Data Mining Applications in Manufacturing

(1)

Data Mining Applications in Manufacturing

Dr Jenny Harding Dr Jenny Harding

Senior Lecturer

Wolfson School of Mechanical &

Manufacturing Engineering, Loughborough University

(2)

New P

roduct New Designs

Additional Facilities?

Redesigned Processes?

Extra Resources?

Existing Products

Existing Designs

New Business Objectives

New Technology?

Identification of Knowledge - Context

Intelligent Support

Systems Problem Areas

Tools:

Information Modelling Process Modelling

Artificial Intelligence Simulation

Data Mining???

How do we find and use the

“Right” Knowledge?

(3)

Sources of Manufacturing Knowledge

People

Expertise and Experience

Facts or Folklore? Under Exploited?

Processes

(Operational Data)

Production Processes

(4)

Background History

Why Data Mining?

What is the “Right” Knowledge?

Or is the knowledge hidden somewhere else as well?

Are “experts” the only sources?

Is there any substance in folklore?

(5)

Reported Applications of DM in Manufacturing

Earliest - Fault diagnosis (1987) Substantial increase in

(manufacturing) data mining publications in last 2 years

1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 ^M_a

teria l Prop

erties Sh

op Flo

or C ontro

l La

you t Des

ign M

ain ten

ance Dec

isio n S

up port En

gin eerin

g D esign Ma

nu factu

rin g Sy

stem s 2

1 4

1 1 4

2 8

23 3 7

6 1

1

3 2

3

1 2 2 7

1

12 1 1

3 1 2 2

4

1 3

2 1 1 3

1 3

1 1 5

2

1 1 11 2 1 1

1 1

1

1 1

1 112 2

1 1¹ 2 0

1 2 3 4 5 6 7 8

No. of Papers

1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 ^M_a

teria l Prop

erties Sh

op Flo

or C ontro

l La

you t Des

ign M

ain ten

ance Dec

isio n S

up port En

gin eerin

g D esign Ma

nu factu

rin g Sy

stem s 2

1 4

1 1 4

2 8

23 3 7

6 1

1

3 2

3

1 2 2 7

1

12 1 1

3 1 2 2

4

1 3

2 1 1 3

1 3

1 1 5

2

1 1 11 2 1 1

1 1

1

1 1

1 112 2

1 1¹ 2 0

1 2 3 4 5 6 7 8

No. of Papers

Harding, J A, Shahbaz, M, Srinivas and Kusiak, A, 2006, "Data Mining in Manufacturing: A Review", American Society of Mechanical Engineers

(ASME): Journal of Manufacturing Science and Engineering.

(6)

Reported Applications of DM in Manufacturing (2)

Use of data mining applications in Manufacturing is relatively new

To date – most of the research done has focused on semi-conductor industry (fault detection and

productivity improvement)

Summary:

To date – most of the reported research uses decision trees, clustering and neural networks

(7)

Challenges for DM in Manufacturing

? Investment – very wide range of tools and techniques and not obvious in most applications which will produce best results and rewards (DM preparation costs are high).

Why slow adoption?

? Availability and quality of data – substantial data cleaning and pre-processing required.

? Combinations of expertise needed – Data Mining Expertise + Expertise in Processes / Technologies under consideration

? Inadequate Exploitation of Resulting Knowledge -

“one-off” applications addressing specific problems

(8)

Challenges for DM in Manufacturing (2)

? Uncertainty and risk - there are no guarantees that useful knowledge will be discovered. In highly

competitive situations high risks may be unacceptable

Why slow adoption?

? Complexity and diversity - extremely difficult to devise generic data mining processes even for particular types of manufacturing problems

(9)

Case Studies

Experience from literature and

industrial case studies shows……….

(10)

An Example

Width

Height

aa ab ac ad

Thickness aa ab ac ad ae af

aa ab ac

Width

Height

aa ab ac ad

aa ab ac

Width

Height

aa ab ac ad

aa ab ac

Width

Height

aa ab ac ad

aa ab ac

A simple sample product

(11)

Data Cleaning

• Very important! (But very time-consuming)

•Iterative Process

Process 1

Process 2

Process n

Data Cleaning

Module Data

Data

consolidated Error-free

table Process 1

Process 2

Process n

Data Cleaning

Module Data

Data

consolidated Error-free

table

General Classifications

Identical Records

Some Duplication in Records Confusing Records

Missing Values in Records etc

Possible Causes

Breakdowns / Restarts Rework cycles

Operator Error

Operator Training Etc., etc.

(12)

Example Thoughts during Cleaning

Should Reworked records be included or ignored?

Can Missing values be replaced or should the whole record be rejected?

Can domain experts of more detailed files be consulted about replacing the abnormal data?

All the replaced values need to be documented for future consultation.

(13)

Data Transformation

01 02 03 04 05 06 07 08 09 10 11 σ

σ σ σ σ

σσ σ σ

σ σ σ

01 02 03 04 05 06 07 08 09 10 11 σ

σ σ σ σ

σσ σ σ

σ σ σ

Lower Limit LoutLower H-Lower M-Lower S-Lower Nominal S-Upper M-Upper H-Upper Upper

Upper Limit Uout

Products - split

tolerance range for dimension.

Equal divisions???

Process variables

Normal distributions?

Equal divisions???

(14)

Data Mining

Association Rules the Apriori Algorithm has been most successful

Regression analysis used first – are there any simple relationships that can be exploited?

Several data mining approaches tried

A complete set of dimensional / manufacturing data for each product needs to be collected and transformed to make each record

(15)

Data Mining

In Manufacturing - Data mining decisions should be made in the context of the physical realities of the data that is being examined.

In very highly controlled manufacturing

processes, unusual combinations of results (representing problems) may be rare.

Support level kept very low to reduce risk of

missing unusual but important combinations of manufacturing outputs

(16)

Association Rule

Used in mostly Market Basket Analysis Apriori Property:

“Any subset of a large itemset must be large”

Rules are in the form of

A ŁŁŁŁ B or A ŁŁŁŁ (B & C) etc.

e.g. Chips ŁŁŁ Coke Ł

•Each rule is based on certain Support and Confidence level.

(17)

Rule Quality

Need to check Quality of the generated rules – use Support, Confidence, Statistical methods AND check with domain experts that the rules make sense in the Manufacturing Context.

Support: Support is the frequency of the occurrence of items in a transaction or record.

Confidence: Confidence is the certainty of the discovered rule.

(18)

Example Rules

Types of Manufacturing Rules that may be identified include:

Product Design Errors or Constraints Process Errors or Constraints

From Product Data - Rules of the form

Good Dimension output ŁŁŁŁ Good Dimension output Bad Dimension output ŁŁŁŁ Bad Dimension output Good Dimension output ŁŁŁŁ Bad Dimension output

From Product and Process Data

Process variable values that produce Good Product Results Process variable values that produce Poor Product Results

(19)

Example Rules

•Rules that indicate one dimension being “nominal”

results in another dimension being “nominal” are reassuring – but generally not very interesting

•Rules that indicate one dimension being “nominal”

results in another dimension being out of tolerance or away from “nominal” are more interesting.

Possible Causes include:- Design Error

Process Error

Process Constraints

(20)

Results

• Data Mining can be used to extract Manufacturing process limitations.

• Data Mining can be used to discover design constraints and then help to improve the product design, (if manufacturing process is acceptable).

• Data Mining can be used to improve the output quality by controlling the manufacturing process variables.

(21)

Data Mining projects tend to be “one-off”

applications to solve a particular problem

Future Considerations……….

But how do we make sure the acquired

knowledge is made use of when it has been discovered?

Greater benefits would be gained if they were

part of an ongoing Knowledge Reuse

programme

(22)

Data Mining Network

Main Pool of Knowledge Integration &

DM Engine

Local Site / Process 1

Local Databases and Files

Main Data Warehouse

Local Data Mining Engines

Local Pool of Knowledge

Local Site / Process 2

Local Pool of Knowledge Local Site / Process n-1

Local Site / Process n

(23)

Any Questions

Thank You Thank You Thank You Thank You

[email protected] [email protected]