Breakthrough Process Improvement

Optimizing Lining Material inventory Lean Six Sigma Project

DMAIC Methodology

Introduction

Dubai Aluminium ("DUBAL") owns and operates one of the world's largest aluminium smelters

Vision: To be one of the best companies in the global Aluminium industry

Products and services:

– 330 individual products to 300 customers

– More than 1 million metric tonnes /50 countries

– 3800+ employees

– 34 nationalities

In line with the vision , continual improvements are

the driving factors for success in DUBAL

Potroom Operation & Supply in DUBAL

DUBAL OPERATION

SMELTER

CARBON POTROOMS

CAST

HOUSE

POWER PLANT

SUPPLY

WAREHOUSE PURCHASE

GREEN MILL,

BAKING KILN,

RODDING & DOCK

POTROOMS WAREHOUSE STRATEGIC

MATRIAL

PURCHASE

M A I C D Project

Beginning

Define

Measure

Analyze

Improve

Control

Focus on value addition through operational and service excellence

CORPORATE SOCIAL

RESPONSIBILITY

SECURE A

DIVERSIFIED ENERGY

SUPPLY

RECOGNITION AS EMPLOYER OF CHOICE

ACHIEVE OPERATIONAL

EXCELLENCE

VISION

To be one of the best companies in the global Aluminium industry

Driv

er

s

Ou

tco

mes

P2 P3

P1

MAXIMIZE LONG TERM SHAREHOLDER’s VALUE

MARKET & BUSINESS

DEVELOPMENT

P10

P8

P9

L1 L2 L3 L4

P4

CH

P7

P5

P6

EXCEED CUSTOMER EXPECTATIONS

C1 C2 C3

P11

P&D Smelter Support EHS

F1 F2 F3 Optimize

Working

Capital

F4

Improve Cash flow

Project was aligned to Corporate

Strategic objective

The Business Case: Project Selection

Reducing lining material inventory

variation and Optimize Inventory

Inventory Plays a Major Role in Aluminium Smelting

Define M A I C

Pot Lining and Electrolysis Cell

Adapted from: Relining Rationale.ppt, Morten Sorlie, Elkem

Electrolyte

Molten Aluminum

Cathode with Collector Bar

Frozen Bath

Dense

Refractory

Bottom Insulation

(Lining Material)

Strong Steel Support Structure

Carbon cathode block Carbon

anode

Each set of lining material cost’s approximately 0.6 mAED

Different Designs - D18 - CD20 - D20

- DX

Project Charter

Problem Statement:

In H1-2009, Lining materials inventory variation against the target (20 pot sets for each design), in warehouse was very high (parameter of

difference distribution mean=20.8, std. dev.29.5) and this variation was leading to Having variation between different materials in one BOM design

High inventory in some items

Very low inventory in other items

Generating expiry of some shelf life items

Define

-10

10

30

50

70

90

110

130

150

April

Min Inv

Variation in inventory

Large variation in inventory of lining materials against target.

High variations in

the lining inventory

against target

Define

Cell-lining PROCESS S

UP

PL

IER

S

CU

ST

OM

ER

S

Cathode Blocks

Collector Bars

Ref ractories

Pig Iron

Ramming Material

INPUTS PROCESS OUTPUTS

Requirement

Schedule from

Potline Ops &

Process Control

MRP by Warehouse

and PR to purchase

PO from purchase

to supplier

Lining material

supplies from

supplier to

warehouse

External Supplier

f or lining material

Purchase

Issue of Lining

Material and Inv entory

stock at warehouse

Cell Lining Inventory

Cell Lining

Project Charter -SIPOC

Material Request from

area Material Issue

Through SIPOC all stakeholders and areas related to the problem

identified

Project Charter

Project Target:

Reduce average excess inventory from avg. 21 to ~12 pot sets by

Reducing lining material inventory variation by 50% and

Optimizing inventory

Define

•Warehouse, Purchase, Shipping, Pot repair. Project Scope

•Potlines Project out of

Scope

•Lining inventory CTQ

•Variation in inventory at the time of physical verification

Process Indicator

•Reduction in total inventory cost by ~5 million AED

Measurable Benefit

Project Team

Team Members

Project Leader

Project Sponsor Dr Ali H.A.M. Al Zarouni

V.P. Smelter operations

Dr Maryam Mohamed AlJallaf

Snr Manager PCPR & PCCL

Define

Ibrahim Baggash Snr Manager Smelter Production Services

Najm Alawadhi Snr Manager Purchasing – Strategic Materials

Abdul wahid Al Madani Snr Manager Supply Chain

Abdulla .A Manager Cell Relining

Asim Hassan Manager Warehouse

Mohamed Tawfik Boraie HOD Process control Cell Lining

Haiedeh Marashy Warehouse

Saif Mohamed Snr Planner Smelter Reductions

Daniel Whitfield Snr Engineer Process Control Potrooms

Andries Louw Snr Engineer Process Control Potrooms

Narendra Singh Rana Plant Industrial Engineer

Team selected is a cross-functional team involving all areas

related to the project.

Stake Holder Model : ARMI Chart

Approver Resource Member Interested Party

Dr Ali H.A.M. Al Zarouni Warehouse Dr Maryam Mohamed AlJallaf VP-Smelter

Purchase Ibrahim Baggash VP-Finance

Smelter Services Najm Alawadhi VP-Supply

Process Control. Abdul Wahid Al Madani

Industrial Engineer Abdulla .A

Asim Hassan

Mohamed Tawfik Boraie

Haiedeh Marashy

Saif Mohamed

Daniel Whitfield

Andies Louw

Narendra Singh Rana

Roles of Stakeholders have been identified through ARMI chart.

Project Schedule

Phase Deliverables End Date

Define • Project Charter

• SIPOC & Process Map

• Goal Setting

• ARMI Chart

• Schedule

June’09

Measure • Unit Definition

• Data Collection Plan

• Baseline

July’09

Analyse • Brainstorming

• Pareto August 09

Improve • Solution Selection Matrix

• Implementation Plan November 09

Control • Control Charts

• Control Plan

• Review Meeting

December’ 09

Define

Defined project tollgates and milestones – agreed by

the team

Potroom

Process Control

Warehouse

Purchase

Pot Repair Supplier

Purchase

material

Deliver material

to inventory

Issue material

for production

Finished goods

Build ready

shells to

inventory

Issue for pot

replacement

Process Map

Cell Relining Process

Point of CTQ Measurement

Data Collection Plan

Data Frequency Source Responsibility

Variation in inventory from target Monthly

PCCL Records Mohamed Tawfik

Measure

•Pot set (Quantity of lining materials available at end of the month)

Unit Definition

• (One month consumption + buffer) ± 40 sets (set by pot failure prediction) Specification

•Any data point out of specification (± 40 sets). Defect Definition

Data Collection Plan

MSA – Data Verification

Physical verification of lining material inventories performed monthly.

Consumption data verified between PCCL databases and iRPMS.

GR&R is not applicable as the measurement is count of inventory. Hence just data verification was done to ensure data accuracy

Measurement system/ Data found satisfactory

Baseline

Cpk 0.27

PPM (Exp. Overall) 278,115

Very Low Cpk and High defect ppm – high number of

data points out of specifications

Identify Possible causes Analyze

Problem

• Variation in inventory & non-optimized inventory

Tool Used

• Brain Storming

Possible Causes

1.Prediction Model Error

2.Variation in delivery

3. Change in schedule for early delivery by procurement planning

4.Lead time

5.High Buffer level

Filtering the Root causes Analyze

Possible Causes 1.Prediction Model Error

2.Variation in delivery

3. Change in schedule for

early delivery by procurement planning

4.Lead time 5.High Buffer level

Tool Used 1. Process Capability

2. Why why & Histogram

3. Why why & VSM

4. Process Knowledge

5. Inventory Model &

Control chart

Why Used 1. To check Model is

capable of giving

desired accuracy

2. To identify root cause and validate

3. To identify root cause and validate

4. To study controllable or uncontrollable cause

5. To identify whether

buffer level acceptable or not

This summarises the tool used for filtration of root causes from

possible causes

Possible Cause 1: Prediction Model Error

(Process Capability of prediction)

Analyze

Conclusion:

As Capability analysis and Control chart suggest that prediction error is

within control, hence it is not a root cause..

Hence prediction error is well with in the tolerance of ± 10 pot sets. Expected error ppm is 4044, which is acceptable for the above process.

Possible Cause 2: Variation in Delivery

(Why-Why Analysis)

Analyze

Why • Why variation in delivery occurs?

• Due to variation in delivery date.

Why • Why variation in delivery date occurs?

• Due to poor commitment from supplier.

Variation in delivery (Delay) of lining material shipments Histogram

2008

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

-180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180

Delay (Days)

% o

f s

hip

me

nts

Series1

Hence, Poor commitment from supplier is root cause for inventory variation.

Root cause Validation

High deviations in

delivery dates

Possible Cause 3: Change in schedule for

early delivery by procurement planning (Why-Why Analysis)

Analyze

Hence No visibility of Finished Goods and WIP inventory to procurement planning

is root cause for inventory variation .

Why • Why change in schedule occurs?

• Due to lower than planned inventory at warehouse.

Why • Why lower than planned inventory at warehouse?

• Due to no visibility of WIP and Finished Goods inventory

to procurement planning.

High WIP& F.G. inventory (VSM Analysis)

Root cause Validation

Possible Cause 3: Change in schedule for

early delivery by procurement planning (Why-Why Analysis)

Analyze

Hence No visibility of Finished Goods and WIP inventory to procurement planning

is root cause for inventory variation .

High WIP& F.G. inventory (VSM Analysis)

Root cause Validation

Possible Cause 4: High Lead Time

Analyze

Hence, lead time causes inventory variations but it is not a controllable root cause,

although some other actions like - Making some material as consignment stock, were proposed by the team.

Why • Why high lead time causes inventory variation?

• Due to high max to min difference between inventory levels

Why • Why high difference between inventory levels?

• Due to supplier delivering material in bulk and we need to cover for long lead time.

Why • Why supplier delivering material in bulk?

• Due to material being exclusive to us supplier produces material at certain quantity.

Lead Time Mostly 4-6 months

Possible Cause 5: High Buffer level

Analyze

Hence no scientific basis for buffer level setting is a root cause for high

inventory.

Variation in actual process +/- 10 sets

* Value depends on material type and its lead time.

Why • Why high buffer level?

• Due to wrong setting of buffer level

Wrong Setting

Root Cause Validation

Why • Why wrong setting of buffer level?

• Due to no scientific basis for buffer level setting

Actual pot failure variation is within +/- 10 pot sets and we are keeping a buffer of 20 pot sets

Filtering of Root causes Analyze

This summarises the filtration of root causes from possible causes

Variation in inventory and high

buffer level

1.Prediction Model Error

2.Variation in delivery

Poor commitment from supplier

3.Change in schedule for early

delivery by procurement

planning

No visibility of WIP and Finished Goods

inventory to procurement

planning

4.Lead Time

5.High Buffer level No scientific basis

for buffer level setting

Summary of final root causes

1. Poor Supplier Commitment

2. No visibility of Finished Goods & WIP

3. No scientific basis for buffer level setting

Brainstorming: List of Possible Solutions

Root Causes Counter-Measures/ Solutions

1.Poor commitment

from supplier

A. Agree with supplier at year beginning for full year delivery

schedule.

B. Follow up with supplier before each shipment.

C. Communicate to user about any expected variation.

D. Change packing standard.

2. No visibility of

Finished Goods and

WIP inventory

A. Change concept of inventory to overall process inventory,

considering: Raw material, WIP and Finished Goods.

B. Consignment stock from major lining material and to be paid

only when used.

3. No scientific basis

for buffer level setting

A. Dynamic buffer for each design to be calculated every year

based on past process variation.

Improve

Solution Selection Matrix Improve

Root Cause Solution Cost Difficulty in Implementa

tion

Rating (Cost x

Difficulty)

Poor commitment from supplier

Agree with supplier at for full year delivery

schedule.

5 3 15

Follow up before each shipment 4 4 16

Communicate to user about any expected variation

5 5 25

Change packing standard. 2 1 2

No visibility of Finished

Goods and WIP

inventory

Change concept of inventory to overall process inventory, considering: Raw material, WIP and Finished Goods.

5 3 15

Consignment stock from major lining material.

3 1 3

No scientific basis for buffer level setting

Dynamic buffer for each design to be calculated every year based on past process variation.

4 4 16

RATING SCALE COST

HIGH 1 2 3 4 5 LOW

DIFFICULT IMPLEMENTATION

Implementation Plan Improve

Actions were monitored and implemented as per the plan

No. Particular Action Points Responsibility Target Date

Completion Date

1 Poor commitment from supplier

Agree with supplier at for full year delivery schedule.

Najm AlWadahi Nov 2009 Dec 2009

Follow up before each shipment

Najm AlWadahi Jan 2010 Jan 2010 (Continuous )

Communicate to user about any expected variation

Najm AlWadahi Nov 2009 Nov 2009

2 No visibility of Finished Goods and WIP inventory

Change concept of inventory to overall process inventory, considering: Raw material,

WIP and Finished Goods.

All Jun 2009 Jun 2009

3 No scientific basis for buffer level setting

Dynamic buffer for each design to be calculated every year based on past process variation.

Maryam Mohamed Aljallaf

Oct 2009 Oct 2009

Process performance before and after implementation

BEFORE AFTER

Before After

Cpk 0.27 1.15

PPM (Exp. Overall) 278,115 4,326

Improve

Significant improvement in excess lining inventory variations

Supplier Commitment

Process performance before and after implementation

Lining Martial Physical inventory actual vs Target

Improve

Significant improvement in excess lining inventory variations

Control

Sharing of draft annual plan with all the stakeholders in year beginning.

Regular Bi-weekly meeting between stakeholders.

Department objective for controlling lining material inventory reported and reviewed through BSC.

MRP procedure modified for not reacting to common cause variations.

BPR Report with month by month objective vs actual performance on KPIs

To sustain the improved results following control methods were institutionalized in the system and being continuously monitored.

Modified MRP Procedure

Control Plan

Parameter Measurement

Method

Responsibility Actions: If Out Of Control

Buffer level Difference between Actual failure and predicted failure within control limit

Cell Lining Process Control, HOD

Plant Industrial Engineer, Smelter

Sr. Mgr., PC PR & CL

Re calibrate the level if some unexpected event happens.

Physical Inventory

Physical Inspection Cell Lining Process Control, HOD

- Inform all stakeholders and decide if changes are required in future delivery schedule

Control

OCAP and responsibilities identified in case of out-of-

control process

Stakeholders involvement in Control Phase

Stakeholders Area Role in control

Top mangement Reviewing performance.

Dr Maryam Mohamed AlJallaf Process Control potroom and Cell Relining

- Setting dynamic buffer. - physical stock taking, monitoring

and reporting.# - Monitoring and communicating

consumption.

Najm Alwadahi Purchase Strategic Materials Implementing MRP recommendation with supplier.

Asim Hassan Warehouse MRP control based on data from all sources.

Ibrahim Baggash Smelter services Reporting consumption in SAP.

Results Achieved & Financial Benefits

Working capital reduction by 11.3 million AED

audited by finance.

OTHER BENEFITS:

Reduced greatly fire fighting situations.

Reduced greatly expiry of shelf life items.

Improved storage space utilization.

Avg 20.8 excess

pot sets

Avg 2 excess

pot sets

Project Closure

• Learning from the project

– Enhanced understanding of the process through Value Stream Mapping.

– Lean and Six Sigma integrated project a unique case study.

– Cross functional team involvement.

• Horizontal deployment

– Learning from the project applied to other process materials.

• Further refinement

– Automated MRP project in SAP 2013.

• Team recognition

– Team was recognized monetarily through

performance enhancement project (PEP) program.

– Presented in quarterly review meeting and

appreciation from VP smelter operation.

Why this project is an excellent improvement example

• Impacted Dubal’s cash flow positively and reduced it by 11.3 mAED.

• Complex project involving multiple functions, having specialized and

customized high value materials (minimum 51Million AED at 2009).

• Achieved reduction in inventory in spite of limitation of 6 months lead

time.

• Transformation in inter department relationship and team work

(namely Smelter operation, warehouse and purchase).

• Great success story in service area (not only technical).

• This project is a live example of how lean application with six sigma can

provides a unique blend of results.

Q & A

Thank You

Lining Mat. inventory before power outage (Oct 2008)

0

10

20

30

40

50

60

70

80

90

100

Board

,insu

latin

g

232x

210x

20 c

alcium

silic

ate

Brick,

insu

latin

g 23

0 x 23

0 x 64

mm

Brick,

insu

latin

g 30

5 x 11

4 x 76

mm

Brick,

fire

230

x 11

4 x

64m

m

Brick,

fire

305

x 11

4 x

76m

m

Brick,

SIC 4

85 x

258

x 7

6mm

Car

bon

Sidew

all b

lock

, Gra

phitic

Insu

latio

n Boa

rd -

Multip

le

Ref

ract

ory

Brick -2

30x2

20x7

6

Car

bon

Sidew

all b

lock

, Gra

phitic

Insu

latio

n Boa

rd

Slab,

fire

250 x 5

00 x 7

6mm

Ref

ract

ory

Brick -1

20x2

10x1

00

Gra

phitiz

ed C

atho

de 3

150

Insu

latio

n bo

ard s

et

Gra

phitiz

ed C

atho

de 3

250

Col

lect

or B

ar,ro

und

4320

Slab,

fire

250 x 5

00 x 6

4mm

Brick,

fire

230

x 11

5 x

50m

m

Brick,

fire

232

x 21

0 x

100m

m

Brick,

fire

500

x 50

0 x

64m

m

Semigra

phitic

Cat

hode

315

0x440

mm

Col

lect

or B

ar,ro

und

Shape

d Ref

ract

ory

Brick -2

32

Ref

ract

ory

Brick -1

20x2

20x1

00

Slab,

fire

500 x 5

00 x 7

6mm

Brick,

fire,

HD, 2

30 x

115

x 6

0mm

Brick,

fire

195

x 11

5 x

60m

m

Brick,

SiC 3

80 x 27

3 x 76

mm

Sidew

all b

lock

,Gra

phitise

d

Oct

Min Inv

• High inventory Vs min Required

• High Variation (not usable)

• Not considering ready shells

Lining Mat. inventory after power outage (March-April 2009)

0

10

20

30

40

50

60

70

80

90

100

Boa

rd,in

sulatin

g

calcium silica

te (sh

ape

d)

Bric

k,insu

latin

g 23

0 x 64

mm

Bric

k,insu

latin

g 30

5 x 76

mm

Bric

k,fire 23

0 x 11

4 x 64

mm

Bric

k,fire 30

5 x 11

4 x 76

mm

Sem

igra

phitic Catho

de 3

100x

400

Insu

latio

n Boa

rd - M

ultip

le

Refra

ctory Bric

k 230

x76

Slab,fire 250

x 7

6mm

Refra

ctory Bric

k 120

x210x

100

Graphitiz

ed C

atho

de 315

0

Insu

latio

n bo

ard set

Graphitiz

ed C

atho

de 325

0

Collector B

ar,ro

und 43

20

Slab,fire 250

x 5

00 x 6

4mm

Bric

k,fire 23

0 x 11

5 x 50

mm

Bric

k,fire 23

2 x 21

0 x 10

0mm

Bric

k,fire 50

0 x 64

mm

Collector B

ar,ro

und

Sha

ped Refra

ctory Bric

k 232

x100

Refra

ctory Bric

k 120

x220x

100

Slab,fire 500

x 5

00 x 7

6mm

Bric

k,fire,HD, 2

30 x 115

x 60m

m

Bric

k,fire 19

5 x 11

5 x 60

mm

Bric

k,SiC

380 x 27

3 x 76

mm

Car

bon Sidew

all b

lock

,Graph

itise

d

March

April

Min Inv

NOT ONLY TO COMULATE INVENTORY TO COVER THE RISK, BUT LOWERING RISK BY IMPROVING DELIVERY CONTROLS.

Proc Cont. & Supply

P1

Cathode Casting

P2

shell Demolition & Repair

115 Cathode 6.05 pots

Lining material Supplier Material

Potrooms

P3

Cell Construction

D18

18 pots

CD20

18 pots

D20

14 pots

P4

Cell Change Over

846 Cathode

49.76 pots

320 Cathode

16.84 pots

21454

SiC 124.7 pots(22.5+50.3+51.8)

1412 Col Bars

78.4 pots (24.4+54)

521 Col Bars

27.42 pots

D18 CD20 D20

D20 76

4 pots

115 Cathode

6.05 pots

846 Cathode

49.76 pots

CD20

28 1.47 pots

D18 118

6.35 pots

D18 (Days) 0.7 NVA (days)

VA (days)

P rod. T ime= 9.9 days

Door to Door lead time

=590.7 days

POT LINING CURRENT STATE MAP

NVA (days)

VA (days)

P rod. T ime= 9.9 days

Door to Door lead time

=255.4 days

NVA (days)

VA (days)

P rod. T ime= 9 .9 days

Door to Door lead time

=231.7 days

CD20 (Days)

D20 (Days)

58 POT SETS

56 POT SETS

49 POT SETS

P1

C/T Day 1

C/O hr 12

Parts# No. N/A

UT % 85

#M No. 1

#O No. 7

#S No. 2 x 12h

WT min 660

BS set ~ 0.22

FTQ % 99

P2

C/T 5.3

C/O 1

Parts# N/A

UT 80

#M 1

#O 5

#S 2 x 12h

WT 660

BS 100%

FTQ 100

P4

C/T 0.7

C/O 3

Parts# 0.00%

UT 85

#M 1

#O 12

#S 2 x 12h

WT 660

BS 100%

FTQ 100

P3

C/T 2.9

C/O 2

Parts# 0.00%

UT 95

#M 3

#O 16

#S 2 x 12h

WT 660

BS 100%

FTQ 99.5

Raw Mat. 161 1.0 WIP. 6 5.3 +

2.9 FG. 78 0.7

P1 P2+P3 P4

Raw Mat.

141

1.0 WIP. 18 5.3 +

2.9 FG. 63 0.7

P1 P2+P3 P4

Raw Mat.

332

1.0 WIP. 69 5.3 +

2.9 FG. 180

P1 P2+P3 P4

C/O.

0.5

C/O.

0.1

C/O.

0.1

C/O.

0.5

C/O.

0.1

C/O.

0.1

C/O.

0.5

C/O.

0.1

C/O.

0.1

POT LINING FUTURE MAP

D18 (Days)

P1

NVA (days)

VA (days)

P rod. T ime= 9.9 days

Door to Door lead time

=105 days

NVA (days)

VA (days)

P rod. T ime= 9.9 days

Door to Door lead time

=84 days

NVA (days)

VA (days)

P rod. T ime= 9.9 days

Door to Door lead time

=79 days

CD20 (Days)

D20 (Days)

Proc Cont. & Supply

P3

Cell Construction

P4

Cell Change Over

P2 shell

Demolition & Repair

Cathode 20.6 pots

Lining

Supplier Material

Potrooms

P1

Cathode Casting

Cathode 14.8 pots

Cathode 11.7 pots

SiC 38.3 pots (14.8+20.6+11.7)

Col Bars 35.4 pots (14.8 +20.6)

Col Bars 11.7 pots

D1

8

CD2

0

D20

Supplier Development

Min 3

pot sets /design

2 pot

sets /design

OEE Improvement

Inventory Reduction

19.8 POT

SETS

25.6 POT

SETS

16.7 POT

SETS

P1

C/T Day 1

C/O hr 12

Parts# No. N/A

UT % 85

#M No. 1

#O No. 7

#S No. 2 x 12h

WT min 660

BS set ~ 0.22

FTQ % 99

P4

C/T 0.7

C/O 3

Parts# 0.00%

UT 85

#M 1

#O 12

#S 2 x 12h

WT 660

BS 100%

FTQ 100

P3

C/T 2.9

C/O 2

Parts# 0.00%

UT 95

#M 3

#O 16

#S 2 x 12h

WT 660

BS 100%

FTQ 99.5

0.7 Raw Mat. 71 1.0 WIP. 10 5.3 +

2.9 FG. 14

P1 P2+P3 P4

C/O.

0.5

C/O.

0.1

C/O.

0.1

0.7 Raw Mat. 60 1.0 WIP. 6 5.3 +

2.9 FG. 9

P1 P2+P3 P4

C/O.

0.5

C/O.

0.1

C/O.

0.1

0.7 Raw Mat. 48 1.0 WIP. 8 5.3 +

2.9 FG. 12

P1 P2+P3 P4

C/O.

0.5

C/O.

0.1

C/O.

0.1

P2

C/T 5.3

C/O 1

Parts# N/A

UT 83

#M 1

#O 5

#S 2 x 12h

WT 660

BS 100%

FTQ 100