Improving product recovery decisions through enhanced product ...

Improving product recovery decisions through enhanced product information Ajith Parlikad, Duncan McFarlane, Anand Kulkarni

05th April 2005

© 2005, all rights reserved

1

Overview • Introduction to the problem • Linking product information with product recovery decisions • Enhancing product information with “networked RFID” • Improving product recovery decisions with enhanced product information • Conclusions


2

The Problem

Government Regulations – WEEE Directive – RoHS Directive Other financial benefits…


3

WHAT CAN BE DONE?

• Recycle

Reuse Future

– Not cost-effective – Not environment-friendly

Refurbish Cannibalise

• Refurbish/Reuse

Recycle

– Desired option – Problem: Information availability

Current

Dispose Volume


4

Research Question

How does ready availability of product information affect the effectiveness of product recovery decisions? How can we quantify the impact of information availability on product recovery decisions?


5

Case Studies • 12 companies – 3 computer remanufacturers – 3 photocopier remanufacturers – 1 phone remanufacturer – 3 computer dismantlers – 2 white goods recyclers


6

Product recovery operations


7

Information Requirements • Current specifications (‘Decision enablers’) – Material content – Original specifications – Later modifications

• Residual life (‘Decision verifiers’) – – – – –

Reliability Age Usage Maintenance Current condition


8

Information “Loss” Aero engines, Caterpillar,... Photocopiers, Automobiles,...

High

Toasters, Blenders,... Information associated with product Remanufacture

Resale

Low Production

Retail

Usage

Product lifecycle stage


9

Disposal

Percentage of companies

Product Identification 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Access to Manufacturer

Product History Logs

Inspection & Testing

Recycling/Dismantling


10

"Expert" knowledge

Remanufacturing

No information collected

Information availability Information availability Information Required

Pre-sorting

Grading

Material content

8

depending on access to design data

Original specification

9

9

Later modifications

8

9

Reliability

8

depending on access to design data

Age

8

depending on access to sales & maintenance data

Current condition

8

9

Usage

8

8

Maintenance

8


11

depending on the availability & accuracy of maintenance logs

Linking Product Recovery Decisions to Information Quality

• Uniqueness – to enable individual information trails for each unique object throughout its lifecycle and across the whole supply chain.


12

Linking Product Recovery Decisions to Information Quality Decision – Cost/Revenue Value

Threshold Time

• Timeliness – to ensure that information is readily available for decisionmaking and execution process with minimal need for manual inspection or testing.


13

Linking Product Recovery Decisions to Information Quality Dell P2 500MHz; 128MB RAM; 20GB HDD…

Dell P2 500MHz; 512MB RAM; 100GB HDD…

• Completeness – to ensure that all relevant information is available for optimising decisions.

• Accuracy – to reduce or eliminate inaccurate representations of current and historical product information.


14

A Networked RFID Solution for PLM D584.S421.CC21.AA21 F127.C238.DF1B.C7CC

CD135.EEA2.DCF5.12DD

BA132.FFA2.DEF1.12ED 1E1D.DD13.14CC.AC3B


15

EPCtm Network Technology Building Blocks Application

EPC URL

EPC

ONS

PML

Savant

Local

EPCIS

RF Reader


16

Remote

EPCIS

Information Quality EPC • Uniqueness – Enable unique product “footprint”

––> EPC

• Completeness – Ensure availability of relevant product info

––> EPCIS/ONS/XML

• Timeliness – Ensure “ready” availability of product info

––> RFID/Filtering

• Accuracy – Reduce/eliminate errors in info management ––> RFID/XML


17

Networked RFID for PLM

EPCIS Discovery Service

EPCIS

EPCIS

EPCIS

EPCIS

Date of manufacture Parts/ materials used (Dis)assembly recipe

Manufacture


Retail

Usage/Maintenance

18

End-of-Life



EPCIS Date of manufacture Parts/ materials used (Dis)assembly recipe

Manufacture


EPCIS

EPCIS

EPCIS

Date of sale Warranty details Parts replaced

Retail

Usage/Maintenance

19

End-of-Life




Manufacture


EPCIS

EPCIS


On-board data Usage history Parts installed

Retail

Usage/Maintenance

20

EPCIS

End-of-Life




Manufacture


EPCIS

EPCIS



Retail

Usage/Maintenance

21

EPCIS Parts/materials identified Disassembly history

End-of-Life

Networked RFID for PLM 1

EPC xyz

“Where can I find information about EPC xyz?”


EPC xyz 2

ONS

“Tell me which part was replaced” “100 GB HDD”


Manufacture


EPCIS

EPCIS



Retail

Usage/Maintenance

22

EPCIS Parts/materials identified Disassembly history

End-of-Life

RFID-enabled Product recovery

Information gathered

Functional Condition

EPCIS or Component-level tagging

Usage Rate

Age

EPCIS or Sensor Data

Current Specifications Original Specifications

History Logs

Model

Testing

Brand

Inspection

Product Type Manual


Barcode

23

Book-in RFID

A Bayesian Product Recovery Decision model Decisions

D

Utility of a decision = f (certainty, payoff)

U

States/Hypotheses

Utilities

Product

Components Component 1

Features/ Evidences

Feature 1


Component 2

...

Feature 2

24

...

Component n

Feature m-1

Feature m

EPC enhanced product recovery decisions Quicker decisions

Certainty

Higher Utility Utility (t)

Payoff

Test 1

Information

11

22

Test 2 33

44

Test 3 55

66

77

Test 4 88

EPC enabled information


25

Time

CONCLUSIONS • Process improvements – Quick and possibly automated identification & sorting – Error reduction

• Decision improvements – Better estimation of residual life and value – Better estimation of recyclable material content – Rich information leads to better decisions


26

Thank You [email protected] http://www.ifm.eng.cam.ac.uk/cdac http://www.autoidlabs.org.uk


27