1) Go for preventive maintenance!
2) Eliminate process problems
3) Segment your spare parts portfolio
4) Evaluate spare parts criticality
5) Spare parts management starts with good forecasting
6) Use special methods for intermittent demand items
7) Consider the whole life cycle of your equipment
8) Computers and technologies will help you
Eight rules for efficient SPM
Go for preventive maintenance!
STANDARD
PLANNED
PLANNED
MAINTENANCE
CORRECTIVE
(REPAIRS)
UNPLANNED
PREDICTIVE
PROACTIVE
DEFFERED
CORRECTIVE
PLANNED PROCUREMENTINVENTORY
MANAGEMENT
Go for preventive maintenance!
Eliminate process problems
SPARE PARTS PROCESS
Who is responsiblefor spares stock?
Who/how oftenapproves?
Procurementspecification missing
Insufficient sparesidentification
Are returns to warehouse allowed?
Spare parts process – benchmarking
Internal best practice
PLANT 1 PLANT 2 PLANT 3 PLANT 4
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
1 2 3 4
Comparison of inventory turnover
Eliminate process problems
SP NEED
IDENTIFICATION
No direct responsibility of maintenance engineers/technicians for
“their” items and spare parts levels.
REQUEST
FOR ORDER
RFO created by someone else, not the technician who requested part.
The step of RFO may not be necessary in the process.
APPROVAL
How often are orders approved? Who approves?
Approving in IS workflow or by signing a paper copy? Or both?
Approving both RFO and then issued order again?
Too many approvers, complicated procedure and hierarchy.
Approving on high levels of management.
PROCUREMENT
Insufficient information available to procurement, poor spare parts
identification – the buyer hardly knows what should be bought,
additional communication with maintenance technician is needed.
Missing or incomplete procurement specification in the IS.
EOQ
MAX
Ordering systems
s|Q
s|S
Fixed period
Fixe
dq
uan
tity
Var
iab
leq
uan
tity
B|Q
B|S
Variable period
Ideal world: inventory management modelConstant demand and lead time
STOCK
400
200
0Days 10 20 30 40 50 60
Order sent Order sentAverage
stock
Goods deliveredGoods delivered
Order more often!
STOCK
400
200
0DAYS 10 20 30 40 50 60
Order sent Order sentAverage
stock
Goods deliveredGoods delivered
No impactto risk!!
Real example: Material 210015: from order quantity 200 000 pcs down to 80 000 pcs: reduction of average stock by 60 000 pcs = 265 546 Kč
Optimal order size – EOQEconomic Order Quantity
Ordering cost(transportation)
Inventory holding cost
(warehousing)
Total cost
No impactto risk!!
Orderingdaily
Ordering oncea year
Order size [pcs]
Eliminate process problems
CONSUMPTIONSlow spare part issues in case of sudden need.
Issued spare parts are not consumed in fact. What happens then?
Consumption of external material even in case the part is on stock.
WAREHOUSE
RETURNS AND
REFURBISHED
SPARE PARTS
Refurbished parts return to warehouse while new are bought.
Accounting price of refurbished items is much higher (or lower) than
the non-realistic value of items on stock.
Problematic or impossible returns of parts issued but not consumed.
Insufficient control of parts dismantled from the maintained object
(the information system has no information about these).
RECEPTIONProblems with missing (undelivered) documentation for the received
material (certificates, declarations).
Lost spare parts documentation – only “paper-based” archiving.
WAREHOUSING
Insufficient identification of spare parts in the warehouse.
Problems to find items stored.
Inventory count discrepancies, physical stock different from
information system data.
Non-real value of stock in the information system.
Out-of-system stocks.
How are your bearings doing?
How are your bearings doing?
How are your bearings doing?
Segment your spare parts portfolio
Service level x locked-in capital
Inventoryvalue
90 % 95 % 98 % 99 % 100 %?
Service level / availability
What items are really important?
What items are really important?
Dominant categoriesC and D are typical forspares parts
On-hand inventory value
Consumed quantity
ABC analysis – on hand inventory value
Value of available inventory
ABC analysis – consumed quantity
In spare parts, C and D categories typically prevail
Quantity
Spare parts inventories segmentation – consumption frequency(in quantity)
Spare parts inventories segmentation – consumption frequency (in value)
WHEN?
HOW MANY?
WHEN?
HOW MANY?
Strategic segmentation of spare parts
CONSIDER CONSIGNATION
CONSIDER CONSIGNATION
CONSIDER CONSIGNATION
CONSIDER CONSIGNATION
BUY
BUY
BUY
BUY
CONSIDER CONSIGNATION
CONSIGNATION
BUY BACK
BUY
TYP
E O
F FO
REC
AST
CONSUMPTION FREQUENCY
Example of buyback application
20 pieces were purchased for turnaround in 2010 for 41.2 M CZK, but these spare parts
were not used during the turnaround and will be stored until the next turnaround in 2014.
Buy-back in this case can save 9 M CZK (360k EUR) on storage and locked-in capital cost
Between shutdowns
Evaluate spare parts criticality
Effects of failure
Production losses
Links to asset register, critical assets (RCM)
Delivery time
Repairability
Number of items in use
Maintenance planning
Part lifespan
Failure probability
Failure characteristics
Failure anticipation
Price
% cost of capital
Cost ofinventory
holdingFailureprobability
Impacts ofspare part unavailability
Leadtime andother
parameters
Spare parts criticality – 4 bits of information
𝐶𝑖𝑛𝑣 = 𝐶𝑢𝑛 ∗ 𝐿𝑇 ∗ 𝑓
10 000 EUR x 10% 100 EUR / day 100 days 1 per 2 years
1 000 EUR 5 000 EUR
KEEP ON STOCK
<
Criticality calculation
INVENTORY HOLDING UNAVAILABILITY LEAD TIME FAILURE RATE
CMMS spare partsmaster data
RCM dataSpare parts
consumptionforecasts
Preliminarysegmentationof spare parts
Quantitativeevaluation of
criticality
Qualitativeevaluation of
criticality
PreviousSP criticalityassessments
(if available)
Cost ofunavailability
(from RCM or other sources)
Planning Wizarddata analysis
CMMS spare partstransactions
data
Assessment by maintenance engineers and
technicians
Potentiallycritical items
Critical items (criticality score)
Non-criticalitems
Non-criticalitems
1
2
2-levelevaluation of criticality
Criticality evaluation
Critical SP Non-critical SP
Intermittentconsumption
Regular consumption
Stock level set byBootstrapping
Longleadtime
Standard modelsof stock
management+ Safety stock
Standard modelsof stock
management
Shortleadtime
Stock level is set based on level of
criticality and quantity of
installed pieces
Stock level is validated by maintenance
technician
Criticality
Consumption frequency
Lead time
Potentially critical SP Non-critical SPPreliminary identification
of potentially criticalitems
Recommended method of stock control (stock level
calculation)
All active spare parts (SP) Spare parts segmentation
For items under Framework contractswith vendors, automated MRP ordering
in ERP can be applied
Qualitative criticality assessment: online app
Login
Login = Příjmení
Heslo = Jméno
Příklad:
tokac miroslav
Selection of items to evaluate
Answering few simple questions
Answering few simple questions, send
Done – items evaluated
Spare parts criticality analysis result
Criticality score
Spare parts items
Simple segmentation: FSN, VED and SDE
39 http://www.yourarticlelibrary.com/business-management/marketing-management-business-management/ved-analysis-sde-analysis-and-fsn-analysis/69363
V
E
D
F
S
N
Fast-moving
Slow-moving
Non-moving
Vital
Essential
Desirable
S
D
E
Scarce
Difficult
Easy
Consumption CriticalityAvailability
on themarket
http://www.yourarticlelibrary.com/business-management/marketing-management-business-management/ved-analysis-sde-analysis-and-fsn-analysis/69363
Spare parts management starts with good forecasting
Quantitative methods x Common sense
Quantitativemethods
Commonsense
MAKE EFFICIENT!MAXIMIZE!
Quantitative methods x Common sense
Quantitative methods
Common sense
Unexplained / Random
Uncertaintyof future
consumption
Maximize
Make efficient
Minimize
Forecasting step-by-step
Visualisation of time seriesFor better understanding of the time series
Calculation of accuracyAbsolute and relative errors, evaluation on testing season
Calculation of forecasts using all available methods
Selection of the best methodBest accuracy and reliability
1
3
2
4
What forecast method is best for spare parts?
1% 1%2%
3%5%
9%
79%
Konstantní model
Regresní model
Holtovo exp. vyrovnání
Jednoduché exp.vyrovnáníKlouzavý průměr
Winters
Forecasting není možný
Constant model
Regression model
Holt’s exp. smoothing
Simple exp. smoothing
Moving average
Winters
Forecasting impossible?
???
Use special forecasting methods for intermittent demand items
Týd
enn
í sp
otř
eby
ND
Historie týdenních spotřeb (týdny)
4 13 17 300
1
2
3
4
81 18 25 2726
32
Spare parts – intermittent demand
Monthly consumption history (months)
Mo
nth
lysp
are
par
tsco
nsu
mp
tio
n (
pie
ces)
QUESTION: What reorder level should be set in order to ensure required availability of a spare part?
Týd
enn
í sp
otř
eby
ND
Historie týdenních spotřeb (týdny)
4 13 17 300
1
2
3
4
81 18 25 2726
32
BootstrappingBootstrapping = random sampling from history of consumptions.
SP consumption for lead-time period is sampled from history
Sample 1: Consumption in
6 months = 5 pcs
Example: SP lead time is
6 months
Monthly consumption history (months)
Mo
nth
lysp
are
par
tsco
nsu
mp
tio
n (
pie
ces)
Týd
enn
í sp
otř
eby
ND
Historie týdenních spotřeb (týdny)
4 13 17 300
1
2
3
4
81 18 25 2726
32
Bootstrapping
Vzorek 2: Spotřeba za
6 týdnů = 0 ks
Sample 2: Consumption in
6 months = 0 pcs
Weekly consumption history (weeks)
Spar
e p
art
con
sum
pti
on
(p
iece
s)
Monthly consumption history (months)
Mo
nth
lysp
are
par
tsco
nsu
mp
tio
n (
pie
ces)
Bootstrapping
Vzorek 3: Spotřeba za
6 týdnů = 12 ks
Týd
enn
í sp
otř
eby
ND
Historie týdenních spotřeb (týdny)
4 13 17 300
1
2
3
4
81 18 25 2726
32
Sample 3: Consumption in
6 months = 12 pcs
Weekly consumption history (weeks)
Wee
kly
SP c
on
sum
pti
on
Monthly consumption history (months)
Mo
nth
lysp
are
par
tsco
nsu
mp
tio
n (
pie
ces)
Bootstrapping
Vzorek 4: Spotřeba za
6 týdnů = 2 ks
Týd
enn
í sp
otř
eby
ND
Historie týdenních spotřeb (týdny)
4 13 17 300
1
2
3
4
81 18 25 2726
32
Sample 4: Consumption in6 months = 2 pcs
Weekly consumption history (weeks)
Wee
kly
SP c
on
sum
pit
on
Monthly consumption history (months)
Mo
nth
lysp
are
par
tsco
nsu
mp
tio
n (
pie
ces)
Če
tno
sti s
po
tře
b v
jed
no
tliv
ých
inte
rva
lech
(z
e s
imu
lace
)
Spotřeba během LT – intervaly (ks)
0 5 10 15
60 000
0
30 000
Ku
mu
lati
vní p
ravd
ěp
od
ob
no
st s
po
tře
by
100 %
Example of 100 000 simulations of SP consumption
Target: SP availability (Service level) = 99%
OPTIMUM INVENTORY = 9 PCS
Consumption during leadtime (pcs)
Cu
mm
ula
tive
pro
bab
ility
of
con
sum
pti
on
Freq
uen
cies
of
con
sum
pti
on
(fro
m s
imu
lati
on
)
Bootstrapping application – a case study
Original inventory: 17 000 EUR(49 pcs)
Spare part lead-time: 32 days
Recommended inventory
29 pcs
10 000 EUR
Savings
7 000 EUR
99.9% availability
Intermittent demand
Minimum level 2 pcs
Life cycle thinking:Consider the whole life cycle of your assets
PROCUREMENT
MAINTENANCE
RENOVATION
COST
PROFIT?PROFIT
MAINTENANCE CREATES VALUE!
ASSET LIFE CYCLE
Asset life cycle (Kari Komonen, EFNMS EAMC)
INVESTMENT UTILIZATION
$$
PROCUREMENT
MAINTENANCE
RENOVATION
COST
PROFIT?PROFIT
MAINTENANCE CREATES VALUE!
ASSET LIFE CYCLE
PROFIT
Efficient spare parts management:
Computers, information systems and technologies will help you!
Efficient spare parts management – 8 rules
Preventive maintenance Smooth SP processes
Segment your SP portfolio Assess criticality
Good forecastingSpecial methods for
intermittent demand items
Life cycle thinkingGood information system
and smart tools
Good information system for spare parts management
Quantitative methods
Common sense
Uncertainty
Maximize!
Make efficient!
Minimize!
ForecastingInventory levels
Ordering
Input of technicians and procurement
Criticality analysis
Forecast accuracyand reliability
Identification of spare parts in smartphone
Even you can have an app in your phone that solves your process or problem.Today it is easy. And cheap.
Forecasting of spare parts consumption with neural networksAI and advanced statistics work for us.
Artificial intelligence is working for us:Diagnostics from acoustic emission using neural networks
Analysis ofdata in place
of origin
Remotediagnostics
https://www.youtube.com/watch?v=yuZtnEZEOxw
Prevention, inspectionand cleaning:Jettyrobot
Even more preventive
maintenance
3D PRINTING OF SPARE PARTS IS REALITY
Today, blades for gasturbines can be printed
Keepspares on printer!
Culture Eats Strategy
For Breakfast,
Innovation For Lunch,
And Transformation
For Dinner
CULTURE STRATEGY
Efficient Spare Parts Management
MAINTENANCE VOLUME
SPARE PARTS INVENTORY
MAINTENANCE VOLUME
SPARE PARTS INVENTORY
Thank youLogio, s.r.o.Evropská 37160 41 Praha 6
Tel +420 731 151 276Email [email protected]: https://www.linkedin.com/in/tomashladik/
www.logio.cz
https://www.linkedin.com/in/tomashladik/