Upload
jacknickelson
View
1.178
Download
0
Tags:
Embed Size (px)
Citation preview
1
USING ANALYTICAL TOOLS TO IMPROVE ASSET MANAGEMENT FOR T&D
May 5-7, 2004Hyatt Regency BostonBoston Massachusetts
Sponsored by EUCI
S. Chapel AssociatesLee Merkhofer Consulting
VMN Group LLC
2 2
Outline
• Introduction
• T&D Project Prioritization
• Managing Aging T&D Assets
• T&D Strategic Reliability
• Summary
OverviewPurpose of WorkshopUsing Models
ProblemDecisionsAnalytic ToolsDataOrganizational Issues
TheoryData & ModelsOrganizational Issues
1
Using Analytical Tools to Improve Asset Management for T&D
INTRODUCTION
Steve Chapel, S. Chapel Associates
Lee Merkhofer,Lee Merkhofer Consulting
Charles D. Feinstein and Peter A. MorrisVMN Group LLC
May 2004
2 2
Agenda
• Overview– Some Asset Management Beliefs: fact or fallacy?– History of our Asset Management Research – T&D Asset Management Issues
• Purpose of Seminar• Using Models
3 3
Some Asset Management Ideas – True or False?
• Problems can be solved by organizational change and asset management teams
• "Cost Benefit analysis, the "tried and true" method for ranking projects is a perfectly reasonable way to select projects to fund."
• The first important step is to gather data• All projects can be valued using the same aggregate
measures (i.e. $)• Projects are risky because of uncertain financial
consequences • Beta is an appropriate way to measure project risk
4 4
Some Asset Management Ideas – True or False?
• An important objective is near-term profitability • The biggest problem for asset management is
insufficient data. • ROI is a good metric for evaluating investments. • Ranking is a good way to prioritize projects. • Strategic alignment is a good metric for prioritizing
projects. • Balanced scorecards are a good way to prioritize
projects. • Hurdle rates are a good way to account for risks.
5 5
Agenda
• Overview– Some Asset Management Beliefs: fact or fallacy?– History of our Asset Management Research – T&D Asset Management Issues
• Purpose of Seminar• Using Models
6 6
T&D Research 1997 – 2002
• December 1996 Chicago Focus Group• 1997 - 1998
– Create the portfolio – distribution planning – Produce V1.0 Area Investment Planning Tools
• 1999– Start work on Customer Needs & Reliability projects– V1.5 Strategy Model & V2.0 LoadDynamics – Methodology design for Project Prioritization (AEP)
7 7
T&D Research 1997 – 2002 cont.
• 2000– Aging Assets - Started– Project Prioritization – method & software designed – Customer Needs & Reliability – EPRI white papers– Area Investment Planning – tech transfer
• 2001: Focus on Aging Assets and Project Prioritization• 2002:
– Continued focus on Aging Assets and Project Prioritization – Some work on Measuring & Valuing Reliability
• 2003:– Aging Assets – refine existing software & explore better ways to
analytically address the problem– Project Prioritization: Refine existing software & continue improve
methods for measuring T&D value
8 8
Utility Infrastructure Decision Making – The Focus of Our Work
Underlying Objective: Minimize the lifecycle costs of distribution infrastructure while meeting customer needs for reliability & power quality
Area Infrastructure Planning
Area Infrastructure Planning
Managing Aging Assets
Managing Aging Assets
Project Prioritization / Capital Budgeting
Project Prioritization / Capital Budgeting
Managing existing assetsManaging existing assets
Expand distribution infrastructure to meet future load
Expand distribution infrastructure to meet future load
Measuring & Valuing Reliability
Measuring & Valuing Reliability
Assessing Customer Needs
Assessing Customer Needs
Investment EconomicsInvestment Economics
Supporting Research
Decision Tools & Information
Utility Infrastructure Decisions
9 9
Our Focus
• Analytical Tools & Information– Managing Distribution Aging
Assets– Project Prioritization– Measuring & Valuing
Reliability – Assessing Customer Needs– Infrastructure Investment
Strategy
• Supporting research:– Measuring & Valuing
Reliability– Assessing Customer Needs– Investment Economics
10 10
Agenda
• Overview– Some Asset Management Beliefs: fact or fallacy?– History of our Asset Management Research – T&D Asset Management Issues
• Purpose of Seminar• Using Models
11 11
T&D Asset Management is Complex
• What about reliability - how far can we push the system before falling off the edge?
• How do we get the right-sized capacity when it is needed?
• Can we save $ by deferring maintenance?• What risks are we taking by practicing “doing-less”
decision making (do we really know how close we are to the edge)?
12 12
T&D Asset Management is Mostly About Solving Investment Problems
• 40% to 50% of electric utility net investment
• Business issues– Minimizing investment costs – Having “right” infrastructure to
meet customer needs– Making money
• Key strategic needs– Managing assets– Linking investment and O&M
decisions to customer needs
Net Invest.Gen. = $8.7B
Tran. = $4.5B
Dist. = $13.5B
Total = $26.7B
Generation Generation
A
B
230 KV
Backbone Transmission Grid
C
230/116
115/12 115/12
230/69
69/469/12
Distribution PlanningArea 1
Distribution PlanningArea 2
GENERATIONPLANNING
TRANSMISSIONPLANNING
DISTRIBUTIONPLANNING
13 13
T&D Asset Management
• Two driving facts– Very low revenue to asset
ratio– Large embedded asset base
• Substantial care and feeding is required– Repair / Replace– Expand / Prepare for future– Reliability Net Invest.
Gen. = $8.7B
Tran. = $4.5B
Dist. = $13.5B
Total = $26.7B
Generation Generation
A
B
230 KV
Backbone Transmission Grid
C
230/116
115/12 115/12
230/69
69/469/12
Distribution PlanningArea 1
Distribution PlanningArea 2
GENERATIONPLANNING
TRANSMISSIONPLANNING
DISTRIBUTIONPLANNING
14 14
Opposing Trends are Changing the Business
• Increasing levels of key drivers– Restructuring & deregulation– Regulatory & Corporate
scrutiny – Customers demand for
reliability and service quality– Technology change
• Reduced budgets & increased corporate control
• Together these are creating a collection of problems
•Competition•Customer Needs•Technology
•Budgets•Local Control
Time
Leve
l
15 15
Nature of Business Planning Problem
Local AreaDecisions
Local AreaDecisions
TechnologyDemand for
Services
Regulators
ExistingAssets
CorporatePolicy
CorporatePolicy
Req’ts
Req’ts
Alternatives,Consequences
Budgets,Policy
16 16
T&D Asset Management – Project Planning & Financial Planning
• Fundamental engineering economic problems– Maintenance / Repair /
Replace– Capacity expansion– System Risk Assessment &
Mitigation
• Financial planning problems– 3 to 5 year Capital
Budgeting / Project Planning– Long-Term Financial
Planning
17 17
In Summary T&D presents special challenges
• Huge investment in assets• Regulations / Oversight• “Right” infrastructure for
customer needs• Competitive pressures - extract
maximum value from every asset• Performance vital to other
industries & economy– Recent NE power blackout estimated
to cost $7-10B– Loss of electric power to city of New
York costs about $36m/hr
18 18
Agenda
• Overview– Some Asset Management Beliefs: fact or fallacy?– T&D Asset Management Issues– History of our Research
• Purpose of Seminar– What do you need to know?– What do you hope to get out of the seminar?– What are the main asset management problems?– What have we missed?
• Using Models
19 19
Agenda
• Overview– Some Asset Management Beliefs: fact or fallacy?– History of our Asset Management Research – T&D Asset Management Issues
• Purpose of Seminar• Using Models
T&D Analytic Tools
Using Analytical Tools to Improve Asset Management for T&D
May 5-7, 2004Hyatt RegencyBoston, Mass
Workshop on
Using ModelsLee MerkhoferLee Merkhofer Consulting22706 Medina CourtCupertino, CA 95014
408 446 4105www.prioritysystem.com
3c 2003, 2004 Lee Merkhofer Consulting
Fundamentally, good asset management is about making good decisions
“Asset management is a business process that allows a utility to make the right decisions on the acquisition, maintenance, operation, rehabilitation, and disposal of assets.”
(Stephen Kellogg, Vice President, CDM)
4c 2003, 2004 Lee Merkhofer Consulting
Fundamentally, good asset management is about making good decisions
• Asset management decisions are critical because:– The assets that are owned and how those assets are used strongly
affects the revenues that are generated.– The costs of acquiring, installing, maintaining and retiring assets are
typically large and directly affects the bottom line.– The capacity, reliability, and quality of the company’s assets strongly
affect customer satisfaction along with the perceived and actual levels of service.
“Asset management is a business process that allows a utility to make the right decisions on the acquisition, maintenance, operation, rehabilitation, and disposal of assets.”
(Stephen Kellogg, Vice President, CDM)
5c 2003, 2004 Lee Merkhofer Consulting
Numerous analytic tools are being promoted for asset management
Do such tools really help? What characteristics does a good tool need to have?What type of tool is best?
• Simulation models• Scoring tools• Executive dashboards• Narrow-focused applications• Enterprise-wide applications• Generic software• Custom software
Examples:
6c 2003, 2004 Lee Merkhofer Consulting
The quality of a decision tool depends on the quality of its decision model
“All models are wrong, but some are useful.”(George E. P. Box, University of Wisconsin)
AlternativesDecision
Model
Value
• Tools rely on analytic models.
• The model serves as a means for testing and evaluating alternatives.
7c 2003, 2004 Lee Merkhofer Consulting
More needs than resourcesHigh stakesChoices produce impacts over many yearsRisk and uncertaintyCompeting objectivesMultiple stakeholdersInterdependencies
Models are cost-effective for complicated decisions
Power Delivery Asset Management Decisions
Characteristics of Complex Decisions
YesYesYesYesYesYesYes
Etc.
8c 2003, 2004 Lee Merkhofer Consulting
Models are useful because they address key limitations of human problem solving
Empirical research shows people:- Have limited information processing skills- Can be biased- Can be inconsistent at making choices
…computers to do the things people have trouble doing.
…people to do the things that they are good at.
But people are good at:- Recognizing structure- Making “small,” well-defined judgments- Being creative
Models allow…
9c 2003, 2004 Lee Merkhofer Consulting
If well-designed, a decision model can provide 3 types of benefits1. Use limited resources more effectively by
– identifying value-maximizing choices– aligning decisions with corporate objectives– promoting better spending alternatives by clearly communicating performance
metrics– clarifying sources of value
10c 2003, 2004 Lee Merkhofer Consulting
If well-designed, a decision model can provide 3 types of benefits1. Use limited resources more effectively by
– identifying value-maximizing choices– aligning decisions with corporate objectives– promoting better spending alternatives by clearly communicating performance
metrics– clarifying sources of value
2. Improve the decision-making process by – providing a framework for collecting and incorporating relevant information into
decision-making process– documenting performance goals and improving accountability– reducing opportunities for “gaming”– providing mechanisms for involving stakeholders in decision making– promoting consistent logic for valuing spending options and “leveling the playing
field”– providing discipline for making tough decisions– controlling the role of politics
11c 2003, 2004 Lee Merkhofer Consulting
If well-designed, a decision model can provide 3 types of benefits1. Use limited resources more effectively by
– identifying value-maximizing choices– aligning decisions with corporate objectives– promoting better spending alternatives by clearly communicating performance
metrics– clarifying sources of value
2. Improve the decision-making process by – providing a framework for collecting and incorporating relevant information into
decision-making process– documenting performance goals and improving accountability– reducing opportunities for “gaming”– providing mechanisms for involving stakeholders in decision making– promoting consistent logic for valuing spending options and “leveling the playing
field”– providing discipline for making tough decisions– controlling the role of politics
3. Improve decision-making defensibility by– documenting underlying assumptions and decision logic– promoting consensus over spending decisions– answering “what if” questions
12c 2003, 2004 Lee Merkhofer Consulting
The dilemma: Complex decisions require sophisticated models
“There is a simple solution to every complex problem. Unfortunately, it is wrong.”
(H. L. Mencken)
An inadequate tool can mislead decision makers, resulting in poorer decisions than those that would be made without it.
How can users be assured that analytic tools are adequate decision aids? The workshop addresses this question.
T&D Project Prioritization
Lee MerkhoferLee Merkhofer Consulting22706 Medina CourtCupertino, CA 95014
408 446 4105www.prioritysystem.com
2c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
3c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
4c 2003, 2004 Lee Merkhofer Consulting
Prioritization is an important concept for asset management
• Because resources are limited, not all asset investments can be made now– Prioritization is needed to determine which projects to do now, and
which should be postponed, scaled back, or eliminated.
Assets• Infrastructure• Property• Buildings• Equipment• Vehicles• People• Intellectual assets
Asset Investment “Projects”• Rehabilitate/repair• Replace/improve• Dispose• Create/acquire
5c 2003, 2004 Lee Merkhofer Consulting
Project prioritization guides asset management
• The goal of project prioritization is to select the set of projects that create maximum value.
6c 2003, 2004 Lee Merkhofer Consulting
Project prioritization guides asset management
• The goal of project prioritization is to select the set of projects that create maximum value.
• A “project” is any investment of organizational resources.
7c 2003, 2004 Lee Merkhofer Consulting
Project prioritization guides asset management
• The goal of project prioritization is to select the set of projects that create maximum value.
• A “project” is any investment of organizational resources.• The goal of asset management is to manage assets to
create maximum value.
8c 2003, 2004 Lee Merkhofer Consulting
Project prioritization guides asset management
• The goal of project prioritization is to select the set of projects that create maximum value.
• A “project” is any investment of organizational resources.• The goal of asset management is to manage assets to
create maximum value.• Thus, a priority system can be used to guide investments
for asset management.
PRIORITY LISTPROJECT CUM. COSTS
1. PROJ. W $1M2. PROJ. X $5M3. PROJ. K $8M4. PROJ. G $12M
27. PROJ. H $37M------------------------------------------28. PROJ. M $43M
9c 2003, 2004 Lee Merkhofer Consulting
Project prioritization fits the modern view of asset management
• Technical– Asset management is assuring high reliability and quality service
through adequately building and maintaining poles, wires, transformers, turbines, etc.
Old philosophy
10c 2003, 2004 Lee Merkhofer Consulting
Project prioritization fits the modern view of asset management
• Technical– Asset management is building and maintaining poles, wires,
transformers, turbines, etc., to ensure high reliability and quality of service.
• Economic– Asset management is efficient capital rationing (using hurdle rates,
NPV, ROA, etc.).
Old philosophy
Recent philosophy
11c 2003, 2004 Lee Merkhofer Consulting
Project prioritization fits the modern view of asset management
• Technical– Asset management is building and maintaining poles, wires,
transformers, turbines, etc., to ensure high reliability and quality of service.
• Economic– Asset management is capital rationing through financial evaluation
of options (hurdle rates, NPV, ROA, etc.).
• Strategic– Asset management is about obtaining maximum value from assets,
which requires integrating policy (strategic-level) information, systems-network understanding, and operations-management knowledge.
Old philosophy
Recent philosophy
New philosophy
12c 2003, 2004 Lee Merkhofer Consulting
The “paradigm” of modern asset management• Organizations exist because they create value
– For owners (who provide investment capital).– For customers (who purchase products and
services at a price they are willing to pay)
13c 2003, 2004 Lee Merkhofer Consulting
The “paradigm” of modern asset management• Organizations exist because they create value
– For owners (who provide investment capital).– For customers (who purchase products and
services at a price they are willing to pay)• The capacity to produce outputs of value to the
customer and, thereby, to create value for owners is directly related to the performance of the company’s production assets.
14c 2003, 2004 Lee Merkhofer Consulting
The “paradigm” of modern asset management• Organizations exist because they create value
– For owners (who provide investment capital).– For customers (who purchase products and
services at a price they are willing to pay)• The capacity to produce outputs of value to the
customer and, thereby, to create value for owners is directly related to the performance of the company’s production assets.
• Investments in the asset base (capital acquisition, operation, maintenance and renewal “projects”) should, therefore, be chosen with the goal of maximizing value.
15c 2003, 2004 Lee Merkhofer Consulting
Similarities between asset management and project prioritization
• “… should not just create value, it should create the greatest possible value.”
• “…requires seeing the value of managing assets from a strategic perspective.”
• “…is largely about information management.”
• “…involves seeking common solutions and practices across different lines of business and classes of asset types.
..seeks “value-maximizing”project portfolios.
…evaluates projects based on their impact on the fundamental objectives of the organization.
..uses rigorous techniques to obtain and analyze information.
..seeks accurate, logically defensible, and consistent solutions
Value-maximizing goal
Strategic perspective
Importance of information management
Search for consistent approaches/solutions
Asset Management Project Prioritization
16c 2003, 2004 Lee Merkhofer Consulting
However, formal project prioritization is a big step for most T&D companies
In most cases, companies:• Do not quantitatively evaluate and compare all distribution
projects.• The value of doing a particular project is not compared
with the values of competing projects. • Ignore uncertainty, project interdependencies, and the
strategic “option value” of investments.• Inadequately address process issues--including
information quality, credibility, and trust.
17c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
18c 2003, 2004 Lee Merkhofer Consulting
A priority system is a decision support tool
Rank Project7 ~~~~~8 ~~~~~9 ~~~~~
10 ~~~~~
Decision Support
scoresform
scoresform
TECHNICAL DATA
Software
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Data collection processes
Evaluationprocesses
Also known as tools for– Resource allocation– Multi-project management– Project portfolio management– Capital allocation– Enterprise asset management
Used to – Collect and consolidate project
data– Evaluate and compare project
alternatives– Provide decision
recommendations– Explain and defend choices– Create a level playing field for the
competition for resources
19c 2003, 2004 Lee Merkhofer Consulting
What is your biggest challenge?
Prioritizing which projects to do
Coordinating efforts across the company
Moving at the speed of the market
Finding qualified people
Insufficient resources
Convincing executives of importance
Identifying strategy
40%
36%
30%
22%
20%
20%
16%
Priority systems are popular because many managers view prioritization as a top challenge
(PricewaterhouseCoopers survey of managers, 10/00)
20c 2003, 2004 Lee Merkhofer Consulting
Project prioritization is increasingly being recognized as a key component of business success
“Of all the things a manager can do, figuring out the best way toallocate capital is probably the most important.”
(Fortune Magazine).
Choosing projects is of critical importance…
21c 2003, 2004 Lee Merkhofer Consulting
Project prioritization is increasingly being recognized as a key component of business success
“Of all the things a manager can do, figuring out the best way toallocate capital is probably the most important.”
(Fortune Magazine).
Choosing projects is of critical importance…
…as is justifying the resources required.
“Regardless of industry, people are asking for more rigorous cases that illustrate both the short-term impact and the long-term benefits of projects.”
(Randy Hancock, Senior VP of strategy at Mainspring Inc.)
22c 2003, 2004 Lee Merkhofer Consulting
Project prioritization is increasingly being recognized as a key component of business success
“Of all the things a manager can do, figuring out the best way toallocate capital is probably the most important.”
(Fortune Magazine).
Choosing projects is of critical importance…
…as is justifying the resources required.
Thus, project prioritizationis key.
“Overwhelmingly, enterprises that regularly prioritize projects and services by their business impact are increasing shareholder value.”
(Doug Lynn, MetaGroup)
“Regardless of industry, people are asking for more rigorous cases that illustrate both the short-term impact and the long-term benefits of projects.”
(Randy Hancock, Senior VP of strategy at Mainspring Inc.)
23c 2003, 2004 Lee Merkhofer Consulting
Priority systems address important questions for asset management
How should we spend our resources?
• What total budget should be sought from funding sources?• How should funds be allocated across organizational units?• What level of funding should be provided for each project?
24c 2003, 2004 Lee Merkhofer Consulting
Where should we be headed?
• Is the current set of projects the best way to accomplish our mission?
• To improve alignment with strategy, which projects should be accelerated and expanded, and which should be slowed, scaled back, or eliminated?
Priority systems address important questions for asset management (continued)
25c 2003, 2004 Lee Merkhofer Consulting
Priority systems address important questions for asset management (continued)
What is the right project mix?
• Balance between– Low-payoff “sure things” versus high-payoff “gambles”?– Near-term versus long-term payoffs?– Maintenance versus growth?– Risk versus return?
26c 2003, 2004 Lee Merkhofer Consulting
Continualimprovement
Identify needs
Eval
uate
pro
ject
s
& pr
ojec
t por
tfolio
s
Mon
itor p
rogr
ess
& ef
fect
iven
ess
Implementselected projects
Defineprojects
Priority systems are intended to support the on-going process of project portfolio management
27c 2003, 2004 Lee Merkhofer Consulting
Numerous needs,limited resources
Priority systems address key challenges
Competing objectives
Marketshare
Customer service
ReputationRegulatorycompliance
Financialperformance
Dispersed information/Multiple stakeholders/
Differing opinions
?
Risk/Uncertainty
D ecisions
Critical review
Biases/Gaming
Health, safety &environment
28c 2003, 2004 Lee Merkhofer Consulting
Project mix inconsistentwith strategy
≠Poor-performing portfolios
Projects that won’t die
Resources spread over too many projects
Over-emphasis on “low-hanging fruit”
…and overcome common problems of project portfolio management
R I S K
Insufficientunderstanding
of risk
“Fair-share” Allocations
29c 2003, 2004 Lee Merkhofer Consulting
The computational goal of a priority system is to identify the “efficient frontier” of project portfolios
• Concept: Some project portfolio (e.g., Portfolio D) are undeniably better than others.
30c 2003, 2004 Lee Merkhofer Consulting
The efficient frontier identifies the project portfolios that create the most value for the least cost
31c 2003, 2004 Lee Merkhofer Consulting
There are 4 basic types of priority systems
1. Project ranking systems2. Budget-allocation systems3. Combined (tiered) systems4. Problem/opportunity ranking systems
32c 2003, 2004 Lee Merkhofer Consulting
• Projects ranked according to benefit/cost• Typically used with an all-or-nothing, fund-from-the-top-down decision rule
• Produce a funding curve (efficient frontier) showing which activities to cut (add) and specific benefits lost (gained) if budget is decreased (increased).
Project ranking system
Fun
ded
PROJECT CUM. COSTS1. PROJ. W $1M2. PROJ. X $5M3. PROJ. K $8M4. PROJ. G $12M
27. PROJ. H $37M------------------------------------------28. PROJ. M $43M
Project costs and benefits evaluated
Projects ranked based on B/C
scoresform
scoresform
TECHNICAL DATA
Tota
l Ben
efit
Added costAdded value
Total value
Total cost
Organizational element funding
Ranking produces total benefit versus total cost curve (efficient frontier)
33c 2003, 2004 Lee Merkhofer Consulting
Advantages and disadvantages of project ranking systems
• Ranking activities is a relatively simple, easy to understand concept
• Enables the organization and project proponents to defend individual project choices
• Most appropriate for organization with centralized decision making
• Ranking logic can be extended to handle project interdependencies
• All-or-nothing funding is unrealistic—may need to define and evaluate alternative “versions” of activities (e.g., base-case and reduced scope).
34c 2003, 2004 Lee Merkhofer Consulting
Budget Allocation
Funds Unit A Unit B Unit C Unit D$15M 2 3 5 5$16M 3 3 5 5$17M 3 4 5 5$18M 3 4 6 5$19M 4 4 6 5$20M 4 4 6 6$21M 4 5 6 6$22M 5 5 6 6$23M 5 5 6 7$24M 5 5 6 8$25M 5 5 7 8$26M 5 6 7 8$27M 5 7 7 8
::
Budget allocation system• Provides an optimal allocation of a fixed budget among competing organizational
units • Each organizational unit proposes activities to be conducted under several
different, pre-specified funding cases (e.g., base-case funding, reduced funding, enhanced funding, etc.).
• Each funding case is evaluated (based on the activities to be funded).
Unit A (E-bus) Funding CasesFunding Case
Activity Min Reduced Base MaxB2B site $1M $2M $2M $2MWeb store $1M $1M $2M $2MKid’s lifestyle portal 0 0 $1M $1MRecruiting site 0 0 0 $1M
Total $2M $3M $4M $5M
Each unit prioritizes activities, indicates funding under alternative unit funding
cases, and scores funding cases
Scores used to determine optimal allocation of total funds
• “Bang for the buck” used to allocate funds across sites
35c 2003, 2004 Lee Merkhofer Consulting
Advantages and disadvantages of budget allocation systems
• Not as intuitive.• Allows organizational unit to retain authority for choosing
which activities to conduct—most appropriate for decentralized decision making
• Interdependencies among activities do not create a problem (since groups of activities are scored).
• Allows consideration of partial funding of activities (doesn’t assume all-or-nothing decisions).
• Doesn’t provide benefit estimates at the level of individual activities.
36c 2003, 2004 Lee Merkhofer Consulting
• Individual organizational units use a common ranking tool to evaluate and prioritize projects while upper organizational levels use a budget allocation system to allocate resource across the organizational units.
• The value-maximizing allocation matches slopes:
Combined (tiered) system
Integrated Budget Allocation/Activity Ranking System
Budget allocation system
Projectranking
tool
Project ranking
tool
Enterprise level
Individual organizational elements
Project ranking
tool
• Project prioritiesTotal benefit vs. total funding curves
• Funding case guidanceBudget allocation criteria/weights
Element 1 funding
Ben
efit
Cost
Element 1Element 1
Element 2 funding
Ben
efit
Cost
Element 2Element 2
37c 2003, 2004 Lee Merkhofer Consulting
Advantages and disadvantages
• Automatically provides the necessary activity ranking information in the form needed for the budget allocation system
• More complex than either of the individual systems• Usually results as an expansion from a project ranking or
budget allocation system
38c 2003, 2004 Lee Merkhofer Consulting
Fun
ded
PROBLEMS1. PROBLEM W2. PROBLEM J3. PROBLEM K4. PROBLEM G
27. PROBLEM. H
Problem/opportunity ranking systems• Rank problems or opportunities, not projects or investments• Tell you where you should be focusing your attention, not what you should
be doing• Simpler and less constraining than other approaches
– No inputs required on costs or effectiveness of investments– Defends decision makers without constraining them—decision makers decide how
much to spend on top-ranked items • Typically used as “screening” systems
Each problem/opportunity is scored
Scores used to rank problems/opportunities
scoresform
scoresform
TECHNICAL DATA
39c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
40c 2003, 2004 Lee Merkhofer Consulting
The project prioritization problem
Find the set of projects that• Produces the greatest possible value (benefit)• Without exceeding the available budget
41c 2003, 2004 Lee Merkhofer Consulting
The project prioritization problem
Find the set of projects that• Produces the greatest possible value (benefit)• Without exceeding the available budget
Mathematicians refer to this a the “knapsack problem”
42c 2003, 2004 Lee Merkhofer Consulting
Mathematical formulation of the basic project prioritization problem
N = number of project opportunities
xi =
bi = benefit obtained if i’th project is fundedC = total available budget
0 if the i’th project is not funded
1 if the i’th project is funded
Maximize Σ bixi
Subject to Σ cixi < C
i = 1
i = 1
N
N
This is known as a “zero-one integer programming problem”
43c 2003, 2004 Lee Merkhofer Consulting
Prioritization is surprisingly difficult to solve mathematically
• The problem is the number of alternative project combinations.– If there are N possible projects, there are 2N possible project
portfolios– E.g., with 30 projects there are roughly 1 billion alternative project
portfolios!
• Most software packages for integer programming use the “branch and bound” solution technique.
• The solution is time consuming, even with modern computers.
• If there are more than about 200 possible projects, even the fastest computers may not be able to find the solution.
• Thus, most priority systems do not use branch-and-bound
44c 2003, 2004 Lee Merkhofer Consulting
Fortunately, a theorem from economics often applies • Suppose the projects are independent
– Any combination of projects can be selected– Neither the costs nor benefits of any project depend on what other projects are funded
45c 2003, 2004 Lee Merkhofer Consulting
Cumulative cost
Cumulative benefit
B/CBenefitCostProjectRank
2371.50.524M71969.5111J61868.5263H51562.53217G4841.54.54.51K37375255X22126122w1
Fortunately, a theorem from economics often applies • Suppose the projects are independent
– Any combination of projects can be selected– Neither the costs nor benefits of any project depend on what other projects are funded
• Then, ranking projects by the ratio of benefit-to-cost and funding from the top down approximately gives the value-maximizing project portfolio
46c 2003, 2004 Lee Merkhofer Consulting
Fortunately, a theorem from economics often applies • Suppose the projects are independent
– Any combination of projects can be selected– Neither the costs nor benefits of any project depend on what other projects are funded
• Then, ranking projects by the ratio of benefit-to-cost and funding from the top down approximately gives the value-maximizing project portfolio
Cumulative cost
Cumulative benefit
B/CBenefitCostProjectRank
2371.50.524M71969.5111J61868.5263H51562.53217G4841.54.54.51K37375255X22126122w1
Budget constraint:
$20 million
Funded projec
ts
47c 2003, 2004 Lee Merkhofer Consulting
The project ranking shows the order in which projects are added
Cum
ulat
ive
bene
fit
Cumulative cost
0.5123
4.556
B/C
71.569.568.562.541.53712
Cumulative benefit
7654321
Rank
MJHGKXw
Project
23191815872
Cumulative cost
$20 M
5 10 15 2520
Added costAdded value
10
30
20
50
0
40
70
80
60
Budget constraint
…which can be plotted to show the “Efficient Frontier”
48c 2003, 2004 Lee Merkhofer Consulting
Important considerations
• the proper ranking metric is “bang-for-the-buck.”– Not “bang”– Not “alignment with strategy”– Not “balance”– Not “points”
Note that:
49c 2003, 2004 Lee Merkhofer Consulting
Important considerations
• the proper ranking metric is “bang-for-the-buck.”– Not “bang”– Not “alignment with strategy”– Not “balance”– Not “points”
• ranking gives and approximate solution, not an exact one.
Note that:
50c 2003, 2004 Lee Merkhofer Consulting
Important considerations
• the proper ranking metric is “bang-for-the-buck.”– Not “bang”– Not “alignment with strategy”– Not “balance”– Not “points”
• ranking gives and approximate solution, not an exact one.• ranking will give the wrong answer if there are
interdependencies
Note that:
51c 2003, 2004 Lee Merkhofer Consulting
Case study: Getting the concepts right is critical
Example—Suspiciously performing priority system
– Company was using a scoring approach recommended by consultants:
» Candidate projects were awarded points based on performance in several dimensions (financial reward, strategic fit, leverage, probability of success)
» Points were added and results used to rank projects
– Something was wrong» Large project nearly always ranked near the top
52c 2003, 2004 Lee Merkhofer Consulting
Case study: Getting the concepts right is critical
Example—Suspiciously performing priority system
– Company was using a scoring approach recommended by consultants:
» Candidate projects were awarded points based on performance in several dimensions (financial reward, strategic fit, leverage, probability of success)
» Points were added and results used to rank projects
– Something was wrong» Large project nearly always ranked near the top
– Approach had numerous “fatal flaws”» Most notably, failing to divide by cost (to get bang-for-the-buck) meant
that the logic was wrong. The system was biased toward large projects.
53c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
54c 2003, 2004 Lee Merkhofer Consulting
What exactly is “value”
“Asset management should not just create value, it should create the greatest possible value”
(Asset Management Software Vendor)
55c 2003, 2004 Lee Merkhofer Consulting
What exactly is “value”?
“Asset management should not just create value, it should create the greatest possible value”
(Asset Management Software Vendor)
• Such financial metrics fail to capture the true objectives of organizations
Two views on value:- shareholder value- stakeholder value
• Value is NOT – ROI, – ROA, – IRR, – NPV, – pay-back period, – etc.
56c 2003, 2004 Lee Merkhofer Consulting
Many US business leaders and management scientists believe the one and only goal of business is to create value for shareholders
“Maximizing shareholder value is becoming the number one priority for most publicly listed corporations.”
CFO, Price Waterhouse, 1997.
57c 2003, 2004 Lee Merkhofer Consulting
Many US business leaders and management scientists believe the one and only goal of business is to create value for shareholders
“Maximizing shareholder value is becoming the number one priority for most publicly listed corporations.”
CFO, Price Waterhouse, 1997.
“Maximizing shareholder value is now embraced as the “politically correct” stance by corporate board members and top management in the United States.”
Alfred Rappaport, Creating Shareholder Value, 1998.
58c 2003, 2004 Lee Merkhofer Consulting
Many US business leaders and management scientists believe the one and only goal of business is to create value for shareholders
"Managing for shareholder value...pays off... We refined the goal over the years, constantly trying to come up with measures that better reflected the intrinsic value of the company andeach of the businesses. But we always maintained a single overall objective: generating greater value for the shareholder"
Brian Pitman, CEO of Lloyds Bank, HBR, April 2003.
“Maximizing shareholder value is becoming the number one priority for most publicly listed corporations.”
CFO, Price Waterhouse, 1997.
“Maximizing shareholder value is now embraced as the “politically correct” stance by corporate board members and top management in the United States.”
Alfred Rappaport, Creating Shareholder Value, 1998.
59c 2003, 2004 Lee Merkhofer Consulting
Arguments for adopting shareholder value as the ultimate metric
• For publicly held companies, the value assigned by markets is, arguably, the final measure of business success
60c 2003, 2004 Lee Merkhofer Consulting
Arguments for adopting shareholder value as the ultimate metric
• For publicly held companies, the value assigned by markets is, arguably, the final measure of business success
• Market valuations are objective and applied consistently across all publicly held companies and markets– Assured by principle of no arbitrage
61c 2003, 2004 Lee Merkhofer Consulting
Arguments for adopting shareholder value as the ultimate metric
• For publicly held companies, the value assigned by markets is, arguably, the final measure of business success
• Market valuations are objective and applied consistently across all publicly held companies and markets– Assured by principle of no arbitrage
• Companies not managing for shareholder value are candidates for investor takeover
62c 2003, 2004 Lee Merkhofer Consulting
• Shareholder value is NOT the risk-adjusted, discounted value of projected future cash flows.
Managing for shareholder value and managing for profitable cash flows are two, very different things
63c 2003, 2004 Lee Merkhofer Consulting
Managing for shareholder value and managing for profitable cash flows are two, very different things
• Shareholder value is NOT the risk-adjusted, discounted value of projected future cash flows.
• Example (4 power delivery companies):
Gap (=“option value”)
NPV of projected earnings (discounted at WACC)
64c 2003, 2004 Lee Merkhofer Consulting
Managing for shareholder value and managing for profitable cash flows are two, very different things
• Shareholder value is NOT the risk-adjusted, discounted value of projected future cash flows.
• Example (4 power delivery companies):
• What causes these differences?
Gap (=“option value”)
NPV of projected earnings (discounted at WACC)
65c 2003, 2004 Lee Merkhofer Consulting
Stakeholder Value represents the alternative view
“It is probably safe to say there remains, worldwide, a large segment of the people that highly doubts the business goal of maximizing shareholder value is aligned with the interests of society at large.”
Bartley Madden, CFROI Valuation, 2000.
66c 2003, 2004 Lee Merkhofer Consulting
Stakeholder Value represents the alternative view
“It is probably safe to say there remains, worldwide, a large segment of the people that highly doubts the business goal of maximizing shareholder value is aligned with the interests of society at large.”
Bartley Madden, CFROI Valuation, 2000.
The alternative view is that businesses have obligations to other stakeholders as well:
• customers• employees• the community• suppliers• business partners• etc.
Company leaders must decide whether stakeholder values matter and what weight (if any) to assign to each stakeholder.
67c 2003, 2004 Lee Merkhofer Consulting
What kind of metrics reflect impact on value?
• Most organizations take the wrong approach. They measure what is easy to measure, not what is important.
68c 2003, 2004 Lee Merkhofer Consulting
What kind of metrics reflect impact on value?
• Most organizations take the wrong approach. They measure what is easy to measure, not what is important.
• They take a bottom-up approach—they define interesting metrics, but then can’t come up with algorithms for computing value based on those metrics.
69c 2003, 2004 Lee Merkhofer Consulting
What kind of metrics reflect impact on value?
• Most organizations take the wrong approach. They measure what is easy to measure, not what is important.
• They take a bottom-up approach—they define interesting metrics, but then can’t come up with algorithms for computing value based on those metrics.
• Unless there is a way to logically combine the metrics to determine the value added by projects, the metrics will not be of much use for prioritizing projects.
70c 2003, 2004 Lee Merkhofer Consulting
Use “value modeling” to define metrics
Value modeling• Subject of numerous books and articles, taught
any many graduate school business and engineering programs.
• Also called− “multiattribute utility analysis”− “multicriteria decision making”− “multi-objective decision analysis”
• Goal is to construct the most accurate and objective measure of value possible
System Performance
System Performance
SafetySafety PowerQualityPowerQuality ReliabilityReliability
Net RevenueNet Revenue EnvironmentEnvironment
71c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
72c 2003, 2004 Lee Merkhofer Consulting
The philosophy underlying the decision analysis approach to prioritization
• Resources are insufficient to do everything at once.• Things that are most valuable and most urgent should be done
first.• The value of an investment or other alternative is determined by
the degree to which it contributes to the achievement of objectives.
• Objectives and the weights to be assigned to them should be established by policy makers.
• In the absence of hard data and reliable models, estimates of the ability of alternatives to achieve objectives should be providedby those most knowledgeable about the alternatives and the needs that they will address.
• Techniques should be used to reduce biases and discourage gaming.
73c 2003, 2004 Lee Merkhofer Consulting
Basic steps for developing a priority system
1. Identify and structure the objectives– Specify what you want
2. Define performance measures– Specify how you will measure the degree to which an
alternative would achieve each objective
3. Derive the aggregation equation and decision rule– Specify how you will combine the performance measures
– Specify relative importance or weights
– Equation may be linear or non-linear and is derived using value modelling principles
– Define how projects will be prioritized
74c 2003, 2004 Lee Merkhofer Consulting
A simple example
What to pack for vacation?
AlternativesLimited resource: space in suitcase(5,000 cubic inches)
75c 2003, 2004 Lee Merkhofer Consulting
Structuring objectives as a hierarchy shows how achieving lower-level objectives (that are impacted by choices) enable higher-level objectives to be achieved.
Step 1: Identify and structure objectives
Avoid bringing heavy items
Avoid bringing big items
Bring items whose use enhances enjoyment
Bring items likely to be used
Maximize enjoyment of
vacation
Maximize fun Minimize work
76c 2003, 2004 Lee Merkhofer Consulting
Step 2: Define performance measures
Objectives Performance Measures Required Estimates
Avoid bringing items big items
Avoid bringing heavy items
Bring items whose use enhances enjoyment
Bring items likely to be used
Estimated volume (length x width x height, in inches)
Estimated weight in pounds
Benefit score:
Estimated number of times item would be used
Item size
Item weight
Benefit per use
Frequency of use
4 = Major benefit added3 = Significant benefit added 2 = Moderate benefit added
(about ½ that judged “major”)1 = Minor benefit added0 = No benefit added
77c 2003, 2004 Lee Merkhofer Consulting
Step 3: Derive the aggregation equation and corresponding decision ruleExample equation:
Ranking Measure = [ 80 x (Frequency of Use x Benefit per Use) - 20 x Weight ] / Size
78c 2003, 2004 Lee Merkhofer Consulting
Desired characteristics
• Logically sound (defensible) and accurate• Complete
79c 2003, 2004 Lee Merkhofer Consulting
Project benefit = Value (with project) –Value (without project)
OR
Project benefit = Value (project done now)-Value (project delayed)
Wn$ W1
1. Structure objectives
3. Estimate performance using data, models & judgment
4. Assess weights and other value parameters (scaling)
5. Combine to obtain total project value
Aggregationequation
Maximizeachievement of
organizational objectives
Maximizeshareholder
value
Maximize quality of organizational
image
Community
Maximize stakeholdervalue
Customers Workers
1
High
Low
5
1
High
Low
5
⌧
Value
Minimize adverse regulatory impacts
Overview of the logic for evaluating projects using a value model
2. Develop performance models/scales for lowest- level sub-objectives
2000 2001 2002 2003 2004Summary Income StatementPremiums $22,695.2 $23,998.8Net Investment Income 681.0 693.9Fees and Other Income 2,045.4 1,977.5
Total Rev. Excluding Cap. Gains $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1annual growth rate 22.8% 4.9% 5.0% 5.0% 5.0%
Hard Benefit Incremental Revenue, Portfolio 0.0 0.0 0.0 0.0 0.0Hard Benefit Incremental Revenue, Single Initiative 0.0 0.0 0.0 0.0 0.0
Total Revenue $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1
Current and Future Benefits (COGS) 19,340.2 20,433.4 21,562.9 22,641.0 23,773.0Gross Margin $6,081.4 $6,236.8 $6,440.9 $6,762.9 $7,101.0
percent of revenues 23.9% 23.4% 23.0% 23.0% 23.0%
W2
Financial model Scoring scales
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Performance models
"Hard" financialvalue
Platform forfuture success
Maximizeshareholder
value
Sub-objectives
Customers Workers Community
Maximizestakeholder
value
Minimize adverseregulatory impacts
Sub-objectives
Maximize qualityof organizational
image
Maximizevalue
80c 2003, 2004 Lee Merkhofer Consulting
Step 1: Structuring objectives
• The objectives hierarchy should consist of fundamental (not means) objectives.
For the value model to be logically sound, the objectives must satisfy certain rules…
81c 2003, 2004 Lee Merkhofer Consulting
Step 1: Structuring objectives
• The objectives hierarchy should consist of fundamental (not means) objectives.
• To ensure a simple aggregation equation (weight and add), the objectives should be “preferentially independent”– the importance of achieving any objective must not depend on the level of
performance achieved on any objective at the same level of the hierarchy
– this facilitates the evaluation of alternatives by ensuring that the weights specifying the importance of objectives will be constants
For the value model to be logically sound, the objectives must satisfy certain rules…
82c 2003, 2004 Lee Merkhofer Consulting
Step 1: Structuring objectives
• The objectives hierarchy should consist of fundamental (not means) objectives.
• To ensure a simple aggregation equation (weight and add), the objectives should be “preferentially independent”– the importance of achieving any objective must not depend on the level of
performance achieved on any objective at the same level of the hierarchy
– this facilitates the evaluation of alternatives by ensuring that the weights specifying the importance of objectives will be constants
• Lower-level objectives that are not preferentially independent cannot be weighted. Equations/models and other methods must be used to relate performance on these objectives to performanceon the higher-level objectives.
For the value model to be logically sound, the objectives must satisfy certain rules…
83c 2003, 2004 Lee Merkhofer Consulting
Case study: One T&D company’s view of value
• The first step was to get senior executives to agree on “value to whom?”
Example – Defining stakeholder value
Maximize total stakeholder value
Achievecompany financial
objectives
Provide quality,Low-cost service to
customers
Improve localcommunityperceptions
Reducecosts
Increaserevenue
Meet customer
commitments
Provide reliableservice
Charge reasonable
price
Timely & Accurate outagecommunicationShareholders Customers Citizens
84c 2003, 2004 Lee Merkhofer Consulting
Case study: One T&D company’s view of value
• The first step was to get senior executives to agree on “value to whom?”• The second step was to identify sub-objectives
Example – Defining stakeholder value
Maximize total stakeholder value
Achievecompany financial
objectives
Provide quality,Low-cost service to
customers
Improve localcommunityperceptions
Reducecosts
Increaserevenue
Meet customer
commitments
Provide reliableservice
Charge reasonable
price
Timely & Accurate outagecommunication
85c 2003, 2004 Lee Merkhofer Consulting
Sample objectives hierarchy developed for a county water utility
Provide healthy & & safe environment
and enhancedquality of life
Protect health &safety
Protectthe
Environment
Satisfy current& future
water needs
Prevent adverseeconomicimpacts tocommunity
Providerecreationalopportunities
86c 2003, 2004 Lee Merkhofer Consulting
Step 2: Defining metrics for measuring the degree to which objectives are achieved
• Performance measures quantify the degree to which objectives are achieved
• Performance measures should be “observables”– It’s 5 years in the future. What data could you collect to assess
performance against objectives?
• Choose measures that span the space of possibilities– Capture the attributes of what you expect to happen, and what might
happen.
87c 2003, 2004 Lee Merkhofer Consulting
• Objective: Customer satisfaction• Sub-objective: Minimize customer power outages• Performance measure: Outage duration• Possible events: Transformer trip, transformer fire,
explosion, fallen wire,…• Minimum duration--0 hours• Maximum duration--several days• Performance measure is a constructed scale based on
natural units.• Define scale using natural units.
Scale: 0 1 2 3
Example—Outage duration
88c 2003, 2004 Lee Merkhofer Consulting
Influence diagrams may be used to identify measures for creating constructed scoring scalesProcess for using influence diagrams:
1. Identify factors that influence achievement of the objective.2. Determine the relationship and relative importance among key factors
*Opportunity
for newservices &markets
Negotiatingleverage
Timeto market
Strategiccompetitivepositioning
In-sourcing/outsourcingpartnership
opportunities
*Technology
Ability totailor
Scaleability
Cable ofevolving
(adaptability)
Connectivity
Security
Electronicrecords
Abilityto conduct
surveysInformationrelevance
Informationscope (compre-hensiveness)
Robustgiven futureuncertainties
*Positive
visibility of actions tomarketplace
Info forunderstanding
value chain
Sub-objective:Provide
platform for future success
Illustrative
89c 2003, 2004 Lee Merkhofer Consulting
Based on value drivers in the influence diagrams, scales are developed for measuring each type of soft project benefit
*Opportunity
for newservices &markets
Negotiatingleverage
Timeto market
Strategiccompetitivepositioning
In-sourcing/outsourcingpartnership
opportunities
*Technology
Ability totailor
Scaleability
Cable ofevolving
(adaptability)
Connectivity
Security
Electronicrecords
Abilityto conduct
surveysInformationrelevance
Informationscope (compre-hensiveness)
Robustgiven futureuncertainties
*Positive
visibility of actions tomarketplace
To qualify for a score of 3, the initiatives must produce a significant positive impact on the company’s platform for success by allowing opportunities to provide new services or enter new markets and/or increase negotiating leverage, decrease time-to-market, facilitate profitable insourcingand outsourcing arrangements, and improve competitive position. This conclusion is based on a judgement that the initiatives will do one of the following(indicate which):1 Provide a significant new technology that is readily tailorable, scaleable, robust given future uncertainties, adaptable, and secure. For example, the technology will provide the greatest possible secure connectivity with all constituency groups, or2 Provide significant new and relevant information that will substantially improve understanding of the company value chain, or.3 Provide a clear signal to Wall Street that will be interpreted by analysts as evidence that company is exceeding expectations.
Info forunderstanding
value chain
Sub-objective:Provide
platform for future success
5
4
3
2
1
Scoring scale & sample definition
91c 2003, 2004 Lee Merkhofer Consulting
Step 3: Use data/models/best judgment to evaluate performance with and without each projectBasic logicKnowledge of…– Customers and markets– Levels of service provided (target and actual)– Existing assets and performance characteristics
2000 2001 2002 2003 2004Summary Income StatementPremiums $22,695.2 $23,998.8Net Investment Income 681.0 693.9Fees and Other Income 2,045.4 1,977.5
Total Rev. Excluding Cap. Gains $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1annual growth rate 22.8% 4.9% 5.0% 5.0% 5.0%
Hard Benefit Incremental Revenue, Portfolio 0.0 0.0 0.0 0.0 0.0Hard Benefit Incremental Revenue, Single Initiative 0.0 0.0 0.0 0.0 0.0
Total Revenue $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1
Current and Future Benefits (COGS) 19,340.2 20,433.4 21,562.9 22,641.0 23,773.0Gross Margin $6,081.4 $6,236.8 $6,440.9 $6,762.9 $7,101.0
percent of revenues 23.9% 23.4% 23.0% 23.0% 23.0%
Financial model
1
High
Low
5Scoring scales
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Performance models
92c 2003, 2004 Lee Merkhofer Consulting
Step 3: Use data/models/best judgment to evaluate performance with and without each projectBasic logicKnowledge of…– Customers and markets– Levels of service provided (target and actual)– Existing assets and performance characteristicsProvides us the ability to predict…– Demand– How and when assets will fail– The consequences of failures
2000 2001 2002 2003 2004Summary Income StatementPremiums $22,695.2 $23,998.8Net Investment Income 681.0 693.9Fees and Other Income 2,045.4 1,977.5
Total Rev. Excluding Cap. Gains $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1annual growth rate 22.8% 4.9% 5.0% 5.0% 5.0%
Hard Benefit Incremental Revenue, Portfolio 0.0 0.0 0.0 0.0 0.0Hard Benefit Incremental Revenue, Single Initiative 0.0 0.0 0.0 0.0 0.0
Total Revenue $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1
Current and Future Benefits (COGS) 19,340.2 20,433.4 21,562.9 22,641.0 23,773.0Gross Margin $6,081.4 $6,236.8 $6,440.9 $6,762.9 $7,101.0
percent of revenues 23.9% 23.4% 23.0% 23.0% 23.0%
Financial model
1
High
Low
5Scoring scales
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Performance models
93c 2003, 2004 Lee Merkhofer Consulting
Step 3: Use data/models/best judgment to evaluate performance with and without each projectBasic logicKnowledge of…– Customers and markets– Levels of service provided (target and actual)– Existing assets and performance characteristicsProvides us the ability to predict…– Demand– How and when assets will fail– The consequences of failuresWhich allows us to…– Evaluate alternative asset acquisitions, operations, maintenance, and
renewal options– Prioritize based on benefit/cost– Optimize the allocation of resources
2000 2001 2002 2003 2004Summary Income StatementPremiums $22,695.2 $23,998.8Net Investment Income 681.0 693.9Fees and Other Income 2,045.4 1,977.5
Total Rev. Excluding Cap. Gains $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1annual growth rate 22.8% 4.9% 5.0% 5.0% 5.0%
Hard Benefit Incremental Revenue, Portfolio 0.0 0.0 0.0 0.0 0.0Hard Benefit Incremental Revenue, Single Initiative 0.0 0.0 0.0 0.0 0.0
Total Revenue $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1
Current and Future Benefits (COGS) 19,340.2 20,433.4 21,562.9 22,641.0 23,773.0Gross Margin $6,081.4 $6,236.8 $6,440.9 $6,762.9 $7,101.0
percent of revenues 23.9% 23.4% 23.0% 23.0% 23.0%
Financial model
1
High
Low
5Scoring scales
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Performance models
94c 2003, 2004 Lee Merkhofer Consulting
Step 3: Use data/models/best judgment to evaluate performance with and without each projectBasic logicKnowledge of…– Customers and markets– Levels of service provided (target and actual)– Existing assets and performance characteristicsProvides us the ability to predict…– Demand– How and when assets will fail– The consequences of failuresWhich allows us to…– Evaluate alternative asset acquisitions, operations, maintenance, and
renewal options– Prioritize based on benefit/cost– Optimize the allocation of resourcesWhich, in turn, enables us to…– Maximize performance at the system level (and, thereby, maximize
value) and explain and defend choices
2000 2001 2002 2003 2004Summary Income StatementPremiums $22,695.2 $23,998.8Net Investment Income 681.0 693.9Fees and Other Income 2,045.4 1,977.5
Total Rev. Excluding Cap. Gains $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1annual growth rate 22.8% 4.9% 5.0% 5.0% 5.0%
Hard Benefit Incremental Revenue, Portfolio 0.0 0.0 0.0 0.0 0.0Hard Benefit Incremental Revenue, Single Initiative 0.0 0.0 0.0 0.0 0.0
Total Revenue $25,421.6 $26,670.2 $28,003.7 $29,403.9 $30,874.1
Current and Future Benefits (COGS) 19,340.2 20,433.4 21,562.9 22,641.0 23,773.0Gross Margin $6,081.4 $6,236.8 $6,440.9 $6,762.9 $7,101.0
percent of revenues 23.9% 23.4% 23.0% 23.0% 23.0%
Financial model
1
High
Low
5Scoring scales
Tota
l Ben
efit |
|||||||||
Total CostTotal BudgetTot
al B
enef
it
Performance models
95c 2003, 2004 Lee Merkhofer Consulting
Step 4: Assign scaling functions and weights to quantify value
• Scaling may be needed to specify the relative value of changes.– How good/bad is a score of 2 compared to a score of 1?
• Scaling can be accommodated by either– Applying a scaling function to the natural scale or to the uniform,
constructed scale– Making the scale non-uniform.
• Example: Project designed to reduce outage duration.
96c 2003, 2004 Lee Merkhofer Consulting
Scaling – Example
0 1 2 3Duration (days)
Is an outage of 3 days 3 times worse than an outage of 24 hours?
Is an outage of 1 hour 6 times worse than an outage of 10 minutes?
97c 2003, 2004 Lee Merkhofer Consulting
Scaling – Example
NaturalscaleNone10 min.1 hr.2 hr.4 hr.8 hr.16 hr.24 hr.3 days> 3 days
0 1 2 3Duration (days)
0 -100-20-15-7-3 -90
Value (0-to100 scale)
Like weights, scaling requires value judgments.
Outage duration (days)
Value
1 2 3
-100
-50
Non-uniform value scale
0-0.1-.6
-1.2-3-7
-15-20-90
-100
Scaling function
98c 2003, 2004 Lee Merkhofer Consulting
A quick way to assign weights
1. Decide which objective is “most important”Suppose a series of “ideal” alternatives existed, each of which would produce a “best possible” impact on one (and only one) objective. All other things equal, which would you choose?
2. Rank the objectives
3. Each participant (individually) assigns weights
4. Comparison of individual weights & discussion
5. Final group weight assignments
99c 2003, 2004 Lee Merkhofer Consulting
The simple approach involves assigning weights that represent the value of going from “worst” to “best”
The weight, w, represents what 100% of the scale is worth
0
0.5
1
Worst
Best
25%
Value added by hypothetical project = 25% x w
Hypothetical project that scores a 0.25
Worst performance we could expect
Best impact we could expect from projects
(Normalized) scores
100c 2003, 2004 Lee Merkhofer Consulting
Weight assignment can be facilitated using poker chips• Assign chips (weights) in proportion to the value of each
impact
• If one impact is judged twice as important as another, assign twice as many chips (etc.).
• Maximum of 100 chips.
• Assume an impact in one area produces no (positive or negative) impact in any other area.
• Number of chips on each objective corresponds to weights.
101c 2003, 2004 Lee Merkhofer Consulting
Step 5: Determine the form of the aggregation equation• Weight and add is only correct if objectives are preferentially
independent (need to check)
• Sometimes, performance measures should be multiplied• E.g., multiplying the likelihood something happens by the outcome if it happens.
• Appropriate ranking metric (if projects are independent and don’t differ much in their “window of opportunity”) is benefit (net value excluding budget year cost) divided by budget year cost
• If projects have limited windows of opportunities, useful ranking metric is dollar loss from delay divided by budget year cost
Project benefit = Value (with project) – Value (without project)
Loss from delay = Value (if project done this period) – Value (if project done next period)
Wn$ W1
Aggregationequation
Value
W2
103c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
104c 2003, 2004 Lee Merkhofer Consulting
Are risk analysts heroes?
• According to a 1997 Business Week review of Peter Bernstein’s book Against the Gods: The Remarkable Story of Risk:
“Statisticians, in the telling of Peter L. Bernstein, arenothing less than Promethean heroes…people whomastered the calculation of probabilities,…and stolefrom the gods something more precious than fire--namely, the understanding of risk.”
105c 2003, 2004 Lee Merkhofer Consulting
The key to addressing risk is “embracing uncertainty”
• Voltaire: “Doubt is not a pleasant condition, but certainty is an absurd one.”
• Francis Bacon: “If a man will begin with certainties, he shall end in doubts; but if he will be content to begin with doubts he shall end in certainties.”
106c 2003, 2004 Lee Merkhofer Consulting
Definition of risk
• Risk A situation wherein different things might happen, and at least one of the possibilities is bad.
• Note– Risk involves uncertainty and value– Risk is not “probability” or “probability times consequence”. These are
measures of risk.
Discrete Risk Continuous Risk
Consequence if risk event occurs = $1M loss
Probability of risk event = 1 in 10,000
107c 2003, 2004 Lee Merkhofer Consulting
Embracing uncertainty requires:
1. Characterizing the risk
2. Adjusting value based on organizational risk tolerance
108c 2003, 2004 Lee Merkhofer Consulting
The first step for characterizing risks is to identify all of the types of potential adverse consequences that are of concern
Types of riskTypes of risk
Publichealth &safety
Publichealth &safety
Workerhealth &safety
Workerhealth &safety
Environmentalresources
Environmentalresources Regulatory
compliance
Regulatorycompliance Corporate
image
Corporateimage Financial
performance
Financialperformance Employee
commitment
Employeecommitment Customer
satisfaction
Customersatisfaction
Risks can adversely impact the achievement of any and all objectives.
109c 2003, 2004 Lee Merkhofer Consulting
The next step is identify the sources of risk
• What are things that, if they happen could lead to the adverse consequences of concern?
• Some sources of risk are internal to the project:– Technology risk– Implementation risk– Schedule risk– Cost risk
• Some sources of risk are external to the project:– Governance risks (e.g., Enron)– Market risks (e.g., the economy)– Weather risks– Legal/regulatory risks
Project risks are often independent (uncorrelated) across projects; the project portfolio reduces the importance of these risks through diversification.
External risk can affect many or all of the projects simultaneously, making them potentially more important.
110c 2003, 2004 Lee Merkhofer Consulting
A simple way to address project risks is to use consequence and likelihood scales
Example (public health risk)
Very unlikely (1 chance in 10,000 or less)
.0001
Unlikely (1 chance in 1000)
.001
Possible (1 chance in 100)
.01
Reasonably likely (1 chance in 10)
.1
Certain or almost certain
1
Probability
No health impacts would occur.0
Only small amounts of minor hazards are involved, or exposures will be highly diluted. Any health impacts that could occur will have minor consequences (e.g., discomfort) and be short term.
.1
Exposures to moderate hazards would occur. Any health impacts that occur would have moderate consequences (e.g., illness) and be temporary.
1
Exposures to significant hazards would occur. Health effect could result in extended hospital stay.
10
If the event happens, the health impact would be most serious (e.g., death, cancer, paralysis), due to exposures to extremely dangerous hazards (large amounts of deadly substances)
100
Consequence
111c 2003, 2004 Lee Merkhofer Consulting
If uncertainties are estimated, they can be propagated through models to quantify the uncertainty over value
Project Performance Risks External Uncertainties
Scenario 1Prob P-1
Scenario 2Prob P-2
Scenario 3Prob P-3
Project A PerformanceProb
abili
ty D
ensi
ty
Projects
•Project A
•Project B
•Project N
Composite Risk Model
Risk-SourceModel
RiskPropogation
ModelConsequence
Model
Prob
abili
ty D
ensi
ty
Number of Health Effects
Portfolio Risk Estimates
112c 2003, 2004 Lee Merkhofer Consulting
Quantifying uncertainty allows confidence bounds to be placed on performance forecasts
113c 2003, 2004 Lee Merkhofer Consulting
Sophisticated methods are available for quantifying risk tolerance
• However, unless project involve “bet the company” gambles, it is usually not necessary to use sophisticated techniques.
114c 2003, 2004 Lee Merkhofer Consulting
A simple (although imperfect) way to account for risk tolerance is through hurdle rates
• The hurdle rate is a risk-adjusted cost of capital used to discount future project costs and benefits
• Higher hurdle rates are applied to projects considered to be more risky
25%High risks (e.g., untested technology)WACCAverage risks
6%Below average risksRisk-free rateVirtually no risk
Project Risk Hurdle Rate
EXAMPLE
• Hurdle rates are a “quick-and-dirty” way of addressing project risk.
115c 2003, 2004 Lee Merkhofer Consulting
Risk and project prioritization
• Characterizing risk is relatively easy (and useful)– E.g., rather than obtain a “point estimate” for some uncertain aspect of
project performance, obtain a range.– If there are rare, but high consequence risks, estimate performance
contingent on the occurrence of the risk, plus a probability.– Be careful of external risks and other risks that may affect several or all
projects simultaneously. They can create portfolio risks.
• Characterizing risk tolerance is usually not that important– Exception is “bet the company” projects, for which methods more
sophisticated than hurdle rates should be used.
116c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
117c 2003, 2004 Lee Merkhofer Consulting
Organizing for project prioritization requires recognizing the importance of managing at the portfolio level
• Most organizations put ample effort into making individual projects successful– Highly motivated project managers
– Once funding obtained, focus is on achieving the costs, schedules, and performance mandates of their specific projects
• …but put insufficient effort into making the entire portfolio of projects as successful as it could be– Either no one has responsibility for the overall portfolio, or
management of the project portfolio is not as effective as it could be.
118c 2003, 2004 Lee Merkhofer Consulting
Problems with project-by-project decision making
• In most companies, the basis of a “go” decision is whether the project passes some hurdle
– Many projects pass the hurdle
– ..but, resources are highly constrained
– ..so people become more than 100% committed
– …which leads to the downward spiral of poor project portfolios.
119c 2003, 2004 Lee Merkhofer Consulting
Need to establish a project portfolio management office• Project portfolio manager has accountability for the
success of the entire project portfolio– PPM given an estimate of funding available– Up to PPM to decide how to allocate the funds across projects (at
minimum, PPM makes recommendations for final approval by executive committee)
120c 2003, 2004 Lee Merkhofer Consulting
Need to establish a project portfolio management office• Project portfolio manager has accountability for the
success of the entire project portfolio– PPM given an estimate of funding available– Up to PPM to decide how to allocate the funds across projects (at
minimum, PPM makes recommendations for final approval by executive committee)
• The portfolio manager is supported by a portfolio management team
121c 2003, 2004 Lee Merkhofer Consulting
Need to establish a project portfolio management office• Project portfolio manager has accountability for the
success of the entire project portfolio– PPM given an estimate of funding available– Up to PPM to decide how to allocate the funds across projects (at
minimum, PPM makes recommendations for final approval by executive committee)
• The portfolio manager is supported by a portfolio management team
• Team is responsible for – Evaluating project proposals– Validating project cost and performance estimates– Accepting/rejecting proposals– Accelerating and decelerating projects– Allocating resources– Continually managing the project portfolio
122c 2003, 2004 Lee Merkhofer Consulting
Need to establish a project portfolio management office• Project portfolio manager has accountability for the
success of the entire project portfolio– PPM given an estimate of funding available– Up to PPM to decide how to allocate the funds across projects (at
minimum, PPM makes recommendations for final approval by executive committee)
• The portfolio manager is supported by a portfolio management team
• Team is responsible for – Evaluating project proposals– Validating project cost and performance estimates– Accepting/rejecting proposals– Accelerating and decelerating projects– Allocating resources– Continually managing the project portfolio
• One member of the team is appointed primary contact for each project manager
124c 2003, 2004 Lee Merkhofer Consulting
Topics• Project prioritization and T&D asset management• Introduction to priority systems• Mathematics of project prioritization• Valuing projects and project portfolios• Creating a priority system• Accounting for risk• Organizing for project portfolio management• Selecting project prioritization tools• Case study examples
– O&M resource allocation– Risk management– T&D project prioritization
125c 2003, 2004 Lee Merkhofer Consulting
Numerous analytic tools are being promoted for asset management
Do such tools really help? What characteristics does a good tool need to have?What type of tool is best?
• Simulation models• Scoring tools• Executive dashboards• Narrow-focused applications• Enterprise-wide applications• Generic software• Custom software
Examples:According to META Group estimates, the project portfolio management tool market was about $85 million in 2002; by 2005 it could reach $540 million.
126c 2003, 2004 Lee Merkhofer Consulting
To evaluate a tool, consider how well it performs against 6 criteria
1. Accuracy2. Logical soundness3. Completeness4. Practicality5. Effectiveness6. Acceptability
The remainder of this section will clarify these considerations and provide observations, examples, and advice.
127c 2003, 2004 Lee Merkhofer Consulting
The tool must be accurate
• Does the tool:– Produce the “right” answer?– Provide outputs within an acceptable confidence level and precision?– Indicate the confidence or uncertainty associated with outputs?
• Is the tool biased toward or against certain projects, interests, or considerations?
• Are results highly sensitive to untested or untestable assumptions?
128c 2003, 2004 Lee Merkhofer Consulting
Observation: Intuitive and well-established models are not necessarily accurate
Example - Models for Ranking Superfund Sites
• Congress required the EPA to evaluate whether its Hazard Ranking System (used to place sites on the National Priorities List) was the best-available site-ranking model.
• The analysis consisted of comparing site rankings produced by various models with a ranking produced by an independent expert panel.
• Results showed none of the models produced rankings that correlated with each other or with the ranking provided by the expert panel (Call and Merkhofer 1988).
If a tool is logically sound and complete, its recommendations are more likely to be accurate.
129c 2003, 2004 Lee Merkhofer Consulting
The tool must be logically soundConsiderations for assessing logical soundness
• Degree to which tool can be justified (in terms of relevant theories and empirical evidence)
• Operational validity – Are there problems with satisfying underlying assumptions?
Note: Mathematical theories for valuing and prioritizing projects exist and are well-accepted within the technical and academic communities (e.g., multi-attribute utility analysis, portfolio theory, real options, etc.).
130c 2003, 2004 Lee Merkhofer Consulting
Observation: Tools not based on the relevant theories are time bombs—they go off when subjected to technical review Example – DOE’s System for Ranking Sites for the Nuclear Waste Repository
• Hanford
• Yucca Mtn
• Deaf Smith County
• Davis Canyon
• Richton Salt Dome
• The DOE initially used a scorecard approach to rank sites being considered as locations for the repository.
• Hanford, Washington, was the top-ranked site.
• Responding to critics, DOE had the National Academy of Sciences (NAS) review the analysis.
• The NAS concluded that DOE’s method was “unsatisfactory, inadequate, undocumented, and biased.”
• DOE was forced to redo the analysis using an academically “correct” methodology (multi-attribute utility analysis).
• The result caused DOE embarrassment and forced it to revise its choice to Yucca Mountain, Nevada. (Merkhofer and Keeney 1987).
131c 2003, 2004 Lee Merkhofer Consulting
The tool must be complete, accounting for allrelevant decision considerations
• Does the tool ignore important considerations?• Is it possible to expand the tool to include omitted
considerations, or are the omissions inherent in the approach?
An enterprise-level view of asset management demands capability to evaluate impacts on all enterprise objectives.
132c 2003, 2004 Lee Merkhofer Consulting
Observation: Completeness requires capability to handle “hard” and “soft” considerations
Example – Prioritizing Investments in a Water UtilityTo fully value investments, it was necessary to
- Account for “soft” benefits, such as public health and safety, reduced flood risk, improved customer perceptions, community recreational opportunities, and increased organizational knowledge and capability.
- Integrate information from numerous databases, including customer information systems, GIS, work management systems, investment modeling applications, etc.
- Include scoring methods to capture local knowledge possessed by field personnel but not captured by available data.
The need for accuracy, logical soundness, and completeness is why asset management tools must be sophisticated.
133c 2003, 2004 Lee Merkhofer Consulting
The tool must be practical to implement and use
Do you have:- Necessary expertise to apply the tool?- Data required for input?- Computation resources for applications?- Sufficient time for applications (including time to collect input data)?
A utility can make it practical to use more sophisticated tools through training, reassigning responsibilities, developing new sources of data, and adjusting budget schedules.
134c 2003, 2004 Lee Merkhofer Consulting
Observation: Experience creates capability and demand for more sophisticated tools
• DOT regulations required client to adopt formal system for identifying risks and prioritizing risk-reducing activities.
• Implemented in 1995, the system was initially perceived by many as “overly complex.”
• However, each year the client expanded the system, making it more complete and sophisticated.
• Experience, better documentation, and internal training built confidence and expertise, promoting desire for more capability and accuracy.
Example - Pipeline Risk Management System
Avoid simplistic tools and tools whose capabilities cannot be expanded as experience and understanding grows.
135c 2003, 2004 Lee Merkhofer Consulting
The tool must be effective at achieving its intended goals
How will the tool be used:- As a backroom decision aid?- As a vehicle for creating internal consensus?- As a vehicle for persuading external parties?
What does the tool need to do well to address limitations of current practices?
136c 2003, 2004 Lee Merkhofer Consulting
Observation: Sometimes, being effective requires a new approach
Example—Valuing Long-Term Energy Contracts
– Utility client needed way to compare short vs. long term (>10yr.) energy contracts.
– Standard valuation methods were overly sensitive to assumed discount rate.
– Delivered system uses real options analysis, a new technique that » quantifies the “option value” of contracts (an energy contract
provides an option to buy throughout the duration of the contract).» does not require risk-adjusted discount rates
Actual Market value
Value computed using traditional methods
Opt
ion
valu
e
137c 2003, 2004 Lee Merkhofer Consulting
The tool must be acceptable to decision makers and other stakeholders
Is the tool:- Compatible with existing processes/organizational structure?- Understandable and understood?
Stakeholders feel threatened by tools that impact funding. For success, they must have confidence that the tool will help them and the business succeed.
Unless all parties want it, understand it, and trust it, the tool will be abandoned.
138c 2003, 2004 Lee Merkhofer Consulting
Observation: Developing the tool as a collaborative effort promotes buy-in and allows the tool to be designed to fit the need
Example – Prioritizing Investments in Information Technology
• Client desired system for prioritizing proposals submitted by different departments.
• Core Systems Department convinced tool would not work, “How do you value investments in capacity?”
• Design gives infrastructure that enables other projects a portion of benefit generated by enabled projects.
• Tool encourages Core Systems to propose infrastructure that serves needs of other departments.
140c 2003, 2004 Lee Merkhofer Consulting
Background
Motivation and objectives• Western electric and gas utility operates
roughly 3000 facilities in seismically active locations
• Asset risk manager concerned that senior management does not appreciate the risks and is not providing sufficient resources for risk management
The developed system• Screens facilities to identify those that warrant formal analysis• Ranks facilities based on total seismic risk• Quantifies the risk-reducing benefits of activities proposed to reduce
risk• Prioritizes activities based on benefits and costs• Provides estimates of total seismic risk and ability to reduce total risk
as a function of risk-management budget
141c 2003, 2004 Lee Merkhofer Consulting
• Projects ranked according to benefit/cost• Typically used with an all-or-nothing, fund-from-the-top-down decision rule
• Produce a funding curve showing which activities to cut (add) and specific benefits lost (gained) if budget is decreased (increased).
Project ranking system
Fun
ded
PROJECT CUM. COSTS1. PROJ. W $1M2. PROJ. X $5M3. PROJ. K $8M4. PROJ. G $12M
27. PROJ. H $37M------------------------------------------28. PROJ. M $43M
Project costs and benefits evaluated
Projects ranked based on B/C
scoresform
scoresform
TECHNICAL DATA
Tota
l Ben
efit
Added costAdded value
Total value
Total cost
Organizational element funding
Ranking produces total benefit versus total cost curve
142c 2003, 2004 Lee Merkhofer Consulting
The objectives hierarchy for the system was developed by the project team and approved by senior management
143c 2003, 2004 Lee Merkhofer Consulting
To assess risks…
• Seismic risk models are used to quantify ground-shaking scenarios
• Structure models are used create damage scenarios• Scoring scales used to convert damage scenarios into
impacts on objectives
1
High
Low
5
Scoring scales
144c 2003, 2004 Lee Merkhofer Consulting
Example - Scale for worker health and safety
How many worker fatalities do you expect to result if the facility experiences the indicated damage outcome? Average for changes in population over time.
Score
0. None
1. 1 chance in 1000 of one fatality2. 1 chance in 100 of one fatality
3. 1 chance in 10 of one fatality
4. One fatality
5. 10 fatalities
6. 100 fatalities
145c 2003, 2004 Lee Merkhofer Consulting
Example - Legal liability scale
Score0. No chance1. 1 chance in 10 or less2. Between 1 chance and 50%3. Between 50% chance and 90%4. Almost certain
If the damage outcome occurs, how likely is it that the company will be held legally liable for impacts attributed to facility damage?
Score0. No chance1. Less than $1M2. Between $1M and $10M3. Between $10M and $100M4. Between $100M and $1B5. More than $1B
What is the most likely total dollar impact to the company if it is held legally liable for impacts?
146c 2003, 2004 Lee Merkhofer Consulting
Example – System functionality
Score0. None1. Less than 10,0002. Between 10,000 and 100,0003. Between 100,000 and 1,000,0004. More than 1,000,000
If the seismic event and damage outcome to the facility occurs, how many customers will lose service?
Score0. None1. Less than an hour2. Between 1 and 6 hours3. Between 6 and 12 hours4. Between 12 hours and 1 day5. Between 1 day and 3 day6. More than 3 days
If the facility damage outcome occurs, what is the likely duration of the lost service that it will cause, or, if the facility is needed to return service, by how long will service return be delayed?
147c 2003, 2004 Lee Merkhofer Consulting
Weights were assigned by company executives
One incidence of 3-day power loss to 3000 residential customers
No outage
One chance in 100 of a public fatality
No risk
150c 2003, 2004 Lee Merkhofer Consulting
Projects are prioritized based on the ratio of benefit-to-cost
151c 2003, 2004 Lee Merkhofer Consulting
Summary
• The priority system demonstrated to senior management that seismic risks are serious.
• Enabled proposed investments to be prioritized based on risk-reduction benefit.
• Forced senior management to take ownership of the problem.
• Provided auditable documentation of efforts to address risk.
Case Study: A Budget Allocation System
Originally prepared by
Elsie Martin, Myers Martin Consulting, LLCLee Merkhofer, Lee Merkhofer Consulting
for the
EPRI Power Delivery Asset Management WorkshopJune 3-5, 2003, New York, NY
153c 2003, 2004 Lee Merkhofer Consulting
Case study materials previously presented in public forums
• Xcel entered the application in the Institute for Operations Research and the Management Sciences (INFORMS) 2000 competition for “Best Application of Decision Analysis”
– Awarded “runner up” status
– Material herein presented during judging of finalists
• This material presented at the EPRI Power Delivery Asset Management Workshop, June 3-5, 2003, New York, NY
154c 2003, 2004 Lee Merkhofer Consulting
Background: Application describes priority system used 1998-2000 by Northern States Power (NSP)
• NSP is a large, Midwestern electric utility, now part of Xcel Energy
• Goal was to help NSP (Electric) allocate its $250 million operating budget
• The priority system was used to help the Asset Management department allocate operating funds to diverse business areas responsible for
– Delivery (Transmission & Distribution)– Retail Services (Call Center, Billing, Marketing & Sales)
155c 2003, 2004 Lee Merkhofer Consulting
By 1998, it was clear that the existing budget allocation process was no longer adequate
• Lengthy discussions, eloquence always won
• Sub-optimization by territory was not consistent with new matrix structure
• Need to deliver more service without increasing costs
• Although NSP had experience using a ranking system for prioritizing capital expenditures, such an approach was determined inadequate for operating budget decisions
156c 2003, 2004 Lee Merkhofer Consulting
Challenge was to create a priority system that would:
• Identify and quantify allsignificant benefits of operating expenditures
• Provide a structured process with appropriate input from all stakeholders
• Level the playing field
• Align budgets with objectives
• Incorporate available data on performance & from surveys as well as best-professional judgment on expenditure effectiveness
• Show what is gained/lostif budget is increased/decreased
• Improve “bang for the buck”
157c 2003, 2004 Lee Merkhofer Consulting
Budget Allocation
Funds Unit A Unit B Unit C Unit D$15M 2 3 5 5$16M 3 3 5 5$17M 3 4 5 5$18M 3 4 6 5$19M 4 4 6 5$20M 4 4 6 6$21M 4 5 6 6$22M 5 5 6 6$23M 5 5 6 7$24M 5 5 6 8$25M 5 5 7 8$26M 5 6 7 8$27M 5 7 7 8
::
System based on a budget allocation design• Provides an optimal allocation of a fixed budget among competing organizational
units • Each organizational unit proposes activities to be conducted under several
different, pre-specified funding cases (e.g., base-case funding, reduced funding, enhanced funding, etc.).
• Each funding case is evaluated (based on the activities to be funded).
Unit A (E-bus) Funding CasesFunding Case
Activity Min Reduced Base MaxB2B site $1M $2M $2M $2MWeb store $1M $1M $2M $2MKid’s lifestyle portal 0 0 $1M $1MRecruiting site 0 0 0 $1M
Total $2M $3M $4M $5M
Each unit prioritizes activities, indicates funding under alternative unit funding
cases, and scores funding cases
Scores used to determine optimal allocation of total funds
• “Bang for the buck” used to allocate funds across sites
158c 2003, 2004 Lee Merkhofer Consulting
The system was designed to allocate the budget across 10 business areas
• Business areas, called “portfolios” addressed various transmission, delivery, and retail services.
• Alternative 5-year funding scenarios were evaluated for each portfolio.
• Each portfolio team proposed between 3 and 9 alternative funding cases (5-year scenarios).
• Each funding case is scored against 7 objectives.
159c 2003, 2004 Lee Merkhofer Consulting
Each portfolio team considered several alternative funding scenarios
• Case 1 = minimum funding (minimum funding required to meet legal obligations)
• Case 5= base case (five years at `99 actuals)
• Case 9 = maximum funding (unconstrained funding) all activitieswith benefits that justify costs
• Cases 2-4, 6-8 = intermediate funding scenarios ($1-2M max between cases)
• Portfolio team specified activities to be funded under each case, then “scored” cases to estimate benefits
Case 5
Activity 3
Activity 1Case Year 1 Year 2 Year 3 ...1 0 0 $15 $1M $1.2M $1.2M9 $7M $7M $7.1M
Activity 2
Case 1
Yr 1 2 3 4 5
Case 9
160c 2003, 2004 Lee Merkhofer Consulting
Policy makers created the objectives hierarchy for evaluating the cases
Better, Faster,Cheaper
Financial CustomerService
Platform forSuccess
MeetCommitments
& BuildRelationships
ReliabilityAccurate &
TimelyBilling
Impact onCost toOthers
Profitsfrom Value-
addedServices
OutageCommunication
SystemReliability
161c 2003, 2004 Lee Merkhofer Consulting
A model was developed for quantifying benefits
• Measures and scales (derived from influence diagrams) were created for each objective to “score” performance.
• Weights for objectives were assessed from senior NSP Electric executives.
• A mathematical function was derived to translate “scores” and “weights”into estimates of “utility,” a measure of the value of each portfolio.
• System software computed the costs and benefits of all possible combinations of funding levels.
162c 2003, 2004 Lee Merkhofer Consulting
Influence diagrams were constructed to help select performance measures
For each funding case, estimate:• frequency of sustained outages• frequency of momentary outages• number of customers with > 4
sustained outages• average duration of outages• measure of power quality
1. Influence diagrams identify factors influencing customer’s perception of system reliability
Example: System reliability objective
2. Key factors define measures and scales used to forecast performance
`99 actuals3. Data on recent
performance provided as benchmark for scoring
Systemreliability
Outages Powerquality
CustomertypeFrequency
Duration
Momentary
Sustained
# ofcustomersaffected
Frequency
Magnitudeof disruption
163c 2003, 2004 Lee Merkhofer Consulting
Portfolios (equivalent
dollars)
Portfolio cost
Case 1
Case 5
Case7Case 9
Example: Billing and Payment Portfolio
Scaling functions were developed to translate performance “scores” into estimates of dollar value
• Scaling functions (e.g., reliability function) derived from experts
• Weights (e.g., swing weight for system reliability) assessed from senior NSP Electric executives
• Results show for eachportfolio how fundingimpacts value added
164c 2003, 2004 Lee Merkhofer Consulting
System software computes the costs and benefits of all possible combinations of funding levels
Total (5-yr) cost (millions of dollars)10 20 30 40 50 60 70
Total value added
0
20
40
60
80
100
120
140
160
180
200
0
Efficient frontier of optimal allocations
*Hypothetical data
Each dot represents a specified funding level (case) for each portfolio
165c 2003, 2004 Lee Merkhofer Consulting
A quality process was used to collect inputs, interpret outputs, and improve the system
• Scoping sessions• One person in Asset
Management leads each portfolio team
• 3 meetings with all 50 participants
• 30-page instructions booklet
• Documented funding cases
• Very rigorous scoring sessions!
• Lesson learned
166c 2003, 2004 Lee Merkhofer Consulting
Outputs include:
• Benefit vs cost curves for each portfolio
• Total benefit vs total cost for the operating budget
• Table showing priority order for incrementing each portfolio’s funding level
• Sensitivity analysis
• Incremental benefits for each objective
• Out-year funding requirements for each department manager
• Database containing activities to be conducted and benefits anticipated for each portfolio for all funding cases
167c 2003, 2004 Lee Merkhofer Consulting
Sample Output: Benefit vs. cost for Line Clearance
0.000
0.010
0.020
0.030
0.040
0.050
0.060
0.070
0.080
0.090
0.100
Case 1 - Case 4 Case 4 - Case 5 Case 5 - Case 6 Case 6 - Case 9
Case Increments
Ben
efit
Impact on Costs to Others Future Profits Accurate & Timely BillingSystem Reliability Outage Communication Meet Commitments & Build RelationshipsPlatform for Success
2.40
2.45
2.50
2.55
2.60
2.65
2.70
2.75
$70,000 $80,000 $90,000 $100,000 $110,000 $120,000 $130,
Cost
Cum
ulat
ive
Util
ity
168c 2003, 2004 Lee Merkhofer Consulting
Total benefit vs. total cost: Optimal allocations (based on 5-year benefits & 5-year costs)
Benefit to Total Cost Curve for Optimized Allocation
0.00
1.00
2.00
3.00
4.00
5.00
6.00
$950,000 $1,000,000 $1,050,000 $1,100,000 $1,150,000 $1,200,000 $1,250,000 $1,300,000
Total Discounted 5-Year Cost ($000s)
Cum
ulat
ive
Util
ity $275,000 in 2001
169c 2003, 2004 Lee Merkhofer Consulting
Year 1 funding steps (based on 5-year B/C optimization) for a $275 million budget
Funding level
Portfolio Abbreviation Cases Incr. Utility
Incr. 2001 Costs
($000s)
Cumulative Costs for 2001
($000s) 1 SOS Case 5 ('99 Actuals) 0.17 462$ 218,502$ 2 C&I Case 2, Case 3 0.55 3,422$ 221,924$ 3 B&P Case 3 0.29 3,046$ 224,970$ 4 C&I Case 5 ('99 Actuals) 0.19 1,505$ 226,475$ 5 SOS Case 6 0.35 688$ 227,163$ 6 SOS Case 7 0.29 750$ 227,913$ 7 Dist Sust Case 3, Case 3.5 0.28 3,453$ 231,366$ 8 EP&S Case 2 - Protect, Case 4-Begin Growth, Case 5 ('99 Actuals) 0.41 7,234$ 238,600$ 9 R&O Case 5 ('99 Actuals) 0.05 622$ 239,222$ 10 B&P Case 5 ('99 Actuals) 0.16 2,556$ 241,778$ 11 R&O Case 5.1 RTS+X1, Case5.2 RTS+GIS1+X1, Case 6 GIS2+RTS+X2 0.78 2,695$ 244,473$ 12 Met Read Case 3 - Limited AES rollout 0.06 2,430$ 246,903$ 13 Met Serv Case 2 0.08 2,248$ 249,151$ 14 Met Read Case 5 - Scheduled AES rollout 0.05 2,400$ 251,551$ 15 Trans Case 5 ('99 Actuals) 0.23 5,398$ 256,949$ 16 Trans Case 6 0.14 1,254$ 258,203$ 17 Dist Sust Case 5 0.10 3,394$ 261,597$ 18 Met Serv Case 4 0.01 642$ 262,239$ 19 LC Case 4 0.07 10,600$ 272,839$ 20 Met Read Case 7 0.07 1,305$ 274,144$ 21 LC Case 5 ('99 Actuals) 0.09 8,003$ 282,147$ 22 LC Case 6 0.05 1,700$ 283,847$ 23 EP&S Case 6-Prepare Compete, Case 9-Full strategy 0.25 5,794$ 289,641$ 24 R&O Case 7 GIS3+TRBL1 0.06 1,718$ 291,359$ 25 Met Serv Case 5 ('99 Actuals), Case 9 0.03 2,192$ 293,551$ 26 Met Read Case 9 - Growth(Data Dream) 0.05 1,873$ 295,424$ 27 SOS Case 9 0.13 4,900$ 300,324$ 28 B&P Case 7 0.06 2,384$ 302,708$ 29 R&O Case 9 TRBL2 0.09 4,262$ 306,970$ 30 B&P Case 9 0.03 1,265$ 308,235$ 31 Dist Sust Case 6.9 0.04 1,900$ 310,135$
170c 2003, 2004 Lee Merkhofer Consulting
What the system did wellNSP managers & experts saw significant (and unanticipated) benefits
Improvedunderstanding
• “I gained an appreciation of the impact on others. If I get something, someone else doesn’t.”
• “I gained an understanding of consequences beyond just economics.”
• “We made improvements in getting information up to management.”
(1 of 4 slides)
171c 2003, 2004 Lee Merkhofer Consulting
Improved accountability and responsibility
because managers must document what
they expect to accomplish
• “We had people digging into costs, finding out what builds cost structure.”
• “We gained an understanding, tying costs to benefits and to impacts of not spending.”
• “Building funding cases promotes a better understanding of what it takes to run your business.”
(2 of 4 slides)
172c 2003, 2004 Lee Merkhofer Consulting
Improved performance due
to clearer objectives
• Managers redefined and shifted dollars among activities to get higher performance
• “People acknowledged that there are alternatives”
• “In developing funding cases, we got a chance to talk with colleagues in our own portfolio about what excellence would be”
(3 of 4 slides)
173c 2003, 2004 Lee Merkhofer Consulting
Provided a framework for evaluating funding
requests that is fair, consistent, and
promotes a complete and thorough
consideration of relevant issues
• “It is a place where there is dialogue that is internally consistent.”
• “It worked! We made the tradeoffs to meet our budget.”
• “Areas could go off to do their own pieces and they would still fit together.”
• “Initially it seemed too overwhelming and complex. But, when we got into it, it worked.”
(4 of 4 slides)
174c 2003, 2004 Lee Merkhofer Consulting
In interviews after 2000’s application, NSP managers & experts saw needs that must be addressed to improve applications
Need for more and better dataNeed for controls to prevent
gaming
Need to better link portfolios with department budgets and
capital budgets
• “Bulk transmission doesn’t have adequate reliability drivers. The current ones are for measuring load serving reliability.”
• “[Distribution Maintenance] cases were too big and too generic. If low-value work gets lumped with big value dollars, they go together.”
• “We are trying to do two things – portfolios and departmental budgets – that don’t sync up.”
• “Capital & Operating allocations are separate processes. We have no way to compare preventive maintenance and a capital expense.”
175c 2003, 2004 Lee Merkhofer Consulting
Why was system canceled rather than modified as in previous years?
• NSP merged to be come Xcel Energy– Xcel management wanted their own management system– They wanted a system that relies on analysis of historical data and
statistical risk and less on professional judgment of those closest to problem
– Focus was on merger activities
176c 2003, 2004 Lee Merkhofer Consulting
Conditions for success of any portfolio optimization tool
• Senior management must understand and sponsor it
• Department heads must be willing to give up some autonomy for good of the whole, trusting that tool is fair
– Company culture is important!
• Scope must be clear
2
Problem DefinitionThe company does not currently quantitativelyevaluate and compare all distribution projects. (A formal, repeatable, and uniform approach for valuing projects does not currently exist.)The value of doing a particular project is not compared with the values of competing projects. For the projects that are evaluated, the company is not satisfied with the current procedures.
3
Scope of Project Prioritization Problem
Large number of projectsMultiple performance measuresProjects done for different reasonsAnalysis of uncertaintyRisk of deferralRespond to budget signals
4
Characteristics of Project Prioritization System
Level playing field for all projectsResolve differences of opinion rationally
Techniques for resolving differences of opinion and determining which differences matter
Defensible logic for peer reviewTransparent analysisCompleteness with respect to performance measures
Multiple performance measures for multiple objectives
Bias- and error-free
5
Characteristics of P2 System -continued
Practically applicable with respect to time and costCompatible with existing business practicesExplicit treatment of uncertaintyAbility to quantify what is lost from insufficient fundingSoftware to manage and compare large numbers of diverse activities – client/server database (Oracle, SQL Server)
6
P2 Value Measurement System
The system is:Multi-year Multi-attribute Value driven
Three key dimensionsObjectives of the project portfolio
minimizing or maximizing important, measurable aspects of system performance
Values capture relative importance of competing
objectives.
Project attributes describe how each project contributes to
attainment of objectives
System Performance
System Performance
Safety PowerQuality Reliability
Net Revenue Environment
7
Required I/O + Transformations
INPUTS
•Corporate budgets
•Projects + Alternatives
•Objectives
•Values
•Attributes
OUTPUTS
•Project Rankings
•Portfolio of projects
•Timing of projects
•Value of additional budget
•Value
•Risks
•Costs
TRANSFORMATION PROCESS
•Attribute + values + objectives
Benefits
•Projects + Alternatives
Budget requirements
•Benefits + budget req’ts
Portfolio
•∆ Budgets
∆ Portfolio
8
Overview of System Structure
Objective Specification
Value Specification
Attributes Specification
Project Definition
Project Analysis
Portfolio Design
Budget Specification
Portfolio
9
Natural Units – measure system performance
Power QualityReliabilityCosts / RevenueEnvironmentalSafety
10
Scales – measure value of change
Linear DollarsNo. CustomersEtc.
Non-linearFrequencyDurationNo MomentariesEtc.
0
0.2
0.4
0.6
0.8
1
$0 $10,000 $20,000 $30,000 $40,000 $50,000
Net Revenue
Scal
ed V
alue
0
0.2
0.4
0.6
0.8
1
0 5 10 15 20
Frequency of Sustained OutagesSc
aled
Val
ue
11
Weights – measure relative value of change
Project Value
Power Quality 75 Reliability 100 Safety 100 Financial 70Environment 90
Momentaries 70
Residential 75
Commercial 100
Industrial 100
Sustained 100
Residential 75
Commercial 100
Industrial 100
Employee 100Residential 50
Commercial 100
Industrial 100
Emergency Days 100
Normal Days 100
Public 100
Low Density 40
Medium Density 75
High Density 100
12
Attribute Structure
At a high level, the attributes are the same
The differ in how the high level attributes are measured – Lets add details
Project Value
Power Quality Reliability Safety FinancialEnvironment
13
Attribute Structure - DetailedProject Score
Power Quality Reliability Safety Change in Revenue
Problem - In One Month Momentaries
Problem - In > One Month
Sustained
Residential
Commercial
Industrial
Overload Non-Overload
Residential
Commercial
Industrial
Residential
Commercial
Industrial
Repair and Replace Strategy for Aging Distribution Systems Assets
Charles D. Feinstein and Peter A. MorrisVMN Group
andSteve Chapel
S. Chapel Associates
May 2004
2
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure-single asset• Control of Asset Population - Cable Case Study• Organizational Issues
3
The Aging Asset Problem• Given
– An asset type (e.g., transformers, cables, poles, etc.)– A set of asset characteristics
AgeConditionFailure modesUncertainties in future performanceObservables and UnobservablesCosts
– A set of alternativesRepairReplaceRebuildRefurbishTestMaintain
• What should we do, when, and under what conditions?
4
Problem Statement
• The current system contains (x miles of cable, y power transformers, z breakers, …). These assets are aging and may present a risk of failing in groups at the same time. The company is not satisfied with its current replacement policy.
• Objective: develop a least-cost strategy for repair/replacement of these assets
• Specify a forecast of the expenses by category associated with this strategy
5
Aspects of the Aging Asset Problem• Optimal management of a single asset • Optimal policy for entire asset population• Cash flows for repair/replace for entire asset
population• Role of diagnostic tests
6
Two Fundamental Problems
• Optimal maintenance and replacement policy– Varies by asset class– Based on age, performance,
and condition information for individual assets
• Cash flow planning– Least cost replacement of
infrastructure inventory– Long term financial planning– Policy based on
maintenance and replacement policy for individual assets
Age
P(Fa
ilure
)
Cash Flow Forecast Versus Replace Policy
0.0
10.0
20.0
30.0
40.0
50.0
2000 2005 2010 2015 2020
Year
Ann
ual R
epla
ce C
ost
($00
0,00
0)
Replace at Age 40 Replace at Age 35 Replace at Age 30
7
Optimal Management of a Single Asset• Repair / replace strategy• Diagnostic tests• Individual performance—hazard function• State of asset
- Observable- Unobservable
• Solution: state-dependent optimal control
8
Diagnostic Tests• Several tests exist for each asset• Example: Underground Cable Tests
- Partial Discharge- Time Domain Reflectometry- Isothermal Relaxation Current- Dissipation Factor- Dielectric Spectroscopy- Wafer Test
9
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure—single asset• Control of Asset Population - Cable Case Study• Organizational Issues
10
Substation Transformer Example
• Decisions– Repair (under what conditions?)– Replace– Maintain– Evaluative screening interval– Number of mobile, backup substations– Number of spares– Transfer load
11
Substation Transformer Example
• Economic Variables– Number of customers– Type of customers
• Residential• Commercial• Industrial/Critical
– Cost of Action• Maintain• Rebuild• Replace
12
Substation Transformer Example
• Observable States– Age– Peak load– Oil condition (result of chemical test)– Oil Temperature (result of temperature test)
• Observable states are “Decision Contingent”– Policies are contingent on observable states– Critical to model decision flexibility
13
Substation Transformer Example• Unobservable States reflect expert knowledge & problem structure
– Hot Spot TemperatureFunction of temperature and peak loadCalculable from current and load
– Oil condition (result of chemical test)– Oil Temperature (result of temperature test) – Degree of Global Embrittlement (the entire transformer)
Condition of insulationInfluenced by oil temp (temp can “half or double insulation life”)
– Degree of Local Embrittlement (around hot spot)Influenced by hot spot temperature
– Effective AgeInfluenced by local and global embrittlement
– Company-specific deviation from Industry Bathtub Curve
14
Relationship among observable and unobservable states
OilCondition
PeakLoad
BathtubCurve
EffectiveAge
Failure
Decision
OilTempHot Spot
Temp
GlobalEmbrtmntLocal
EmbrtmntAge
15
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure—single asset• Control of Asset Population - Cable Case Study• Organizational Issues
16
Inference Example: Why we need Learning Model
• Uncertain Event: “Terrorist attack in USA in next week”
– You believe the probability is about 1%– Simulation: Toss coin ten times, observe zero or one
head
17
Inference Example: Why we need Learning Model
• Uncertain Event: “Terrorist attack in USA in next week”You believe the probability is about 1%
• Now you receive information: FBI/CIA Intelligence Report2 possible results: Threat “Credible evidence of attack ”
No Threat “ No credible evidence ”
18
Inference Example: Why we need Learning Model
• Uncertain Event: “Terrorist attack in CA in next week”You believe the probability is about 1%
• Now you receive information: FBI/CIA Intelligence Report2 possible results: Threat “Credible evidence of attack ”
No Threat “ No credible evidence ”
• Suppose you believe the agencies to be highly reliable: Given Attack 99% chance agencies pick up ThreatGiven No Attack 99% chance agencies say No Threat
19
Why we need Learning Model
Probability calculation:
• Given “Threat” Probability of Attack = ?
– (Or: How many tosses and how many heads?)
• Given “No Threat” Probability of Attack = ?
– (Or: How many tosses and how many heads?)
20
Why we need Learning Model
Simple probability calculation:
• Given “Threat” Probability of Attack = 0.50 (!!)
– One coin, one head
• Given “No Threat” Probability of Attack = .0001
– 17 coins, zero or one head
• Very strong empirical evidence that,“Humans are not very good at information processors”
21
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure-single asset• Control of Asset Population - Cable Case Study• Organizational Issues
22
Aging Asset Model Structure-Single Asset
• Two model forms to fit problem characteristics• Planning Period
– Uncertainties and decisions may change year to year– Have detailed knowledge about near-term issues– One-of-a-kind events may be on the horizon– Stochastic dynamic program yields Time- and State-
dependent policy
• Post-Planning Period– Have less detailed year-to-year knowledge – Still need to capture long term policy– Markov decision process yields State-dependent policy
23
Model linkages
Near Term Long TermTime-varying Model Steady State Model
Age: 16
Condition: Poor
Last Temp: Hi
Age: 14
Condition: Good
Last Temp: Low
Each terminal node of the time-varying model plugs into a state
of the steady-state model
24
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure-single asset• Control of Asset Population – Cable Case Study• Organizational Issues
25
Optimal Population Management –Underground Cable Case Study
• Compare optimal policy (replacement interval plus allowable failure history)
and• Non-replacement (continued repair)• Variation by type• Variation by capital cost, o&m, customer values
26
Population Management Model Logic
Read InputParameters
Create CableInventory Matrix Set Year to 1 Age Inventory
Replace CableBased on Retire
Policy & Fail Policy
Determine CableFailures
Calculate CapitalReplace Costs
Calculate O&MRepair Costs
Calculate CustomerOutage Costs
Increment ForecastYear
If Done
No
Report Results Yes
27
Population Hazard Function -- Example
• A1: Failures:2AWG 1/0 2/0 4/0 500 750
394 152 61 24 * 28• A2: 85% of failures occur in cable installed prior to
1983 (non-tree retardant)• A3: 16% of failed segments experience additional
failure in the same year* 500 assumed to fail as 750
28
Hazard Function (cont’d)
• A4: Two functional forms
hss
T
hss (1+α(t-T))post -1983
T-δ
hss
hss (1+ 2α(t-(T-δ)))
pre-1984
t
t
29
Sample Results Summary – Optimal Management of Cable Inventory
.0150+ Years, 3 Fails
750 New
.2930 Years, 3 Fails750 Old
.0250+ Years, 3 Fails
#2 New
1.040 Years, 2 Fails#2 Old
PV CostsOptimal PolicyCable Type
30
Sample Results - #2 Cable Old
Cable #2- Old
0.0
5,000.0
10,000.0
15,000.0
20,000.0
25,000.0
30,000.0
35,000.0
40,000.0
45,000.0
50,000.0
20 25 30 35 40 45 50 55 60
Policy - Replace Age
Num
ber
Miles Replaced Segment Failures
Cable #2- Old
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
20 30 40 50 60
Replace Age
PV C
osts
Replace Costs ($000) Repair Costs($000)
Customer Costs($000) Total Costs($000)
31
Sample Results – #2 Cable New
Cabke #2 New
0.0
200.0
400.0
600.0
800.0
1,000.0
1,200.0
20 25 30 35 40 45 50 55 60
Policy - Replace Age
Num
ber
Miles Replaced Segment Failures
Cable #2- New
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
20 30 40 50 60
Replace Age
PV C
osts
Replace Costs ($000) Utility Failure Costs ($000)
Customer Failure Costs ($000) Total Costs($000)
32
Sample Results - 750 Cable Old
Cable 750 Old
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
800.0
900.0
20 25 30 35 40 45 50 55 60
Policy - Replace Age
Num
ber
Miles Replaced Segment Failures
Cable 750 Old
0.0
0.2
0.4
0.6
0.8
1.0
1.2
20 30 40 50 60
Replace Age
PV C
osts
Replace Costs ($000) Repair Costs($000)
Customer Costs($000) Total Costs($000)
33
Sample Results - 750 Cable New
Cable 750 New
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
180.0
200.0
20 25 30 35 40 45 50 55 60
Policy - Replace Age
Num
ber
Miles Replaced Segment Failures
Cable 750 New
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
20 30 40 50 60
Replace Age
PV C
osts
Replace Costs ($000) Repair Costs($000)
Customer Costs($000) Total Costs($000)
34
Sample Results - Total Costs - Continued Repair & Optimal Replace – #2 Old
Never Replace Versus Optimal Policy - #2 Old
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022
Year
Tota
l Cos
ts
Never Replace Optimal Policy
35
Sample Results - Total Costs - Continued Repair & Optimal Replace – #2 New
Never Replace Versus Optimal Policy - #2 New
0.50
0.70
0.90
1.10
1.30
1.50
1.70
1.90
2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022
Year
Tota
l Cos
ts
Never Replace Optimal Policy
36
Sample Results - Total Costs - Continued Repair & Optimal Replace – 750 Old
Never Replace Versus Optimal Policy - 750 Old
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022
Year
Tota
l Cos
ts (0
00)
Never Replace Optimal Policy
37
Sample Results - Total Costs - Continued Repair & Optimal Replace – 750 New
Never Replace Versus Optimal Policy - 750 New
0.95
0.97
0.99
1.01
1.03
1.05
1.07
1.09
2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022
Year
Tota
l Cos
ts
Never Replace Optimal Policy
38
Outline
• The Aging Asset problem• Substation transformer example—single asset• The need for a Learning Model – group exercise• Aging Asset model structure-single asset• Control of Asset Population – Underground Cable Case
Study• Organizational Issues
39
Organizational Issues—What is aging asset management?
• What is the objective?• Process or analysis?• Analytic requirements
– Data issues• Failure rates• Equipment characteristics• Customer needs and values• Utility costs• Constraints (budgets, operating, …?)
– Commitment to analysis
2
Background
♦ Air Breakers: – installed in the PSEG 138—230—500 kV transmission system
– 25 or 26 air breakers installed in the part of the system under study
– breakers are all approximately 30 years old. None has failed yet.
♦ Two failure modes
– catastrophic mode: pressurized porcelain tube blow up and shatters, scattering parts up to 700 feet away
– non-catastrophic mode: the breaker can develop air leaks (in either 500psi or 2000psi systems) which will induce a compressor failure and causes and interruption. The cause of this failure mode is worn parts, such as aging gaskets. The critical time for this failure mode is dependent on asset age, and appears to be in the range 8-10 years old.
3
Background cont’d♦ The impact of a failure on customers varies
♦ The current maintenance cycle
– yearly inspection and maintenance
– special procedures every three, four, six, and nine years
– field rebuild every nine years
4
Decision Problem
♦ The problem: PSEG does not have a repair / replace policy for air breakers– when to replace air breakers, – when to rebuild them – when to maintain – how best to treat air breakers that exhibit poor performance
♦Objectives: – Develop repair replace policy old air breakers that minimizes life
cycle costs including customer and utility costs– Evaluate the current policy of rebuilding air breakers on a nine-year
cycle
5
Decision Problem - cont.
♦ Decision - what to do with an old breaker?
– maintain the breaker without rebuilding it
– rebuild the breaker in place, which costs approximately $150,000
– refurbish the breaker in a shop
– purchase a new breaker, which costs approximately $300,000.
♦ Uncertainties (and relevant parameters describing the uncertainties):
– Probability of failure of new and old breakers as function of age
– Probability of failure as a function of time since rebuild
– Probability of failure conditional on a breaker performance - good or problem
6
Analysis - Approach
♦A multi-stage decision tree was used to evaluate and identify the least-cost repair / replace strategy.
♦Under this approach the user specifies all decisions, uncertainties and outcomes
♦ The method identifies the set of decisions over time that meet the objective
7
Analysis - Inputs
♦ Types of technology – Old breaker – New breaker
♦ Policy decisions: maintain, rebuild, replace with new
♦ Technology performance states:– Good– Problem– Failure– C-failure
♦ Decision stages: 8 stages, each stage is 3 years in length– Make a decision on a breaker then revisit every three years– Model behavior for 24 years
8
Inputs con’d - Probability Tables Old Tech. Performance
Technology Performance Probabilities (Old Breakers, Prior Performance = Good)
Technology Performance Probabilities(Old Breakers, Prior Performance = Problem)
Perf. State 1-3 4-6 7-9 10-12 13-15 16-18Good 0.8 0.75 0.6 0.4 0.4 0.4Problem 0.15 0.15 0.2 0.3 0.3 0.3Failure 0.05 0.05 0.15 0.2 0.2 0.2C-failure 0 0.05 0.05 0.1 0.1 0.1
Breaker Age (Decision Stage)
Perf. State 1-3 4-6 7-9 10-12 13-15 16-18Good NA 0 0 0 0 0Problem NA 0.7 0.65 0.5 0.5 0.5Failure NA 0.25 0.3 0.4 0.4 0.4C-failure NA 0.05 0.05 0.1 0.1 0.1
Breaker Age (Decision Stage)
9
Inputs con’d - Probability Tables New Tech. Performance
Technology Performance Probabilities (New Breakers, Prior Performance = Good)
Technology Performance Probabilities(New Breakers, Prior Performance = Problem)
Perf. State 1-3 4-6 7-9 10-12 13-15 16-18Good 0.95 0.99 0.99 0.99 0.99 0.95Problem 0.05 0.01 0.01 0.01 0.01 0.05Failure 0 0 0 0 0 0C-failure 0 0 0 0 0 0
Perf. State 1-3 4-6 7-9 10-12 13-15 16-18Good NA 0 0 0 0 0Problem NA 0.95 0.95 0.95 0.9 0.9Failure NA 0.05 0.05 0.05 0.1 0.1C-failure NA 0 0 0 0 0
10
Inputs cont’d - CostsTechnology Decision Costs
Technology Performance Costs
Decision Old Breakers New BreakersMaintain 50 2Rebuild 150 + O&M NAReplace 300 + O&M 300 + O&M
Technology
Performance Old Breakers New BreakersGood O&M O&MProblem 1 - 5 1 - 5Failure 10 - 50 2 - 50C-Failure 300 - 500 300
Technology
1
Strategic Reliability Analysis
Charles D. Feinstein & Peter A. MorrisVMN Group LLC
AndSteve Chapel
S.Chapel Associates
May 2004
2
Outline
Analytical Modeling of ReliabilityStrategic FocusBeyond Distributions on Averages
The Importance of Rare but Catastrophic EventsPitfalls of Historical Analysis and the need for Scenario ThinkingThe Strategic Reliability Problem
3
Analytic Reliability Modeling
Classic reliability models don’t address the full range of customer outcomes
Analytic models based on classic reliability theory estimate reliability indices, which measure system averages Simulation models simulate variations in system averages
Simulations of average measures like SAIFI greatly underestimate the range of outcomes for customers Simulation’s classic problems: “Simulation is not enumerative and may overlook rare and important events”
4
Analytic Reliability Modeling (cont’d)
Current models are not strategically focused They do not focus on the rare but catastrophic events of most concern to utilities and the publicThey do not represent the true range of uncertaintyThey do not optimize strategy
5
Analytic Reliability Modeling (cont’d)
Distributions on system averages greatly understate the range of outcomes
Mean outage duration = 4, σ = 5SAIDI for 100 customers: µ = 4, σ = 0.5
-0.10
0.10.20.30.40.50.60.70.80.9
1
0 2 4 6 8
Hours per Outage
Distribution of MeanOutage DurationDistribution of OutageDuration
6
The Importance of Rare but Catastrophic Events
Probabilities of rare events can drive strategyInventory problems are an example
Costs are relatively flat for range of policies under normal operationsOptimal policy based on costs and frequency of stockout or service interruption, a rare event
7
The Importance of Rare but Catastrophic Events (cont’d)
Policy consequences of probabilities of rare events—inventory exampleOptimal policy can change greatly as frequency changesIndicates what is important to know (what is true frequency?)
OptimalInventoryLevel
Interruption Rate (Time between interruptions)
1 2 3 4 5 6 7 8 9 10 11 12 13
8
Pitfalls of Historical Analysis
Statistical analysis alone is inappropriate for predicting rare, one-of-a-kind eventsModels must be augmented with expert judgmentProcedures exist for encoding expert judgment into usable model inputs
9
Recent events highlight the importance of scenario thinking
1900 1920 1930 1940 1960 1980 2000
History of Terrorism in the US
Victims ofTerrorism
1000
2000
3000
4000
5000
10
The Strategic Reliability Problem
Catastrophic events result from two fundamental causes
Added stress on subsystem components associated with failure of a subset of components in the the same subsystemLinkages among subsystems permit failure consequences to cascade—complex systems can fail in unanticipated ways
11
The Strategic Reliability Problem (cont’d)
Data is insufficient to characterize catastrophic event ratesScenario planning is requiredDetailed models of subsystem behavior do not address the strategic questionsPolicy failure is a failure of the imagination
12
The Strategic Reliability Problem (cont’d)
Fundamental Modeling IssuesCommon Mode FailuresCascading Failures (the avalanche)Complexity and Dependency
13
Common Mode Failures
MULTI-CIRCUIT PROBABILISTIC RELIABILITY ANALYSIS Two Circuits, each described by a Markov Process
Circuit A up down Circuit B up down
up 0.990 0.010 up 0.980 0.020
down 0.600 0.400 down 0.300 0.700
Specify Special Transition Definition Specify Common-Cause Failure Probability
From To 0.0
Model ResultsPlan Period 60 Years
Steady State Annual Frequency of Catastrophes 0.001Number Probability
0 0.9568718Steady State Average Years Between Catastrophes 976 Years 1 0.0302407
2 0.00904793 0.00269894 0.00080275 0.00023816 0.00007047 0.00002088 0.00000619 0.0000018
10 0.0000005
WHEN THE CIRCUITS ARE INDEPENDENT, THE TIME BETWEEN FAILURES IS ALMOST 1000 YEARS
Probability Distribution
A down B downany state
14
Common Mode Failures (cont’d)
MULTI-CIRCUIT PROBABILISTIC RELIABILITY ANALYSIS Two Circuits, each described by a Markov Process
Circuit A up down Circuit B up down
up 0.990 0.010 up 0.980 0.020
down 0.600 0.400 down 0.300 0.700
Specify Special Transition Definition Specify Common-Cause Failure Probability
From To 0.3
Model ResultsPlan Period 60 Years
Steady State Annual Frequency of Catastrophes 0.035Number Probability
0 0.3745Steady State Average Years Between Catastrophes 29 Years 1 0.1675
2 0.13083 0.09844 0.07185 0.05116 0.03567 0.02438 0.01639 0.0108
10 0.0070
WHEN THE CIRCUITS ARE NOT INDEPENDENT, THE TIME BETWEEN FAILURES IS ENORMOUSLY REDUCED
COMMON-CAUSE POSSIBILITIES ARE FAR MORE CRITICAL TO SYSTEM FAILURE THAN THE RELIABILITY OF THE SYSTEM COMPONENTS
Probability Distribution
A down B downany state
15
Cascading Failures (the avalanche)
Avalanche Model -- An Example of Independent Failures, No Interaction
Number of Linked Assets 10 degree of system complexity (number of connected rocks)
Asset 1 Asset 2 Asset 3 Asset 4 Asset 5 Asset 6 Asset 7 Asset 8 Asset 9 Asset 10Age 20 20 20 20 20 20 20 20 20 20
Independent Failure Rate 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050incidence rates (size of "rocks")
201 Failure 2 Failures 3 Failures 4 Failures 5 Failures 6 Failures 7 Failures 8 Failures 9 Failures 10 Failures
Trigger Rate for Future Events 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000degree of dependency (strength of avalanche)
0.1001.0
24
Catastrophe = 3 or more Failures disaster management flexibility (avalanche warning system)
SystemFailures Probability
0 0.60781 0.25052 0.09513 0.03284 0.01025 0.00286 0.00067 0.00018 0.00009 0.000010 0.0000
Total 1.0000
Catastrophe Rate 0.0465
RESULTS
SYSTEM RISK CHART
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 1 2 3 4 5 6 7 8 9 10
NUMBER OF FAILED ASSETS
PRO
BAB
ILIT
Y
16
Cascading Failures (the avalanche) (cont’d)
Avalanche Model -- An Example of Interrelated Failures
Number of Linked Assets 10degree of system complexity (number of connected rocks)
Asset 1 Asset 2 Asset 3 Asset 4 Asset 5 Asset 6 Asset 7 Asset 8 Asset 9 Asset 10Age 20 20 20 20 20 20 20 20 20 20
Independent Failure Rate 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050 0.050incidence rates (size of "rocks")
201 Failure 2 Failures 3 Failures 4 Failures 5 Failures 6 Failures 7 Failures 8 Failures 9 Failures 10 Failures
Trigger Rate for Future Events 0.100 0.200 0.400 0.800 1.000 1.000 1.000 1.000 1.000 0.000degree of dependency (strength of avalanche)
0.1001.0
24
Catastrophe = 3 or more Failures disaster management flexibility (avalanche warning system)
SystemFailures Probability
0 0.60781 0.09712 0.03323 0.00524 0.00005 0.00006 0.00007 0.00008 0.00009 0.000010 0.2567
Total 1.0000
Catastrophe Rate 0.2619
RESULTS
SYSTEM RISK CHART
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 1 2 3 4 5 6 7 8 9 10
NUMBER OF FAILED ASSETS
PRO
BA
BIL
ITY
17
Complexity and Dependency
Strategic Design Issue: Complexity versus Dependency
Number of Independent Assets ClustersNumber of Links in ClustersStrength of Links Within Clusters
Example: 10 Mildly Linked Assets,
10 Clusters 5 Clusters 2 Clusters 1 Cluster0 Failures 0.608 0.608 0.608 0.608
1 Failure 0.251 0.117 0.109 0.097
2 Failures 0.095 0.210 0.039 0.033
3 Failures 0.033 0.031 0.008 0.005
4 Failures 0.010 0.028 0.001 0.000
5 Failures 0.003 0.003 0.196 0.000
6 Failures 0.001 0.002 0.018 0.000
7 Failures 0.000 0.000 0.005 0.000
8 Failures 0.000 0.000 0.001 0.000
9 Failures 0.000 0.000 0.000 0.000
10 Failures 0.000 0.000 0.016 0.257
Total 1.000 1.000 1.000 1.000
Catastrophe 0.04654 0.06493 0.24433 0.26188
18
The Strategic Reliability Problem (cont’d):Current state of the art is out of balance
Great emphasis on solving tactical problems, supported by lots of data and detailed modelsLittle emphasis on solving strategic problems, very little data, no accepted modeling approaches (that we are aware of), so we are developing:
Strategic ParadigmsAnalytic ModelsKey Data and Judgments
Dominant costs and customer effects are on the strategic side (latest estimate of cost of NE blackout is in the neighborhood of $10 billion!!)Tactical Effort > Strategic Effort, but
Tactical Importance < (<?)Strategic Importance
19
The Strategic Reliability Problem (cont’d)
Simplified versions of some of our models are available in a software demonstration
Using Analytical Tools to Improve Asset Management for T&D
SUMMARY
Steve Chapel, S. Chapel Associates
Lee Merkhofer,Lee Merkhofer Consulting
Charles D. Feinstein and Peter A. MorrisVMN Group LLC
May 2004
2
Summary
What is the T&D asset management problem?What are the specific asset management decisions that companies face?What are the useful analytic tools?What data is required to apply tools and make decisions?What are the organization issues?
3
What Is The T&D Asset Management Problem?
Specific T&D Aspects?Maintaining the existing infrastructureExpanding system to meet future needsMeeting customer needs Minimizing costsMaking money
Analytical Aspects?Defining objectivesMeasuring valueAssessing uncertainty and riskChoosing and appropriate time horizonForecasting
5
What Are The Useful Analytic Tools?
For repair / replace?For prioritization?For capacity decisions?For setting budgets?For forecasting cash flows?
6
What Data Is Required To Apply Tools And Make Decisions?
For repair / replace?For prioritization?For capacity decisions?For setting budgets?For forecasting cash flows?
7
What Are The Organization Issues?
Link between specific decisions and corporate objectives / strategyCorporate appreciation of specific customer and engineering issues Incorporating multiple stakeholders valuesRole of process and role of analysis
8
Some advice for those who use analytic tools
Use a tool with an “enterprise-level” view of asset management.Choose a tool capable of capturing all critical considerations. Commonly ignored considerations include:
soft benefits investment urgency (as opposed to investment value) risk (especially “correlated risks”)sequencing and other project interdependencies
9
Some advice for those who use analytic tools (continued)
Don’t use tools that have no basis in accepted mathematical theories. If in doubt, get an independent review.Plan on the need for education and training. Having a thorough understanding of concepts is as important as knowing how to use the tool.Use a phased approach to implementation. Start with one department, conduct lessons-learned review, and make changes before you move to the next level.
10
Some advice for those who use analytic tools (continued)
Make sure that all stakeholders have necessary “buy in” and confidence—otherwise, insufficient effort will be devoted to generating inputs and/or model recommendations won’t change decisions.If the answer seems wrong, don’t trust it (and get a new model!). A tool is an aid (not a substitute) for decision making.
11
Revisiting Asset Management Beliefs – Fact or Fallacy?
Problems can be solved by organizational change and asset management teams"Cost Benefit analysis, the "tried and true" method for ranking projects is perfectly reasonable way to select projects to fund."The first important step is to gather dataAll projects can be valued using the same aggregate measures (i.e. $)Projects are risky because of uncertain financial consequences Beta is an appropriate way to measure project risk
12
Fact or Fallacy cont.
An important objective is near-term profitability The biggest problem for asset management is insufficient data. ROI is the good metric for evaluating investments. Ranking is a good way to prioritize projects. Strategic alignment is a good metric for prioritizing projects. Balanced scorecards are a good way to prioritize projects. Hurdle rates are a good way to account for risks.
Using Analytical Tools to Improve Asset Management for T&D
SOME FINAL THOUGHTS
Steve Chapel, S. Chapel Associates
Lee Merkhofer,Lee Merkhofer Consulting
Charles D. Feinstein and Peter A. MorrisVMN Group LLC
May 2004
2
What Next?
What are some of the things you might think about doing?Problem assessment: do you have any of these problems?Time assessment: if you own a problem, when do you need a solution?Method assessment: if you need a solution, do you have a method for finding one?
3
We believe…
This EUCI Seminar is not a sales pitchWe attempted to teach and explain principles and methods These are challenging problems both to formulate and to solveMost utilities have a difficult time solving these problems by themselvesIf you have the problem, we can help you solve it
4
What Are Possible Ways to Proceed?
In-house seminarsProblem assessmentMethods and practice assessmentSoftware demosTest CasesWorkshopsCoachingConsulting Projects