“Wall of wire” or “wall of confusion”?

(1)

WHAT ARE THE ISSUES &

WHAT DO THEY TELL US?

BENCHMARK ECONOMICS AND CASTALIA CASTALIA

1 Ageing infrastructure investment:

“Wall of wire” or “wall of confusion”?

ISCR Seminar, 27 May 2008 Wellington

Margaret Beardow, Principal Benchmark Economics

Eli Grace-Webb, Analyst

Castalia Strategic Advisors

(2)

The “wall of wire”

2

$-

$5,000,000

$10,000,000

$15,000,000

$20,000,000

$25,000,000

$30,000,000

$35,000,000

$40,000,000

19 40

19 50

19 60

19 70

19 80

19 90

20 00

20 10

20 20

20 30

20 40

20 50 INIT

IAL + R EP LA CE ME NT CA PE X ($

19 99

) Initial investment Replacement investment

We are here

(3)

In a nutshell…

“Wall of wire “ debate is confused…

it does not measure age based renewals…

It measures probabilty of asset failures

it is a tool for asset managers targetting reliability centred maintenance

Probability analysis - importance for NZ:

Provides guidance on future magnitude and timing of expenditures

Allows cost-effective assessment of trade-off between opex and capex

Allows quantification of likely future capital requirements to provide comfort for regulators

3

(4)

Ageing assets and probability analysis

4 Asset management as core business

(5)

“The wall of wire”

5 A colourful term…

for…

asset mortality analysis

(6)

$-

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

$350,000

$400,000

$450,000

1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 2030

CAPEX + REPEX

YEAR

Failure probability Initial investment

What is the “wall of wire”?

6 Initial investment still in service With an average age of 50 years assets may fail at age 15 but also last past 80years -

Asset managers need guidance on likely timing of maintenance and replacement costs

Probability of failure of existing assets based on mortality/survivor distribution function

(7)

What does wall of wire measure?

7 It does measure…

9 probability of failure of ageing assets 9 each year and trend over time

9 wave of probable expenditure

9 possible trade-off between opex and capex

It does not measure…

µ

“wall of wire”

µ

“age based replacement”

µ

“cliff-edge replacement”

µ

“change in investment”

µ

only asset replacements

Asset aged 40 years may have a remaining life from zero to 40 years

Managers have no certainty

µ

certainty

Enables least cost life cycle asset management

Not certainty, but guidance

(8)

Probability of failure Or,

Asset mortality analysis

Determined by:

y

Mathematically derived probability distribution function

y

Probability of failure based on empirically derived mortality rates

8

8 BENCHMARK ECONOMICS AND CASTALIA

CASTALIA

(9)

² “THE CAPM OF AGEING ASSET EXPENDITURE ANALYSIS”

² WHAT IS IN AN ASSET MANAGER’S INFORMATION SET?

² LIFTIME PROBABILITY DISTRIBUTION FUNCTIONS

² Weibull

² Gompertz

² Gumbel

9 Probability analysis:

Estimation & distribution functions

(10)

BENCHMARK ECONOMICS AND CASTALIA

Reliability-based ageing asset theory

10

Probability

Asset Age High

Low

Young Old

Infant

mortality Wear out

failures

1. Observed rate of asset failures

Probability

2. Probability of asset surviving

Asset Age High

Low

Young Old

Survivor curve

Assets more like to wear out when older

Assets more like to fail when older

Reduces system

reliability

(11)

Describing the probability of failure

11

Density

3. Distribution of asset failures due to wear out

Asset Age High

Low

Young Old

Skewed distribution Normal

distribution

Distributions

Each distribution is capable of describing the failure characteristics of a number of asset types, and the aggregation

of assets into networks.

Weibull distribution function

Iowa curves H-Curves

Gompertz-Makeham distribution function Gumbel distribution Normal distribution

(12)

12 The Weibull distribution

Wallodi Weibull (1939)

Mathematical formula, similar to the normal distribution

Fitted to observed failure rates

Weibull

Scale (η) – similar to a mean

Shape (β ) – defines the skew

Location (γ ) – shifts the distribution to the right

Parameters

Shape 7, Scale 25, Location 5

0 0.02 0.04 0.06 0.08 0.1 0.12

0 5 10 15 20 25 30 35 40

Age of Asset (years)

Probability Density

(13)

Changing the Weibull parameters

Changing Shape Parameter

Shape 1 Shape 3

Shape 7 Shape 10

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

0 5 10 15 20 25 30 35 40

Probability Density

Changing Scale Parameter

Scale 10

Scale 15 Scale 20

Scale 25

0 0.05 0.1 0.15 0.2 0.25

0 5 10 15 20 25 30 35 40

Probability Density

13

Changing Location Parameter

Location -15 Location -5 Location 5 Location 15

0 0.02 0.04 0.06 0.08 0.1 0.12

0 5 10 15 20 25 30 35 40

Probability Density

(14)

14

Asset Description Scale Shape Location

Our example 25 7 5

33 kV pin insulators, Australia

42.8 3.54 8

Wooden distribution

poles, Tasmania 43 3.6 5

poles, Queensland 96 4.17 0

Concrete distribution

poles, Queensland 114 4 20

poles, Canada 44.07 4.21 0

XLPE 15 kV cables,

United States 61.21 5.69 0

HMWPE 15 kV cables, United States

66.3 8.39 0

Different assets are modelled by different Weibull distributions

Fitted Weibull Function

Observed Data

0 0.2 0.4 0.6 0.8 1 1.2

0 5 10 15 20 25 30 35 40

Age of Asset (years)

Survival rate

Example Assets: Survivor curve:

(15)

15 Who uses reliability-based aging asset theory?

US Army

Insurance Companies

Network Operators Accountants

Drug Companies

Asset Reliability

Risk of Insurance Claims

Depreciation

Risk of Recall

System Reliability

Used by: Used for:

Electricity Industry

Water Industry

Many Other Industries Gas Industry

Drug Industry

Industries:

(16)

WHY DID IT ENTER DEBATE

FAILURE CURVES & IMPLICATIONS FOR FUTURE EXPENDITURES

NZ INVESTMENT AND PROBABILITY OF FAILURE

Probability analysis

& New Zealand

16

(17)

17 “…expenditure on ageing assets is recurring …it will be reflected in the historical opex trend “

“…Commission has decided that past trend in capex will be used as the starting-point…” … ESC, Victoria

WHY DID IT

ENTER DEBATE?

To demonstrate that the past was not a guide to the future

UK – distribution capex

Year Replacement estimate

$2007 – New Zealand 1980 $6M

1990 $19M 2000 $62M 2010 $152M

Past expenditures were setting basis for future allowances

…But - Ageing assets were new phenomenon and not captured in past trend -

Threshold effectively imposed price cap – no allowance for ageing asset base

(18)

Failure curves & implications for future expenditures

18 Network assets –

Average 50 years: 20 Æ 80

0%

25%

50%

75%

100%

0 10 20 30 40 50 60 70 80

AGE (Years) Failures Survivors

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 Age ‐Years

Probability of failure rate

50 averge life

Failure rates - Annually

Asymmetric failure rate – normal distribution not appropriate

(19)

New Zealand – Initial investment &

probability of failure

19

(20)

Asset age profile & failure rates

Failure rate differs with asset age

Older assets have higher failure rates

Maintaining weighted average asset age of 25 years provides balance between too much

expendiutre or too much unreliability

20

20 0%

4%

8%

12%

16%

$-

$125

$250

$375

$500

1917 1927 1937 1947 1957 1967 1977 1987 1997

Haz ard -ra te

% -es tiam ted fail ure rate

Cap ex -$2 00 7M

NZ asset investment-LHS Hazard rate - RHS

Age 25 years

(21)

Asset age and failure rates NZ

21

0% 10% 20% 30% 40% 50% 60%

Over 25 years Over 30 years Over 40 years Over 50 years

% of assets

0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 3.5% 4.0% 4.5% 5.0%

Over 25 years Over 30 years Over 40 years Over 50 years

Annual failure rate as % of age cohort

Age-based cohorts

Reliability based cohorts

Assets aged 40 years with average life expectancy of 50 years are not 5 to 10 years from renewal…

…They could fail

tomorrow or last another 40 years

Hence probability distribution functions

(22)

Opex/assets ratio rises as assets age

Opex/assets ratio rises

as assets age …and NZ has 54% of assets

> 25 years

…and NZ has 54% of assets

> 25 years

Asset ageing and maintenance expenditure

22 Opex/assets 0% – 2% first 25 years 2% to 13% 2

^nd

25 years

0%

2%

4%

6%

8%

10%

12%

14%

1 6 11 16 21 26 31 36 41 46

O p e x / a s s e t s

%

Asset age ‐years

0%

4%

8%

12%

16%

$-

$100

$200

$300

$400

1940 1950 1960 1970 1980 1990 2000

Op ex / as set ra tio

Init ial i nv es tm en t $2 007 re pla cem en t co st

NZ asset investment-LH S

Opex/asset ratio (RC) - RHS

Older assets require a higher level of opex than

younger assets

(23)

ASSET MANAGEMENT AS CORE BUSINESS ROLE OF PROBABILITY ANALYSIS IN

ASSET MANAGEMENT

FAILURE ANLAYSIS AND EXPENDITURE PLANING 23

Managing asset based businesses

(24)

Role of probability analysis in asset

management

Objective:

to maintain system reliability

at least life cycle cost

by cost-effective trade-off between maintenance and replacement

Criteria::

cost of outages or failure

life cycle costs of the asset

criticality of the asset to the network

24

Objective Maintain system reliability

Reliability-based ageing asset theory

Decision over assets

Determine optimal system reliability

Cost of outages or failures

Life cycle costs of assets

Criticality of assets

Maintenance

Operating Expenditure

Replacement Refurbishment

Capital Expenditure

Trade off

(25)

Implications Implications Asset investment / revenue

Asset investment / revenue

25 Networks are different :

Asset Management as core business

$0 $2 $4 $6 $8

QBE Lend Lease Woolworths David Jones Wesfarmers Orica One Steel Orgin Blue Scope AGL BHP Billiton Telstra Qantas Contact Energy Macquarie Gen Aust Pipeline Trust EnergyAustralia

Powerlink y

Asset management is THE business:

y

Objective for asset manager is to maintain system reliability - hence reliability centred maintenance

y

Criteria:

{ Reliability demanded by consumers

{ Life time cost of assets

{ Criticality of asset

{

Trade-off between opex & capex

(26)

Lifetime mortality distribution can be estimated many years out Lifetime mortality distribution can be estimated many years out

Failure analysis for expenditure planning

26

- 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141

YEAR

POLE REPLACEMENTS

1st Replacement 2nd Replacement

Base Workload - Calculated

Asset Inspection - 4.5 Years Cycle 11 9 17 12 10 10 9 9

Pole Replacement - Actual Defect Rate 10 10 23 11 13 9 7 14

Pole Reinstatement - Actual Defect Rate 0 0 1 0 0 0 0 1

Other Defects - Actual Defect Rate 16 25 21 50 31 10 11 26

Total 38 44 61 73 54 29 28 51

Trend

Asset Inspection 2 2 1 -1 -1 0 0 -1

Pole Replacement 3 0 9 2 7 -4 -2 -1

Pole Reinstatement 0 0 0 0 0 0 0 0

Other Defects -2 4 -10 -2 6 -12 0 0

Total 3 6 -1 -1 12 -16 -2 -2

Backlog

Asset Inspection 0 0 1 0 0 0 0 1

Pole Replacement 3 2 3 1 1 1 1 1

Pole Reinstatement 0 0 0 0 0 0 0 0

Other Defects 6 8 7 5 1 2 1 0

Total 9 10 11 6 3 3 2 2

Additional Resource Requirement 12 16 11 5 15 -13 0 1

Regions

“Wall of wire” or “wall of confusion”?

WHAT ARE THE ISSUES &

WHAT DO THEY TELL US?

1

Ageing infrastructure investment:

“Wall of wire” or “wall of confusion”?

ISCR Seminar, 27 May 2008 Wellington

Margaret Beardow, Principal Benchmark Economics

Eli Grace-Webb, Analyst

Castalia Strategic Advisors

The “wall of wire”

2

We are here

In a nutshell…

 “Wall of wire “ debate is confused…

 it does not measure age based renewals…

 It measures probabilty of asset failures

 it is a tool for asset managers targetting reliability centred maintenance

 Probability analysis - importance for NZ:

Provides guidance on future magnitude and timing of expenditures

Allows cost-effective assessment of trade-off between opex and capex

Allows quantification of likely future capital requirements to provide comfort for regulators

3

Ageing assets and probability analysis

4

Asset management as core business

“The wall of wire”

5

A colourful term…

for…

asset mortality analysis

What is the “wall of wire”?

6

Initial investment still in service With an average age of 50 years assets may fail at age 15 but also last past 80years -

Asset managers need guidance on likely timing of maintenance and replacement costs

Probability of failure of existing assets based on mortality/survivor distribution function

What does wall of wire measure?

7

It does measure…

9 probability of failure of ageing assets 9 each year and trend over time

9 wave of probable expenditure

9 possible trade-off between opex and capex

It does not measure…

“wall of wire”

“age based replacement”

“cliff-edge replacement”

“change in investment”

only asset replacements

Asset aged 40 years may have a remaining life from zero to 40 years

Managers have no certainty

certainty

Enables least cost life cycle asset management

Not certainty, but guidance

Probability of failure Or,

Asset mortality analysis

Determined by:

Mathematically derived probability distribution function

Probability of failure based on empirically derived mortality rates

8

9

Probability analysis:

Estimation & distribution functions

Reliability-based ageing asset theory

10

1. Observed rate of asset failures

2. Probability of asset surviving

Assets more like to wear out when older

Assets more like to fail when older

Reduces system

reliability

Describing the probability of failure

11

3. Distribution of asset failures due to wear out

Distributions

Each distribution is capable of describing the failure characteristics of a number of asset types, and the aggregation

of assets into networks.

12

The Weibull distribution

Weibull

Parameters

“Wall of wire “ debate is confused…

it does not measure age based renewals…

It measures probabilty of asset failures

it is a tool for asset managers targetting reliability centred maintenance

Probability analysis - importance for NZ: