Using
Data
Analytics
to
Issues,
Trends,
Faul
Equipmen
Automatically “find what
equipment systems R D Remote Dec
o
Automatically
Detect
lts
and
Anomalies
in
nt
Systems
t matters” in the data from
s and smart devices
b 2013 cember 2013
Our world contains bi
Producing vast qua
Producing
vast
qua
sec
ntities of data every
ntities
of
data
every
To create value from
To
create
value
from
to
find
what
mat
m our devices we need
m
our
devices
we
need
…By detecting patt
…By
detecting
patt
deviations,
anom
opportunitie
opportunitie
terns that represent
terns
that
represent
malies,
faults,
and
es for savings
es
for
savings
Too Much Data !!!
Analytics
y
is
the
Key
y
A
l ti
ft
t
Analytics
software
autom
“issues”
(things
that
matt
Equipment
faults,
deviati
performance,
actual
resu
etc
Unlike
efficiency
measure
installation
of
major
capit
works
with
the
data
avail
Relatively
easy
to
add
to
ti ll l
k f
matically
looks
for
ter)
in
our
data….
ons
from
expected
lts
vs goals
or
benchmarks,
es
that
involve
the
tal
equipment,
analytics
lable
from
existing
sources
what
we
have
The
Bigger
Picture:
An
ld
our
world…
It can change the
way we manage
our systems too!
our systems too!
Gather
data
and
find
pa
the issues
that matter
the
issues
that
matter
Device data Utility data Facility data Utility data Weather data Production data
Connect to available data Aggregate Connect to available data
sources
Aggregate available d
tterns
that
represent
Easy to get started – what data do you have?
data do you have?
e and normalize Detect patterns that e and normalize
data
Detect patterns that represent issues
Automatically fi
Automatically fi
correlations i
ind patterns and
ind patterns and
in device data
The
Result:
Know what your
Know what your systems are really
doing
Automatically scans your
data to find what matters Automatically generates
views on issues detected Convert expert domain
knowledge to rules –
your value continues to
y
build
Explore relationships and
l ti ld correlations you would
not have otherwise seen
The “Spa related to
ark detail” page – shows everything o the occurrence of an issue
Vehicle/Fleet
Tracking
Concept:
Concept:
Collect
and
analyze
vehicl
GPS
data
to
determine
operational
p
issues
Examples
of
potential
ana
•
Vehicles
traveling
outsid
“fencepost
area”
•
Vehicles
stopped
for
too
•
Speed in excess of regula
•
Speed
in
excess
of
regula
g
e
alytics:
de
of
o
long
ations
ations
Vehicle Tracking Example Example
Security
Concept:pCollect and analyze security sy represent threats or improp represent threats or improp conditions
Analytic example: Analytic example:
• Combine video analytic d
d t E l b l ft
data. Example: a bag left
Expected response is disp confirm
confirm
• Analytic looks for correla response rates time to re response rates, time to re
ystem data for patterns that
per per
detected “events” with other
tt d d t d
unattended at a door.
patch of guard to area to
tion of responses to events,
esponse etc esponse etc.
Cold
Chain
Manageme
Concept:• Need to insure that food, drugsNeed to insure that food, drugs perishables) have been mainta temperatures – if not they mus by regulation
• Monitor temps, compressors, o times, etc.
• Track patterns and magnitude o faults – time is a major factor
• This is a ggrowingg area of regulag
• Analytics simplifies and reduce analysisy
ent
s (and other s (and other ined at correct st be discarded open doors of deviation toryy compliancep s cost ofTracking
Data
Usage
i
M2M
Applications
Concept:
• Cell‐phone connected devices r server. Each site needs to have
sized based on initial expectatio • Buying more data capacity than
expensive
• Going over your plan allotment • But manually analyzing usage a
very difficult and expensive
• Analytics enabled a partner to a choose a more cost effective pl
n
report data back to a central
a data plan. Data plans are
ons
n needed can be very
t can be even more expensive across thousand of devices is
analyze actual data needs and
Analytics
y
vs Alarms
An alarm is when you are on the g
An alarm is when you are on the g
Analytics are the lab tests you tak stay out of the ER
Alarms require that you fully unde
Analytics find patterns & issues yo
Controller‐based alarms only dea
Analytics combine operational, en corporatep data to show patternsp a portfolio of device data
Correlation examples – equipm
h ff d f
weather effects, production fa
Analytics replace the majority of n
that explain what is happening an that explain what is happening an
vs
gurney in the ER
vs
gurney in the ER
ke every year to
erstand the issue ahead of time –
ou couldn’t have foreseen
l with control system data –
nergy, production, facility and and correlations across youry
ment type, age, material, vendor,
actors, etc
non‐productive alarms with insights
nd why nd why
Summary
‐
Analytics
Summary
Analytics
You can’t control what
You
can t
control
what
Analytics
enables
us
to
equipment
systems
ar
Easy to get started W
Easy
to
get
started.
W
Use
available
data
– re
Its
time
to
generate
va
s Value Proposition
s
Value
Proposition
t you don’t measure
t
you
don t
measure
o
know
how
your
re
actually
operating
hat data do you have?
hat
data
do
you
have?
eal
time
and/or
batch
alue
from
our
data!
Fi d h
Find
wha
k f www.skyft
tt
™
t
matters™
f d foundry.com