Jul 19, 2016 - federal government to improve understanding, quality, and public ... of electric power customers in each
1
The National Cost of Power Interruptions to Electricity Customers An Early Peek at LBNL’s 2016 Updated Estimate Distribution Reliability Working Group 2016 IEEE PES General Meeting Boston, Massachusetts July 19, 2016
Joseph H. Eto Lawrence Berkeley National Laboratory
[email protected] | (510) 486-7284
2
Overview/Background • In 2006, LBNL published the first peer-reviewed assessment of the national cost of power interruptions to electricity customers based entirely on publicly available data
• Found that data were sparse, poorly understood, and often misinterpreted
• Since that time, LBNL has worked closely with industry, regulators, and federal government to improve understanding, quality, and public availability of information on electricity reliability • Actively participated in IEEE DRWG to better understand industry perspectives and research aspects of Standard 1366 • Developed and promoted use of the Interruption Cost Estimate (ICE) Calculator • Reviewed OE-417 and NERC EOP-004 • Supported modifications to EIA form 861 • Provided regular, invited briefings to NARUC Electricity and Electricity Reliability Staff Subcommittees
• This presentation is a preview of LBNL’s major update to its 2006 study
3
In 2006, LBNL estimated that power interruptions cost the US $79 billion annually (2002-$) LBNL’s research was the first and remains the only peer-reviewed analysis based entirely on public data $160
Momentary Interruptions Sustained Interruptions
$135 B
$140
$119 B
$120
U.S. Total: $79 Billion
$20.4 Billion
Billions-$
Industrial 26%
Residential 2% $1.5 Billion
$90 B
$100 $80
$79 B
$60 $40
$22 B
$26 B
$23 B
$20
$56.8 Billion
$0
Commercial 72%
LBNL Base-Case Estimate
Regional Data
C&I SAIDI, SAIFI, SAIDI, SAIFI, All Outages All Outages Customers Valued Valued MAIFI MAIFI Using Experience Using Increased Decreased Fewer, by 1 Costs for a Costs for a by 1 Shorter Summer Summer Standard Standard Outages Deviation Weekday Weekend Deviation Night Afternoon
LBNL also documented significant uncertainties in its analysis, ranging from $30-130 billion Source: Hamachi LaCommare, K. and J. Eto. 2006. “Cost of Power Interruptions to Electricity Consumers in the United States.” Energy, the International Journal. 31:1509-1519.
4
A Customer-Focused Framework for Estimating the National Cost of Power Interruptions m
Cost of Power Interruptions =
n
∑∑ C i =1 j =1
i, j
× E i , j × Oi , j × Vi , j
where, C = total number of electric power customers in each region and customer class sector E = the frequency of power interruption events in one year for each region and customer class sector O = the cost per interruption as a function of outage duration by customer class for each region V = vulnerability factor m = the number of customers in each customer class n = the number of regions i,j = indices for customer class and region, respectively
5
Customer Count by Class (C) What we used for the 2004 study
What we’ve used for the current study
What we would like to use for the next study
Year 2001 customer counts by class using EIA 861
Year 2014 customer counts by class using EIA 861
Better alignment between EIA definitions (C and I) and those used in the surveys that were used to build the ICE Calculator (small/med C&I and large C&I)
Note: In 2003, EIA changed its definition of the ‘Other’ category, which led to a shift the customer count to the commercial from the industrial sector
6
Frequency and Duration of Interruptions(E) What we used for the 2004 study A convenience sample of readily available public data on SAIDI and SAIFI (with major events) and MAIFI averaged over the entire U.S.
What we’ve used for the current study 2013 and 2014 SAIDI and SAIFI (with major events) collected by EIA from hundreds of utilities weighted and averaged separately for each of the nine U.S. Census regions – SAIDI converted into distributions of event durations using data collected from utility websites
What we would like to use for the next study A statistically representative, yet enhanced SAIDI and SAIFI-like metric that distinguishes among customers by type, and provides a more explicit representation of the duration of interruptions experienced by customers
Better information on Publicly available, yet limited regional information MAIFI on MAIFI - provenance is not well understood
7
Re-expressing SAIDI as a Distribution of Outage Durations Fraction of total interruptions
0.03 Pacific West South Central New England
0.025
Mountain East North Central Middle Atlantic
West North Central East South Central South Atlantic
0.02
0.015
0.01
0.005
0
5
55
105
155 Duration (min)
205
255
8
Economic Cost of Interruptions (O) What we used for the 2004 study
What we’ve used for the current study
A meta-analysis that developed customer damage functions (CDFs) based on 24 utility customer surveys administered by 8 utilities between years 1989-2002
An updated metaanalysis that developed new CDFs based on 34 utility customer surveys administered by 10 utilities between 19892012 The CDFs are based on a more advanced, twopart regression model (= the current ICE Calculator)
What we would like to use for the next study A current, statistically representative national survey of customer interruption costs that supports regional variations Augmented by a separate cost analysis of long-duration, widespread power interruptions More careful evaluation of the cost of momentary interruptions
9
Vulnerability (V) What we used for the 2004 study In the absence of any data, we assumed all customers were vulnerable to power interruptions
What we’ve used for the current study
What we would like to use for the next study
We commissioned a study by Frost & Sullivan (2015) to develop regional market penetration estimates of stand-by generators and UPS systems by customer class
Updated and more granular information on stand-by generators and UPS Augmented by better understanding of resilience of modern electricity consuming equipment
10
Building Toward an Updated Estimate $160 $140 $120
PRELIMINARY FINDINGS NOT FINAL $106
$103 Billion 2014-$
$100 $80
$115
$92 $79
$77
$114
Monte Carlo uncertainty ± 1 std dev likely between $60-160B $110 Industrial
$60 Commercial
$40 $20 $0
Residential
Case 1: The 2004 study (2002-$)
Case 2a: Case 2b: Case 2c: Case 3a: Case 3b: Case 3c: Case 4: Case 1 + Case 2a + 2001 Case 2b + 2014 Case 2c + Taking Case 3a Taking Case 3b Taking Case 3d Conversion Customer Customers Updated CDFs + updated + Vulnerability + Distributions to 2014-$ Guesstimate SAIDI, SAIFI, MAIFI
11
Updated National Cost of Power Interruptions $110 billion (2014-$) Ind $30B 28%
Res $3B 3%
PRELIMINARY FINDINGS NOT FINAL
$40 $35
Billion 2014-$
$30 $25
res
com
ind
Com $76B 69%
Momenta ry $60B 54%
Sustained $50B 46%
$20 $15 $10 $5 $0
New Middle England Atlantic
East North Central
South East West North Atlantic South Central Central
West Mountain Pacific South Central
12
$250
A Different Perspective on the Cost of Momentary Interruptions
• Momentary PRELIMINARY FINDINGS interruptions are the NOT FINAL result of technology $211B
Billion 2014-$
$200
$150 $114B $100
[SERIES NAME]
$50
$0
[SERIES NAME] Before (Case 3c)
After
investments by utilities to prevent sustained interruptions • $60B in momentary interruption costs might otherwise be $157B in sustained interruption costs
13
A Look Behind the Numbers
PRELIMINARY FINDINGS NOT FINAL Residential Commercial
Cost per customer (2014-$/customer) Cost per MWh annual sales (2014-$/MWh)
$23 $2
$4,257 $56
Industrial $35,757 $30
Source: U.S. Energy Information Administration, Form EIA-861, "Annual Electric Power Industry Report.", Form EIA-861S, "Annual Electric Power Industry Report (Short Form)" and Form EIA-923, "Power Plant Operations Report“ for Year 2014
14
A Sensitivity Study on the Cost of Power Interruptions to Industrial Customers Industrial customers are often served at sub-transmission and transmission voltages $120.0 $100.0 $7.1B
Billion 2014-$
$80.0
PRELIMINARY FINDINGS NOT FINAL
$60.0
Commercial
$40.0 $20.0
Residential
$0.0
Initial estimate (Case 3c)
Assume: 25% of Industrial customers experience only 5% of system SAIDI/SAIFI
Assume: 75% of Industrial customers experience only 5% of system SAIDI/SAIFI
Source: Assumption of 5% of interruptions due to LOS taken from findings in: Eto, J., K. H. LaCommare, H. C. Caswell, and D. Till, Distribution vs. Loss of Supply: Identifying the Source of Electric Service Interruptions, submitted to IEEE Transactions on Power Systems, May 2016.
15
Summary of Preliminary Findings • LBNL has begun an update to its 2004 study of the national cost of power interruptions • The update is based on a number of improvements in the public information that is now available, in part due to research sponsored by DOE • LBNL now estimates – on a preliminary, not yet final basis that power interruptions cost $110 billion per year (2014-$), an increase of more than 30% since our initial, 2004 study • The ~13% of customers in the commercial and industrial classes account for more than 95% of these costs • The cost of momentary interruptions continues to account for a substantial portion of these costs
16
Concluding Remarks • Power interruptions have economic consequences for customers • Addressing these consequences is a responsibility that is shared primarily between the customer and its utility, but also in some cases the government at large • Managing sustained interruptions is a long-standing responsibility of the utility • Momentary interruptions result from utility actions to manage (i.e., avoid) sustained interruptions • Managing severe major events is a responsibility that is shared with government • Customers always have the option to secure (and pay for) higher levels of reliability
