACTIVE
1900 1940
PREVENTIVE
1950
PREDICTIVE
1978
NOWLAN
& HEAP
1998
ASSET
MANAGEMENT
,,,
HOLISTIC
RM PM PdM RCM AM iso55000
Appropriate Lubrication processes in plant
32%
Average time spent on corrective maintenance
activities
40%
Ref: Machinery Lubrication Reader Survey - March, 2011 (n: 347)
Ref: Think Act - November 2014
STILL THE SAME PROBLEMS
50%mechanical components
are replaced due to premature wear
IS THIS A GOOD REASON TO IMPROVE DIAGNOSTIC ON MONITORED EQUIPMENT?
DATA SOURCES
HOW
WHEN
CO
ND
ITIO
N M
ON
ITO
RIN
G T
OO
LSN
EW SC
ENA
RIO
?
GOOD
1970´s 2015COMMON DIAGNOSTIC
ALERT CRITICALBelow limits Slightly above limits Above limits
32
WHAT IS DIAGNOSTIC?
1 4
EQUIPMENT IN SERVICE
DATA CAPTURE
SOFTWARE ANALYSIS
HUMAN DECISION
47%
of jobs will be automated
by machines or
software
In 20 years
Source :KwonMads, John Moravec
There is enough information (Big Data)
Strange but important events
Multidisciplinary team: Experts + analysts
Industrial companies are
not aware of the potential
on the analysis and their
data
Ref: SAS Report on Big Data 2013
PREDICTIVE
PRONGNOSTICS?
76%
WAYS TO IMPROVE IT
GOOD
COMMON DIAGNOSTIC
ALERT CRITICAL
BIG DATA ANALYTICS
PROGNOSTICS
+
+
BIG DATAIs a new paradigm that represents the search for solutions to
store and analyze structured and unstructured data together
for an affordable and scalable data mode.
PROGNOSTICSAlgorithms to detect interesting patterns in data.
Statement about the way things will happen in the future
PREDICTIVE ANALYTICSA variety of statistical techniques from modeling, machine
learning, and data mining that analyze current and historical
facts to make predictions about future, or otherwise
unknown, events.
SU
MM
AR
Y IK4-TEKNIKERDIAGNOSTIC BACKGROUNDWHY TO IMPROVE THE DIAGNOSTIC?WAYS TO IMPROVE ITCASE STUDY
CASE STUDY
• Critical bearings in Plant: 320
• Schedule Maintenance Shutdown: 4-6 months
• 5 years of grease analysis of each bearing
• 2 different type of greases
• 3 different bearing models
22%
36%
42%
CRITICAL ALERT GOOD
BEARING CONDITION
CAN WE WAIT 4-6 MONTHS?
Critical 71: DO SOME MAINTENANCE!Alert 115: BE AWERE OF CONDITIONGood 134: ENOUGH TIME
INFORMATION
• 1 bearing shutdown = 1.500 € per day
• Time to repair a bearing = 3 days
• Potential production losses = 319.500 €
(critical) + 517.500 € (alert)
INFORMATION
• Around 4.000 reports
• 3 different type of failures detected in the past
BIG DATA ANALYTICS & OIL ANALYSIS
BIG DATA ANALYTICS
HOW is going to fail?
9 bearings of type 2 - Failure type b17 bearings of type 3 - Failure type a
PROGNOSTICS
WHEN is going to fail?
PROGNOSTICS
Critical 71:31% will fail in the next 3 months
Case Study Conclusions
DIAGNOSTICS
PROGNOSTICSBIG DATA
ANALYTICS
More Data doesn´t just let us see more
More data allow us to see new
More data allow us to see better
More data allow us to see different
JORGE ALARCON
To sum up