Application of Lean Six Sigma at DTE Energy

Application of Lean Six Sigma at DTE Energy

Real People

Real Contributions

Real Successes

Real Possibilities

Shawn Patterson

Tom WitteOctober 20, 2005

2

DTE Energy: A Quick Look

• Provider of Energy Products and Energy Services located in SE Michigan

• $6.75B in Annual Revenue• Approximately 10,500 Employees (includes both non-

represented and union employees)

-Detroit Edison-Michcon-DTE Coal Services-Energy Trading-Energy Services-DTE Energy Biomass-Gas Production-Plug Power (33% ownership)

55% DetroitEdison

9% MichCon

25% EnergyServices

4% Energy Trading2% Coal Services

1% Biomass 4% Gas Midstream & Supply

2002 Net Income by Profitable Business

3

DTE Energy CI Timeline

1998 1999 2000 2001

DTE Energy Commits to creating DTE Energy Operating System

Director hired to start development of Operating System

Kaizen program launched to introduce Lean concepts

50 Facilitators Trained across the enterprise

Over 100 Kaizens conducted, Kaizens pushed in the organization

MERGER

4

CI Timeline

2001 2002 2003 2004

Operating Council formed charged with finalizing development of Operating System

Initial Operating System framework created with 2-year deployment plan

Black Belts trained, OS projects accumulate savings of $25MM

MERGERSix Sigma tools formally added to Operating System framework

DTE Energy Operating System becomes a way of life

2005

5

DTE Energy Operating System

HBS

Lean (TPS) Six Sigma

6

Operating System Framework

OperateConsistently

Measure, Learn & Analyze

DTE Operating Principles

• Safety of our employees, customers and general public

• Meet and exceed customer expectations

• Communicate honestly and effectively

• Hold each other accountable and reward success

• Drive decision making to the point of activity

• Gain high agreement on what and how

• Create a learning environment

• Internal collaboration and partnership

• Understand current reality through fact based measures

• Surface problems quickly

• Identify and eliminate waste• Focus on activities,

connections and flows

Continuously Improve

Vision: To establish DTE Energy as the premier regional integrated energy company by providing sustained earnings growth.

Corporate Goals

People Effectiveness

Process Excellence

Customer & Regulatory Relationships

Financial Effectiveness

Vision Principles Actions ResultsOperating SystemFramework

Plan & Engage

Develop People

Change Management

Lean Operations

Employee Engagement

Six Sigma Tools

Strategic Planning Corrective Action

Project ManagementEmployee Development

Operating System Toolbox

How We WorkHow We Think What We GetOperate

Consistently

Measure, Learn & Analyze

DTE Operating Principles

• Safety of our employees, customers and general public

• Meet and exceed customer expectations

• Communicate honestly and effectively

• Hold each other accountable and reward success

• Drive decision making to the point of activity

• Gain high agreement on what and how

• Create a learning environment

• Internal collaboration and partnership

• Understand current reality through fact based measures

• Surface problems quickly

• Identify and eliminate waste• Focus on activities,

connections and flows

Continuously Improve

Vision: To establish DTE Energy as the premier regional integrated energy company by providing sustained earnings growth.

Corporate Goals

People Effectiveness

Process Excellence

Customer & Regulatory Relationships

Financial Effectiveness

Vision Principles Actions ResultsOperating SystemFramework

Plan & Engage

Develop People

Change Management

Lean Operations

Employee Engagement

Six Sigma Tools

Strategic Planning Corrective Action

Project ManagementEmployee Development

Operating System Toolbox

How We WorkHow We Think What We Get

7

Operating System Tools

Lean Tools

•Kaizens•5S’s•7 Wastes•Error-Proofing•Visual Management•Flow Concepts•Andon•Standardized Work

Six Sigma Tools

•Basic Statistics•DOE•ANOVA•Pareto Charts•Histograms•Run Charts•Measurement System Analysis•Statistical Process Control

8

Delivery Methodologies

ProcessRedesign

Incremental

Sco

pe o

f Cha

nge

QuickEngagements

(<1 Week)

Long-TermProjects

(3-6 Months)

Event Duration

Kaizens

Waste Walks

CARs

5S Workshops

AAR’s

Learning Lines

9-Step Projects(DMAIC)

Very Effective ToolFor Service and Transaction Processes

9

DTE Energy Operating System

The DTE Energy Operating System starts with the vision of the corporation and then works through a set of principles and tools that govern how we will accomplish our goals.

Vision

Principles

Tools

Results

10

Real Contributions

20

45

105

120

$0

$20

$40

$60

$80

$100

$120

$140

2002 2003 2004 2005

Plan

Actual

Operating SystemSavings by Year ($MM)

The Operating System has been delivering increasing results since its inception in 2002

2005 YTD thru August

11

Real People – Investing in Continuous Improvement Personnel

DTE’s growth in savings corresponds with our investment in continuous improvement personnel which is now approximately 2% of our total employees

FulltimeCI Personnel

0

50

100

150

200

2002 2003 2004 2005

12

1300

20067

0

200

400

600

800

1000

1200

1400

Trained Specialist Fulltime CIPersonnel

Certified Black Belts

Real People – Investing in Training and Certification

13

Real Successes

Issue: A power plant needed to replace one of their main station transformers.

This process would normally take a minimum of 19 days. The power from this plant would need to be replaced by more expensive power from either inside or outside Detroit Edison.

Generation Planned Maintenance Outages

Solution: Set up Kaizen to develop a plan that would cut the outage time in half:

• Walked through process several times prior to shutdown (Eliminate Motion)• Preliminary prep to new transformer during delivery phase (Eliminate Waiting)• Lighting and other environmental modifications to enable around-the-clock work (Eliminate Waiting)

Outage reduced by 10 days;$3.5 M in savings

14

Real Successes

•The application of the DTE Operating System in our Customer Service organization has allowed this team to analyze and take action on the key inputs that were driving an increase in our uncollectible accounts.

Customer Service Collections

Operating System Tools Utilized

• Benchmarking

• Process Mapping

• Regression

• DOE

• Ideal State Workshops

• Visual Management

15

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

2005 YTD Actual

2004 YTD Actual

2005 YTD Uncollectible Expense as Percentage of Revenue

$45 M in savingsin 2005 YTD

Real Successes

16

Real Successes

•One of DTE’s Gas Operation Service Centers was consistently performing at approximately 20% less in productivity than the other centers.

•A Lean Learning Line was established at this facility.

Gas Operation Efficiency: Worst to First

17

Real Successes

Gas Operation Efficiency: Worst to First• Within 4 months the center was consistently outperforming the other centers by 10% - 20%.

Customers Served

0.70.80.9

11.11.21.31.41.51.6

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Weekly - April / May / June

Cu

sto

mer

s/h

r.

1st Year

2nd Year

18

Real Successes

•The use of Lean and Six Sigma has also been applied to reduce the number of injuries experienced by our employees. DTE Energy’s OSHA recordable rate has dropped every year for the past three years. 50% reduction in injuries.

Reduction in Injuries

0

1

2

3

4

5

6

2002 2003 2004 2005

OSHARecordable

Rate

Regression analysis demonstratedstrong correlation to deployment of:

• Injury root cause analysis investigation• Development & Discipline of SWI’s• 5S for housekeeping

19

Tour

TO M

ar

22.520.017.515.012.510.07.55.0

10

9

8

7

6

5

4

3

2

1

S 1.27162R-Sq 86.4%R-Sq(adj) 83.7%

Fitted Line PlotTO Mar = 13.07 - 0.5441 Tour

O

SH

A R

ate March

Jun

April

May

July

AugustSeptember

OSHA = 13.07 – 0.5441 Score

5 S for Safety

Real Successes

20

Real Successes

•The Electrical Meter Repair Shop applied Lean concepts of waste elimination throughout its facility.

Waste of Motion Waste of Waiting Waste of Transportation

•There were over 40,000 meters waiting to be repaired on the floor at the shop. In just six months, the shop increased productivity by 250% while reducing the cost of the repairs by 75%. Today, there is virtually no backlog of meters to repair saving the company $480,000/year.

Electrical Meter Repair Shop

21

Electrical Meter Shop Improvement

Space Constraints

Job Shop Processing

Growing Backlog > 40K Meters

Operating System Tools

Straight–line Processing

Uniform Work Stations

5S

Employee Engagement

4-Gate/9-Step Project

Productivity: 40/day/FTE Cost/Meter: $12



Phase I

Phase 2

Phase 3

22

Real Successes

• Fossil Generation Used Operating System tools to closely analyze the measurement of SO2 at our Fossil Generation plants.

• Based on this analysis, countermeasures were deployed that resulted in improved accuracy of our emissions measurements and correspondingly a reduction in the amount of SO2 credits which were needed.

• This transformed into a $3.9 MM net savings in 2004 and $4.5 MM is expected to be saved in 2005 and beyond.

Emissions Reporting Accuracy

23

Real Successes

Observation

Perc

ent

Erro

r

280260240220200180160140120100806040201

5

4

3

2

1

0

-1

-2

-3

-4

MPP 1-2 Stack CEMS SO2 Calibration error

Emissions Reporting Accuracy

Reducing variation is saving DTE

$4.5MM per year

24

Real Successes

•The DTE Energy Electrical Shop utilized a series of Lean Kaizen events to eliminate or reduce the waste of transportation, waiting, motion, and over processing in the process to repair pad-mounted transformers. The results of the applying lean concepts to this process were:

Original Process New Process32 days of total process time 22 hours12.5 hrs of processing time 6 hours 5,800 feet of travel per transformer 3,000 feet

Annual Savings > $400,000

Transformer Repair

25

Real Successes

• Re-engineered process

• Total cycle time for repair reduced from 30 days to 1 day

• Cost reduced by 51%

Transformer Repair Costs Cut in Half

26

Real Successes

Returned Mail Accounts

4,096

1,145

247

1,4461,659

375217

997644

846828

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

2/12 5/6 7/5 8/9 9/22 11/2 11/9 11/18 11/29 12/5 12/12

2004

# o

f A

cco

un

ts

Reduced by 3849 incidents per month or 94% since implementation of new process.

Bill not delivered due to inaccurate address:$800,000 in revenue increase

27

Increasing # of accounts in CSB with returned mail blank mailing address

New scan / update technology,

Post OfficeInvalid ReturnedMail

Customer Avoids Paying Bill

Lack of Business Processes to Update Blank Mailing Addresses

Initial Mailing Address Incorrect

Processes not defined or documented

Ownership & Resources

Gas/Electric Matched Accounts with Good Mailing

Business Impact Not

Communicated

No Clear Sponsor to Support Resolution

Account is <60 Days Past Due

Y X 1 X 2 X 3 X 3 X 4 )

Y X 1 X 2 X 3 X 4

= f (

= f ( X 6 )

Invalid Mailing Address Process(5 codes)

Process to resolve valid mailing address issues (14 remaining codes)


Y X 1 X 2)= f (

(Process operational in Address Mgt. – for codes NSS, NSN, OTH, IA, & ODL) )

Y

Account is >60 Days Past Due


X 1 X 2)= f (

(Normal Collection Process will resolve address exception upon customer contact)

Customer Phone # - No Mailing Address

X 5

Project Focus Area
















Ownership & ResourcesOwnership & Resources



Business Impact Not

Communicated

Business Impact Not

Communicated





Y X 1 X 2 X 3 X 3 X 4 )

Y X 1 X 2 X 3 X 4

= f (

= f ( X 6 )







Y X 1 X 2)= f (

(Process operational in Address Mgt. – for codes NSS, NSN, OTH, IA, & ODL) )

Y





X 1 X 2)= f (

(Normal Collection Process will resolve address exception upon customer contact)



X 5

Project Focus Area



Real Successes

28

•DTE’s ability to restore power quickly to our customers drives both higher customer satisfaction and savings for our shareholders. Through detailed analysis of our restoration process, the average time to restore a customer has dropped by 35% during the past year and has resulted in over $2 MM in savings YTD.

Electrical System Restoration

Real Successes

29

Step 1: Project Identification

• Historical CAIDI (Customer Average Interruption Duration Index) Performance:

Customer Average Interruption Duration

233

256276

263

216

244

206

243235

150175200225250275300

1996 1997 1998 1999 2000 2001 2002 2003 2004

Min

utes

4th quartile

30

Step 3: Current State Analysis – Restoration Time

193.3

4.5

386.0

25.0

75.6

43.131.4

13.1

All Events 1000 + 501 - 1000 201 - 500 101 - 200 26 - 100 2 - 25 1

17,884 447 407 475 590 1,953 3,166 10,814

1,643,514 938,755 318,198 153,398 84,377 100,993 36,979 10,814

Event Count

Customers Affected

233 206 237 281 296 311 355 419Restoration Time

Event Size (customers)

• 80% of the Customer-Minutes are caused by 7% of the events

• Improving these events by 10% would reduce the overall restoration time by 19 minutes

Customer Minutes

(Millions)

• 60% of the events are 1 customer outs.

• Reducing time for these 10,814 events to 180 min, would drop overall time by only 1 minute

31

Step 3: Current State Analysis

The Trouble Dispatch project was broken into two projects: Large Outage Management project, and Data Integrity – Audit Process

Restore Customers in 3 Hours or less

Time to Repair

Y = f ( X1 , X 2 , X 3 )

Y = f ( X4 , X5 , X6 , X7 )

Time to Dispatch

Resource Availability

ORC / Dispatcher

skill set

Y = f ( X8 , X9 , X10 , X11 )

Large Outage Mgmt

152 CAIDI = 233 66 15= + +

Resource Availability

Crew Allocation

Crew Alignment

Crew Utilization

Project Focus

15 min Restoration

Improvement Target

Visual Management

Time to Dispatch

Data Accuracy

Large Outage Mgmt

32

Step 4: Define Ideal State

200 - 1000 1000 +l Send Supervisor l Send Supervisorl LS l LS

A l B l Bl If indication of Station

Lockout, send Operator

l If indication of Station Lockout, send Operator

l B (and SS if available) l B (and SS if available)

B l l Call in LSl Call in Supervisor

l LS (and SS if available) l LS (and SS if available)

B l Call in B l Call in Bl l Call in Supervisor

l If B crew on-call, bring in B l

l

Cl l Call in Supervisor

Bring in On-Call and additional crew to make a B and LS

Crews / Personnel on site (typical M-F Day, or shift with multiple crews on site)

B Crew on Shift (no LS on site)

Notify Supervisor (supervisor discretion to come in)

No Crews on site

Notify Supervisor (supervisor discretion to come in)

If LS on-call, bring in LS and also call in a BNotify Supervisor (supervisor discretion to come in)

LS on Shift (no B crew on site)

• Developed Large Outage Dispatch Guidelines

– The guidelines focus on getting the right personnel to the event at the start, rather than waiting to see which resources may be needed.

– Driving accountability and performance by getting field supervisors out to the events from the beginning.

Number of Customers Out

33

Step 5: Identify Gaps & Countermeasures

Process Step

Key Process

Input

Potential Failure Mode

Potential Failure Effects

SEV

Potential CausesOCC

Current Controls

DET

RPN

Actions Recommended

Event Dispatching

Field Crew Availability

Lack of crews to dispatch event to

Increased dispatch time & outage duration

5No crews scheduled to work in field

9 Shift schedules 7 315

ST - dispatch guidelines, LT - Analyze trouble resource allocation

Field restorationField crews, field supervision

Crews not working optimally

Increased outage duration, higher expense

7Lack of field supervision presence, especially on off shifts

7Feedback through OT monitoring, ROC monitoring

6 294Increase field supervisor presence, utilize Lg outage disp guidelines

Event Dispatching

Outage data, ORC/RSS

ROC personnel (ORC, RSS) not aware of large outage event

Increased dispatch time & outage duration

5Periods of high workload, multiple priorities, lack of a centralized notification process

5Individual InService installations

6 150

ROC Visual mgmt project - install central displays to notify/track large otuage events

Event Dispatching

Field Crew Availability

Not utilizing ST crews in adjacent service territories

Increased outage duration, higher expense

5Poor communication between ORC's

3 None 6 90Include in the Lg Outage Disp guideline SWI's

Field restoration Field CrewsOptimal restore before repair opportunities not captured

Increased outage duration

5

Field crews making decisions on less than complete information. Poor communication between ROC and field crew

3Large Outage / Long restoration AAR's

6 90Send field supervisor to as many large outage events as possible

Trouble Analysis InService dataEvent rolls-up to the wrong point

More customers reported out than actual

1Customers not assigned to the correct device

7 CARS, audit process 5 35

Data EntryPartial Restorations

Partial restorations not captured properly in InService

Increase to reported outage duration

1 Data entry errors 7 Audit process 3 21

Inservice data accuracy report project (J Koenders), Data Audit Project (S Putrycus)

Utilized an FMEA (Failure Modes Effects Analysis) to identify gaps to ideal state and ensure that countermeasures were developed to address them.

34

Step 7: Test, Refine, & Implement

• Conducted After Action Reviews with North ROC and South ROC (trouble dispatching and operating centers)

• Focus was on reviewing of specific events where the guidelines were and were not used as designed.

• Developed a report outline for dispatch guideline compliance and effectiveness

35

Step 7: Test, Refine, & Implement

Expect year end restoration time to be under the year end target of 195 min

CAIDI YTD 2004 vs 2005

100120140160180200220240260

Jan Feb Mar Apr May Jun

CA

IDI

(min

)

2004 2005

79 min

36

t-Test: Two-Sample Assuming Equal VariancesRestoration time for large outages

2004 2005Mean 235.3 159.7Variance 61,758.6 22,524.6 Observations 457 310Pooled Variance 45911.1Hypothesized Mean Difference 0df 765t Stat 4.80P(T<=t) one-tail 9.721E-07t Critical one-tail 1.647E+00P(T<=t) two-tail 1.944E-06t Critical two-tail 1.963E+00

2005 large outage CAIDI performance is statistically different than 2004. . . About a 1 in 1,000,000 chance that the change is due to random variation

Step 8: Measure, Verify, and Control

37

Step 9: Celebrate the Success

• Restoration performance this year has significantly improved and should finish the year well below Distribution Operation’s target of 195 minutes.

Customer Average Interruption Duration

235 243

206

244216

263 276256

233

170 YTD June

150175200225250275300

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Min

utes

38

Real Possibilities - Combining the Power or Lean and Six Sigma

Shareholder&

Customer Value

Lean • Eliminating Waste• Focus on Process

Activities, Connections & Flows

• Results in Speed of Processes and Time Reduction

Six Sigma• Driving Consistency• Eliminating Process

Errors and Variation• Results in Capability

& Reliability

$ $

$200 M / Year

39

End of Presentation

Questions ?

Application of Lean Six Sigma at DTE Energy

Documents

framework dte energy

dte energy ci timeline

dte energy electrical

real successes bill

tools lean tools kaizens

original process new

lean concepts

days of total process