Professional Documents
Culture Documents
Gary Higganbotham
Solar Turbines Black Belt
April 16, 2009
Agenda
• Solar Turbines \ Caterpillar
- Brief background of company and products
• What is 6 Sigma?
- Background behind the methodology
• What is 6 Sigma?
- Background behind the methodology
• An enterprise can not move forward unless the need for change
is clearly established
200’
200’
The Burning Platform - Why change?
• What is 6 Sigma?
- Background behind the methodology
• A measurement of variation
Define
TS
Define
PR
EN
OC
EM
Measure Measure
ES
OV
S
PR
Analysis
CR
Explore
IM
EA
S
TIO
ES
Improve Develop
OC
N
PR
Elimination
EliminationofofVariation
Variation
Desired / Nominal
Desired / Nominal Desired / Nominal
- +- + - + - +
• Problem Identification
• Root Cause Analysis
• Select Optimum Solution Quality Function
Business Opportunit
Case y
• Specific Deliverables 1
2
Gantt Chart/Project
Goal
ID
Statem
Plan
Project Plan
ID Customers
ID Needs
Statem
Project
ent
Task Name
Scope
Sep
Qtr 4, 1999
Communications Plan
3 ID CCRs
• Stakeholder Involvement
ent
4 Review
TS
10 Simulation
11 Cost Analysis
In Scope
PR
EN
12 Design Review
ImOCE
13 Procurement
14 Implementation
EMs
OVoces
pro S
veSmC
PRPr
Out of
IMon
Out of
enRE
Seati
scope
t AT
scope
ECSr
IO
OC
N
Process Management
PR
PROCESS OWNER
What is 6 Sigma?
Statistical Proven Quality Project Business
and Analytical Tools Management Philosophy
Methods
Customer
Reduce Satisfaction Delivery
Cost Performance
Manufacturing Product
ISO
Velocity Quality
6 SIGMA
Agenda
• Solar Turbines \ Caterpillar
- Brief background of company and products
• What is 6 Sigma?
- Background behind the methodology
" Business Case " Define Y = f(x) and list ! Variation shown ! Generate Multiple ! Pilot/Solution Results vs.
your measures graphically Solution Ideas Project Goals
" Opportunity (Cp, Cpk, SQL)
Statement " Deployment Map(s) ! Evaluate and select ! Budget and Benefits
that depict measure ! Performance solutions Case
" Goals locations gap(s) quantified
! Communicate team ! Potential Problem
" Scope " Operational ! Problem recommendations to Analysis
Definitions Statement(s) stakeholders
" Team Resourced & ! Detailed Implementation
prepared " Data Collection plan ! Process Map ! Link solutions to root Plan
Analysis causes
" High Level Project " Measurement System ! Communication Plan
(bottlenecks,
Plan Evaluation ! New Future State
value-add)
Process Map(s) ! Standards and
" High Level Process " Output Measure (Y), ! Potential Root Procedures
Map (SIPOC) Baseline Performance Causes quantified ! High Level
quantified and prioritized Implementation Plan ! Training Plan
" Quick Wins
(updated in Analyze)
! Root Causes ! Process Owner(s) ! Key Metrics and Process
" VOC Translated to Control System
CCR’s
" Financial How/How Validated confirmed
(updated in Analyze) Y= f (x)
! Projected Benefits ! BRM
" Communication Plan
" Process Owner(s) ! Financial Baseline ! BRM ! FRS Updated & 2/6 mo
" BRM Identified approved control meetings
" E-tracker updated " BRM ! FRS Updated scheduled
! BRM
! FRS Forecast ! Scheduled Transition to
Process Owner(s)
Entered
! Replication Opportunities
Lean 6 Sigma at Solar Turbines
Voice of the Customer
Voice of the Customer Key Customer Issue Critical Customer Requirements
Actual customer statements The real customer concerns, The specific, precise and
and comments which reflect values or expectations measurable expectation which
their perception of: regarding a product or a customer has regarding a
• An attribute of a product or service. Void of emotion or product or service.
service bias, the statement describes the • Mower starts within two pulls on
• An experience with a primary issue a customer may the cord
have with the product or • Mower starts with an effortless
product or service or its pull on the cord not exceeding
delivery services. Describes the
24” in length
• An encounter or experience surrounding the
attributes of the product or • Add additional menu items to the
experience with a business voice system (bad)
processes or representative service expected or desired by
• Customer reaches correct person
“This mower is way too hard to the customer. the first time within 30 seconds
start” Wants the mower to start quickly and
(good)
painlessly
“I’m always on hold or end up • Every design feature needed is
talking to the wrong person” Wants to talk to the right person built into the package
quickly • The software is fully operational
“This package doesn’t do squat” on the customer’s existing system
The software does what the vendor
said it would do
Y = f(x), - Measurement Plan
<Executing Special Requirements for Package Manifolds>
Business Sales Order
Customer Quality Structures
Managment Engineering
PAE
Reviews
customer spec
PAE
Handoff
X1
to PM/PE
Manufacturing
Elapsed Time • Total Process Time 1 4
Project Team
A Initial Hydro’s
(HMSTAT)
QE
YY Workflow A
PE Released
PID in PDSWIN
ME validates
key long lead
items and
YY
releases
Drawing
Drawing CKOM
(This loop may be repeated via Customer input
Forecast BOM
Package Cost
Release
Package Cost
Release
PE/PM/QE
throughout the project)
Reviews Initial
2
No
End
Subassembly
Inputs ITP Revisions? Yes
2
QE Revises ITP
No
End
PM Reviews/
QE
Revised ITP Approves ITP
Problem Customer
changes?
No
Resolution Yes A
Time
Business Risk Management
BRM 6 Sigma Risk Action Plan
Engineering Drawing Release
Updated On: 3/25/09
Key Risk
Risk Comments Actions Owner of Actions
Category
Risk Map
External
10
9
- Improve efficiency of release process
8
Financial
Delayed drawing releases - Project Team
- Provide ability to find lost/delayed
7 will increase probability of
items
expedite fees. - Group Managers
Significance
Brackets
Manifolds
Business Case
Solar’s volumes were increasing in 2006 and
were to grow by 30% in 2007.
r s
5%%'#.- ho
u ! !!
5%%'#.- 5%%'#.- 5%%'#.-
0 h
5%%'#.-
t
0
1 m o n
e r 5%%'#.-
5%%'#.- 5%%'#.- 5%%'#.-
p
5%%'#.-
Process Mapping Inputs something into text
Turbine Assembly Creates BOM for engine
routing (after test items
M.E. removed from engine after
test?)
ASL Coordinator
Inventory Specialist
Cashes PA engine Loose
ship parts 2 weeks from
PSS date “Final Date”
After Market
Shipping Packer #
Packs Loose Ship.
DATE
Gets Weights & Dimensions
Delivers W &D’s to Lisa &
Veronica, or does Alex?
DEBUR DRILL SAW Loads truck
Exits MCT
Lisa & Veronica Create Preliminary Pack
List
Receive W&D’s from
packer and forwards to TJ
Cox in Logistics
Creates Bill of Lading
office DAN'S TABLE BREAK WORK IN RACK 1 Box 2 3 4 5 WIP
Logistics Analyst
(TJ Cox)
Receives W&D’s, creates
shipping instructions
Sends shipping instructions
to Lisa & Veronica to create
the Bill of Lading, Final
Contacts Switzerland of
pickup date
OFFICE
1 20-Sep 69 72 3 2 3 0 9 4 2 2 3 0 0 1 278 13 21.38
Pack List????
Turbine Assembly
Assembles Engine, Test
Cell completes test, returns
engine
Borescopes engine, ???
Buttons up, sends to paint
Returned from paint
108 Gets M.E. to sign off, install
data plate
Photographer
(Joe Canales)
Takes Photo – Posts
on-line
2 21-Sep 82 94 105 10 0 0 0 5 5 1 7 3 0 0 7 319 13 24.54
3 22-Sep 67 90 102 3 1 0 0 6 4 1 5 4 0 1 5 289 7 41.29
Ancillary Shipping Lead Receives engine, ????, Sends Weights & Dims to
Loads truck
(Tommy) eventually packs Veronica & Lisa
Exit Logs
What else did we find? WIP
Weld Supervisor’s
Booths Weld
Office
Hydro Booths
Machine
Test Plasma
Shop
Burn Table
Structures Baseline
Sum mary for C 1
Anderson-Darling Normality Test
A-Squared 3.41
Baseline Performance
P-Value < 0.005
Mean
StDev
Variance
Skewness
20.780
8.001
64.012
3.3112
Brackets
Kurtosis 15.4189
N 50
Minimum 12.000
1st Quartile 17.000
Median 19.000
3rd Quartile 23.000
Median
18 19 20 21 22 23
Cycle time of
20.7 days
Baseline Performance Summary for Weld booth exits
A n d e rs o n -D a r lin g No r m a lit y T e s t
Pipes
A-Squa re d 0 .8 6
P -Va lu e 0 .0 2 6
Mean 5 .8 1 4 8
S tD e v 2 .9 7 2 0
Va ria n c e 8 .8 3 3 0
14/day @ 7 days/wk
S k e wn e s s 0 .6 3 9 2 0 9
Ku rt o s is 0 .1 6 0 0 7 2
N 54
M in im u m 2 .0 0 0 0
1 s t Q u a rt ile 3 .0 0 0 0
M e d ia n 5 .5 0 0 0
3 rd Qu a r tile 8 .0 0 0 0
3 6 9 12 15 M a x im u m 1 5 .0 0 0 0
9 5 % Co n fid e n c e In t e r v a l fo r M e a n
5 .0 0 3 6 6 .6 2 6 0
Median
4 5 6 7
Baseline Performance
16/day @ 7 days/wk
21
Baseline Part Shortages
700 40
Pipe 30
Data
Backlog 20
10
0
Small Gen_1 MarsTitanCSMD_1 C40-T70 CSMD_1 Large GS_1
Variation Analysis
Drilling Analysis
Summary for C1
Drilling
Anderson-Darling Normality Test
A-Squared 2.90
P-Value < 0.005
Mean
Outlier 71.625
StDev
33.398
Variance
Skewness
5130.118
3.3280
Analysis:
Kurtosis 11.6178
N Outlier due to common
14
Minimum 1.450
cause variation.
Median Summary
Copingfor C1
0 20 40 60 80
TIME STUDY Anderson-Darling Normality Test
A-Squared 0.21
COPING P-Value 0.832
Mean 50.314
Program # Diameter Tools needed Old StDev 25.815
Variance 666.432
O20 2"RUN 1.5"PIPE T18 14.33 Skewness 0.278043
O21 3"RUN 1.5"PIPE T18 14.52 Kurtosis -0.780435
Caterpillar: Confidential Green 33 N 14
O22 4"RUN 1.5"PIPE T18 23.32
Minimum 14.330
O23 6"RUN 1.5"PIPE T18 54.02 1st Quartile 31.458
O24
O25
8"RUN 1.5"PIPE
3"RUN 2"PIPE
T18
T18
87.30
34.17 20 40 60 80 100
Median
3rd Quartile
Maximum
47.775
69.228
97.240
Bracket
HistogramCycle Time
of MCT BEFORE
95% Confidence Interval for Mean
O26 4"RUN 2"PIPE T18 41.53 35.408 65.219
14
O27 6"RUN 2"PIPE T18 36.52 95% Confidence Interval for Median
Frequency
time per pipe
8
0
15 20 25 30 35
MCT BEFORE
Solar Success
PSO Success
PSO Packaging Value Streams
Structures
StructuresMulti
MultiGeneration
GenerationProject
ProjectPlanv
Plan
Safety +5S Order Velocity Chase Waste
Management
set,
Safety
sort, Cycle Time NVA Activities
shine, WIP caps reductions
Employee
standardize, Waiting
Engagement Visual Control Process Overproducing
sustain Boards Rework
Optimization
set, sort,
Safety Motion
shine, Green Change Response Processing
Employee
standardize, Management Time Inventory
Engagement
sustain Transportation
Lean 6 Sigma
Integrity Excellence Teamwork Commitment
;24<(0%%)
5S+1
;.-+-%2%.'
Quality
Safety
/($')6 /%')6 /7#.%
/'+.8+$8#9%)! /:1'+#.
Safety Velocity
Ergonomic
Heijunka
Scheduling
Heijunka Methodology
Visual Order
Management
WIP Caps
Over 1,200 pipes
in finished goods.
After
Employee Lead Kaizen Event
One Bin System
80%
Gas Fuel Module
50% cycle time reduction
cycle time
reduction Removed 14 feet of
weld per module
4,200
feet of
welding
4,000 feet annually!
Customer Interface Panel
annually
Feet of welding
Labor hours
Lean Machining
2 minutes A-Squared
P-Value <
3.48
0.005
Mean 0.46071
StDev 0.31936
Variance 0.10199
Skewness 1.56562
Minimum
0.50138
14
0.30000
1st Quartile 0.30000
Median 0.30000
3rd Quartile 0.48750
0.4 0.6 0.8 1.0
Maximum 1.05000
95% Confidence Interval for Mean
0.27632 0.64511
95% Confidence Interval for Median
Machinist
Eliminated
daily trips %) ? 8 8% 8)
( .! >+<:
=
2008
Barstow
8 Deadly Wastes
99% First Pass X-ray Inventory
Overproduction
Unused
Creativity / Over
Capability Defects Inventory Production
Lost opportunities due Production or rework of out- Excess raw material, Excess supply beyond
to poor safety and an of-specification parts work-in-process the requirements of
underutilized workforce or finished goods the next process
Excess Over
Waiting Motion Transportation Processing
Lost time due to Wasted movement Excess movement of Work that adds no value
poor product flow — made while working work-in-process to the customer or business
shortages, bottlenecks,
down machines MOTION
3EE#&%
Burn Barn
B
?
C * +<<%'1
D
/
@:$.)'+A<%
Hand
Writing
/ 7%8
"(:A<%)
1 #8%8
Waiting 110
Processing
74
Phase 1 Phase 2
Metrics Board
Monthly Forum for new
Employee ideas and
Communication feedback to
Meetings sustain process
control
Voice of the Customer
In 2006 the Structures department struggled to meet increased production levels;
their inability to support the package takt rates put the year's financial plan at risk.
The Lean 6 Sigma process improvements not only allowed Structures to get
caught back up and support our 2006 volumes, the improvements have allowed
Structures to successfully supported record levels in 2007, including our 2nd
quarter performance of one package a day, 7 days a week, for over 100 days
straight! This has been a phenomenal improvement!
Director
Packaging Systems Operations
Manager
Customer Services Project Management
"%1#-.)E($)
5+.:E+&':$#.-) >+<:%)*$(4(1#'#(.
K5L1
HHI)E#$1')4+11)J!$+0)
'%1')1:&&%11) C0&<%)'#2%)
$%8:&'#(.1)
G#1'%.#.-)'()(:$)
4%(4<%);3/)
;2A%88#.-)F)/#-2+)
&:<':$%)
5%%'#.-)(:$)
&:1'(2%$1).%%81)
Controls Assembly
Major 5S+1 Event
Lean Work Centers & Product Flow
After
Off-Skid Offices On-Skid Controls
Lean Work Lean Work
Support
Center Central Center Lean Work
Supermarket Center
Floor Stock
Assembled
Project
Assembled
Tested Kitting Tested
Panel Build Up
Final Test
Cross Wiring
Kit
Before After
Motion
Results and Benefits
By changing our internal product flow we were able to
change our external product flow
Before After
Off-Skid
Packaging Packaging
Off-Skid
On-Skid On-Skid
Controls Controls
17 20
Results and Benefits
32
11
7.5
8.5
Results and Benefits
37 days
12 days
Controlling the New Process
Mgr. Spvr.
PAC
ME Schedulers
Results and Benefits
32 7.5
8.5
11
65%
65 13%
87%
37 days 67%
12 days
29%
Drawing Release Project
Project Goals
Improve cycle efficiency
Purchased Part
Configuration
Management
Compliance
Technology
Information
Structures
Modeling
Drafting
Drafting
Project Drawings x x x
AutoCAD
Source Control Drawings x x x
Source Control Drawings x x x x
Specfication Control Drawings x x x x
CADDS
Procurement Control Drawings x x x
Purchased
Project Drawings x x x
Procurement Control Drawings x x x x
Project Drawings x x x x
Pro-E
Source Control Drawings x x x x x
Specfication Control Drawings x x x x
Procurement Control Drawings x x x x
Word Project Drawings x x x x
Source Control Drawings x x x x
x x
Manufactured
External Supplier
CADDS
Structures / PSOSUR
x x x
Structures / PSOSUR
x x x x
Pro-E
External Supplier
x x x
x x x x
Assy
Pro-E
Shop Floor Drawings
CADDS
Shop Floor Drawings x x x
Multiple Drawing Types and Processes
Access file on the S : drive. Info posted (read only) in livelink for
users (Ref http ://solarweb2.solar.cat.com/engstandards /
purchasedpartdrawings /index.html “Where is my drawing now” )
FIFO
OUTPUTS : new DEF item review
flag
OUTPUTS : AutoCAD Word Drawings Word Drawings
OUTPUTS : OUTPUTS : OUTPUTS : OUTPUTS :
Word Drawings WIP ________
OUTPUTS : OUTPUTS :
Drawings
VALUE ADDED TIME :
VALUE ADDED TIME : I I Dave’s Log
I I
VALUE ADDED TIME : VALUE ADDED TIME : VALUE ADDED TIME : VALUE ADDED TIME : VALUE ADDED TIME :
I
VALUE ADDED TIME :
Exits ________
Word and AutoCAD WIP ________ Au Do
toC WIP _______
Drawings AD es WIP ________ WIP ________
WIP ________ Dave’s Log
Word Drawings
a A
If Required tif fi D ra D a n or
Initiate LOA and copy le fr w ing d E s tr
a ig
Create Model if required om xce INPUTS :
Planner and Buyer s l ht
th e to
s ub C Manufacturing OUTPUTS :
m M?
INPUTS : Lead time and need date
INPUTS : itte r Configuration AutoCAD
Package DE Engineer Inbox Purchased Purchased
OUTPUTS : LOA – supplier begins Management and Excel VALUE ADDED TIME :
Inbox Part Part
Pro-E drawings tha t require modeling
I
VALUE ADDED TIME :
FIFO
Pro-E
Pro-E Drawings
Exits ________
CM Group initiates
release in ARP
Manufactured Parts
Pro-E and CADDS
Drawings for
INPUTS :
IT Inbox
gs
55 items in Kanban
w in
OUTPUTS :
10/21/08
D ra
D225 Drafting
Inbox Package DE
Inbox
Create part number in Prepare PPRD / RFC
Windchill and drawing packet DE reworks design as
Make log entry and Structures ME initiates
required per Structure’s IT checks structure , releases
Modeling complete drafting item setup in BAAN
INPUTS : INPUTS :
Required?
FIFO NEW
feedback drawing in Windchill and
INPUTS : INPUTS : INPUTS : sets DEF flag to Done
OUTPUTS : OUTPUTS : Pro-E Drawings Pro-E and CADDS
No modeling required OUTPUTS : OUTPUTS : Drawings OUTPUTS :
INPUTS :
DEF
Manufacture VALUE ADDED TIME :
Part
I PLN ENG
WIP ________ Purchased
Part
Exits ________ Drafting Log
WIP ________
OUTPUTS
:
:
INPUTS
OUTPUTS
:
:
INPUTS
OUTPUTS
:
Manufactured Part Flow (s) VALUE ADDED TIME : VALUE ADDED TIME : VALUE ADDED TIME :
: –
t
i
• Time Study
• Value Add Time
• Span Time
All data points are median values. • Cycle Efficiency
MS Word Based Purchased Parts
: –
n = 60
:
Time Span Cycle
(Minutes) (Hours) Efficiency
Design Services 1 0.5 0.0667
/
: –
t
i
Simple Process Change
Excel Excel
Design
Compliance
,
Services
INPUTS: INPUTS:
:
OUTPUTS: OUTPUTS:
:
: –
t
Design
Compliance
,
Services
INPUTS: INPUTS:
:
OUTPUTS: OUTPUTS:
:
: –
t
Release process
Drawing Types
- process flow
- decision flow
- drawing packet contents
- define drawing types
- data requirements
- applications
- expediting instructions
- nomenclature
- procurement rules
Training Material
Package Hardware
Certification
Requirements R.K. Bostick 5-05
Mech. DE Pkg DE
Copy Standard Copy Standard
Mech. DE Ancillary DE
Prunes Tree Prunes Tree Pkg DE Enclosure DE Frame DE Ancillary DE
Prunes Tree Prunes Tree Prunes Tree Prunes Tree
Generate Generate
In Crate Dwg Out Crate Dwg Design
Freeze
Release Release
In Crate Dwg Out Crate Dwg Generate
MID Drawing
Release MID
One Drawing VS Two
1000
Total Mechanical Drawing Hours
800
600
400
66 hrs
200
0
O&G PG
Bu sin ess Un it
One Drawing VS Two - Regression
Mech Dsn & Dft Hours Dependent Variable = Mech. Design and Draft Hours
18
Predictor P
12
A CS Revisions
ACS Revisions 0.047
6
60
Project Span Time (weeks) 0.566
Total Releases 0.043
45
Project Span (wk s)
Sales Order Items 0.055
30
Custom Features 0.196
4.5
3.5
Total Releases
2.5
300
200
Sales Order Items
100
80
40 Custom Features
0
300 600 900 6 12 18 30 45 60 2.5 3.5 4.5 100 200 300
One Drawing VS Two - Regression
Starting Equation: Total Hours = -140 + 12*(# of ACS Revisions) + 71.3*(# of Dwg Releases)
+ .97*(# of Sales Order Items) + 2.02*(# of Custom Features)
Analysis of variation after accounting for ACS Revs, Releases, Sales Order Items, and CF’s:
Factor P R-sq(adj) Impact of Different Levels
Certification .088 8.4% CE +268 / CENELEC +109 / NEC –66
Other factors (Custom Generator, Custom L/O Cooler, Enclosure, Fuel System, Marine, Inlet, and
Exhaust) showed no statistical significance.
Final Equation: Total Hours = -275 + 12*(# of ACS Revisions) + 71.3*(# of Dwg Releases)
+ .97*(# of Sales Order Items) + 2.02*(# of Custom Features)
+ 336(if CE) +175(if Cenelec) + 82 (if custom frame)
One Drawing VS Two
Boxplot of Normalized Hours
750
500
Normalized Hours
250
-250
-500
O&G PG
Business Unit
Content Differences
Content Differences
Isometric Views
Content Differences
Isometric Views
• Ε Exploded Views
Content Differences
One view
per sheet
Multiple views
per sheet
Content Differences
Elevation Views