Conservation Voltage Reduction (CVR) Voltage Optimization (VO)

Josh RushtonRegional Technical Forum

July 21, 2015

2

Objective

Today, we’re seeking RTF decisions on two standard protocols:• Automated CVR: Recommend deactivation. – Intent is only to move to Custom Protocols– Protocol’s method is still a reasonable approach– Still a lot of CVR savings out there

• Simplified VO: Recommend approval.– Proven category, active status.– Proposal includes major protocol changes

3

Outline

• Background• Simplified VO– The big shift–Reliability–Decision

• Automated CVR

4

Background

112

114

116

118

120

122

124

126

0 2 4 6 8

Volta

ge

Miles from substation

What are these measures about?Basic idea: Some things use less energy at lower voltages

Complication 1: Average ΔV can be hard to estimate

Complication 2:

5 -- Background

Savings factors (%ΔKWh per %ΔV) depend on end-use mix.

• (PNNL, 2010) gives some lab results; • (NEEA, 2008) reflects residential mix circa 2006.

ΔV

Artist’s rendering (fake data)

Two Standard Protocols

Automated CVR Protocol #1: Collect 90 days of alternating day-on/day-off data, use to estimate feeder-specific savings factors• Early version approved by RTF in 2004• Major revision November, 2011 (latest draft dated May 15, 2012)

– Stalled out in subcommittee – Main sticking point was proposed eligibility requirements

Simplified VO Protocol: “Canned” savings factors derived from NEEA’s DEI research (NEEA, 2008)• Early version approved by RTF May, 2010, good background info in April, 2010• Revised version brought to RTF in November, 2012

– RTF sent back to subcommittee for further discussion– Subcommittee took a step back to consider a change of course

Both protocols stalled out at end of 2012 • Both currently Under Review.

6 -- Background

% Δ kWh% ΔV

Recent RTF Direction

• RTF (January, 2015): Find out what it would take to “dress up” the protocols – Clarify terms like mostly residential, review treatment of persistence,

…– Check whether eligibility can be expanded

• Subcommittee: CAT should compare requirements to DEI study/data• Result: No clear answers without major effort; probably none with major

effort

• RTF (January, 2015): Are RTF standard protocols really a good fit for these measures? – RTF Standard Protocols seem too narrow – Maybe some kind of “custom guidance” would be better– Subcommittee (May, 2015): Very limited discussion of this issue

7 - Background

8

Simplified VO Protocol

Measure Overview

• Measure Name: Voltage Optimization

• Method: Standard Protocol

• Category: Proven

• Current Status: Under review

• Proposed Status: Active

• Proposed Sunset Date: July, 2020

• Highlighted areas: Eligibility, reliability

9 – Simplified VO

Two Standard ProtocolsAutomated CVR Protocol #1: Collect 90 days of alternating day-on/day-off data, use to estimate feeder-specific savings factors• Early version approved by RTF in 2004• Major revision November, 2011 (latest draft dated May 15, 2012)

– Stalled out in subcommittee – Main sticking point was proposed eligibility requirements

Simplified VO Protocol: “Canned” VO savings factors derived from NEEA’s DEI research (NEEA, 2008)• Early version approved by RTF May, 2010, good background info in April, 2010• Revised version brought to RTF in November, 2012

– RTF sent back to subcommittee for further discussion– Subcommittee took a step back to consider a change of course

Both protocols stalled out at end of 2012 • Both currently Under Review.• Recommend deactivating Automated CVR Protocol #1• Recommend approving modified version of Simplified VO Protocol

10 – Simplified VO

The “canned” VO factors

• VO factors quantify %ΔkWh per %ΔV• Values in VO factor table based on NEEA Load

Research Project– Vary by climate, saturation of AC and ER heat – Only counts end-user energy savings (distribution losses

calculated separately)– Reflect residential end-use mix

• Tables need periodic review/update to reflect changes in end-use stocks

• May be possible to expand tables to include values for VO factor more commercial loads

11 – Simplified VO

Motivation

Why build an RTF protocol around the VO factor tables?• VO factor tables have potential to simplify

savings estimates for a lot of feeders• An RTF protocol has several functions – Describing how to get reliable estimates from

simplified methods – Maintaining reliability via sunset check-ins – Making conservation credits more predictable

12 – Simplified VO

The tricky part

VO factor (%ΔkWh / %ΔV) only useful if you can estimate ΔV• Why? Savings estimated as ΔkWh = kWhAnnual * VOf * %ΔV• What? ΔV target is average annual change in voltage

experienced by end users • How? That’s the tricky part.

– Easy to estimate ΔV for very linear systems • Performance thresholds increase linearity but restrict eligibility• See Additional Slides for details

– Voltage usually not very linear in the wild– Are reliable ΔV estimates ever possible without meeting all

thresholds? (Answer: Probably sometimes)

13 – Simplified VO

Proposed approach

• Describe the circumstances and method where we know savings is reliable

• Point out some things you need to be careful about when you deviate from that method or those circumstances

• This is a new approach to RTF Standard Protocols.

14 – Simplified VO

How the proposed protocol works

• Proposed Protocol describes Benchmark Method, same as sole Savings Method in previous protocol drafts. – Savings method (with formulas that assume linearity) – Performance thresholds (which support linearity and measure persistence)– Data collection (what’s needed for savings method inputs and reliability

checks) – Details in Additional Slides

• Proposal allows other approaches (besides Benchmark Method) but does not specify alternatives– Totally different, unforeseen, approaches– Minor tweaks (may alter savings method, performance thresholds, data

collection, or any combination)

15 – Simplified VO

Proposed Reliability Framework • Basic Eligibility Requirement: “The practitioner can reliably estimate the

average annual change in voltage that is experienced by end-users as a result of the measure”

• Reliability Standard: “…reliability similar to the Benchmark Method” • Validity Threats: “At a minimum, protocol documentation must

demonstrate that the final calculations address” threats related to four issues: Extrapolation, Annualization, Baseline, Persistence– Protocol notes how performance thresholds address threats in Benchmark

Method – Protocol does not specify what it means to address the threats in general

• Language similar to above used in both Eligibility and Delivery Verification sections

• Intent: Describe circumstances where RTF “certifies” reliability. Leave grey areas of reliability to regulators, evaluators, and utilities, but with “some lines drawn on the playing field”

16 – Simplified VO

Validity Threat: Extrapolation“Is the ΔV estimate based on a reliable approach to extrapolating data collected at the selected metering locations (e.g., source and EOL) to customers along the feeder?

“In the benchmark method, this extrapolation is based on a linear model of voltage decay along the length of each feeder, and the performance thresholds support this linear assumption.”

17 – Simplified VO

Other Validity Threats

• Annualization• Baseline• Persistence• See Additional Slides

18 – Simplified VO

Minor Changes• Clear statement of “right-on-average” reliability standard• “Mostly residential” requirement given specific definition (at

least 80% of feeder load due to residential end-users, the rest light commercial)

• Moved four-step integrated planning process to Appendix • Removed delivery verification items that only related to the

planning process (rather than reliability of savings)• Editorial stuff

19 – Simplified VO

Staff/CAT Recommendation

Recommend approving Simplified VO Protocol with proposed modifications• Status: Active• Category: Proven• Sunset date: July 31, 2020

20 – Simplified VO

Proposed Motions

“I, _________, move that the RTF approves the Standard Protocol for Voltage Optimization as presented (active status, proven category, sunset date of July 31, 2020)”

21 – Simplified VO

22

Automated CVR Protocol #1

Automated CVR Protocol #1: Collect 90 days of alternating day-on/day-off data, use to estimate feeder-specific savings factors• Early version approved by RTF in 2004• Major revision November, 2011 (latest draft dated May 15, 2012)

– Stalled out in subcommittee – Main sticking point was proposed eligibility requirements

Simplified VO Protocol: “Canned” savings factors derived from NEEA’s DEI research (NEEA, 2008)• Early version approved by RTF May 4, 2010• Revised version brought to RTF in November, 2012

– RTF sent back to subcommittee for further discussion– Subcommittee took a step back to consider a change of course

Both protocols stalled out at end of 2012 • Both currently Under Review.• Recommend deactivating Automated CVR Protocol #1• Recommend approving modified version of Simplified VO Protocol

23 – Automated CVR

Two Standard Protocols

Savings method summary

• Data collection: Minimum of 90 days, raise and lower control-zone voltage to get day-on/day-off CVR operation cycles.

• Savings factors: Primary data used to estimate feeder-specific savings factor (%ΔKWh/%ΔV) for each application– Factors capture savings on both sides of meter– Protocol directly measures switchable savings (models used

capture other savings components)• Energy savings: Savings estimated as product of savings

factor, annualized average ΔV, annual kWh (based on historical data).

24 – Automated CVR

Reasons to deactivate• Method (day-on/day-off) is a version of best-practice

– Has traction (people reference it, use as benchmark)– Little or no simplification over custom alternatives– Primary value is consistency of estimation method

• Practitioners want more flexibility– Performance thresholds are controversial– Minor tweaks (e.g., how to normalize for weather or weekends) – Very different methods (e.g., all-year-on study, w/ comparison feeders) – Existence of proven RTF Protocol not supposed to obstruct alternative methods

(but some say it does)• Staff/CAT/Subcommittee resources needed to land the protocol

– Find a path through (or around) performance thresholds– Incorporate model improvements / tweaks – Would push back other measures (for what pay-off?)– Might fail to land anyway

25 – Automated CVR

Effect of deactivation

• Eligible projects will have to go Custom (most already end up there)

• Day-on/day-off method will have no home within a proven RTF document – Doesn’t mean method isn’t reliable– Document will remain on RTF website– Practitioners/regulators may still agree to use the

approach when helpful• Reserve RTF Staff/CAT/Subcommittee resources

for other tasks

26 – Automated CVR

27

Proposed Motion

“I, _________, move that the RTF deactivates the Automated CVR Standard Protocol #1 because the protocol’s value to stakeholders is not sufficient to justify the RTF resources needed to prove out the protocol.– This decision does not imply that the protocol’s

method is unreliable;– This decision does not imply diminished savings

potential for eligible measures.”

28

Additional Slides

Simplified VO: Baseline Notes

From the proposed protocol:“Baseline. A term from the RTF Guidelines. The Baseline is the inefficient-case performance level referenced in savings calculations. Voltage optimization uses a pre-conditions baseline, which means that savings estimates should try to capture the difference in energy consumption between the pre-existing system and the efficient-case system, except when the pre-existing system includes obsolete equipment (in which case, the Baseline is the hypothetical system that would result if essential improvements were made to the pre-existing system).

• “Some measure applications include or coincide with physical system improvements that are designed to drive deeper savings or improve system reliability. These measures can also lead to more predictable system performance, which may improve reliability of the average annual voltage estimates. In such cases, it may be possible to improve savings reliability by estimating Baseline performance using meter data collected after physical improvements are completed, but using operational settings that emulate average voltage conditions of the Pre-Existing System. The Benchmark Method takes this approach.”

• “This protocol does not specify what constitutes obsolete equipment in a distribution system. A generic test is that equipment is obsolete if near-term improvements would be inevitable, even if the energy-efficiency measure was not being implemented.”

29 – Additional Slides

Simplified VO: Validity Threats (1)

Annualization“Is the ΔV estimate based on a reliable approach to annualizing data collected during the metering period?

“In the benchmark method, the meter-period voltage estimates are scaled up or down in proportion to the ratio (average annual demand)/(average meter-period demand). This kind of scaling assumes that voltage normally rises and falls roughly in proportion to demand; this assumption is supported by the performance thresholds.”

30 – Simplified VO

Simplified VO: Validity Threats (2)

Baseline“Does the ΔV estimate reflect the correct baseline? See [Baseline Notes, above].

“If the VCZ includes obsolete equipment prior to the Voltage Optimization measure, then the correct baseline is not the same as the base-case system. Instead, it is the system that would result if the obsolete equipment were replaced with components that would be typical choices in the current market.”

31 – Simplified VO

Simplified VO: Validity Threats (3)

Persistence“Can reasonable assurance be provided that the efficient-case voltage settings will persist? In all applications of this protocol, delivery verification requires that a 3-year persistence plan must be documented to ensure that efficient system operation habits become well-established. However, a persistence plan will not be followed if customers experience adverse low-voltage events during some portions of the year.”

“In the benchmark method, the performance thresholds ensure predictable and reliable system performance throughout the year so that efficient-case operations can be designed to reliably avoid low-voltage events.”

32 – Simplified VO

• Power factor (3-phase total, at source):– Minimum (hourly) greater than 0.96– Average (for week) greater than 0.98

• Phase load balance (3-phase lines, at source) – Per-unit unbalance < 0.15– Neutral < 40 amps

• Max-adjusted voltage drop (3-phase mean)– Max-adjusted drop is mean meter-period drop, times

(annual peak kW) / (mean meter-period kW)– Primary max-adjusted drop < 3.3%– Secondary max-adjusted drop < 4.0%

33 – Additional Slides

Simplified VO: Benchmark Method Performance Thresholds (1)

• Variation between feeder max voltage drops– Compare feeders within substation control zone– Must not differ by more than 2 Volts (on 120 V base)

• Primary line minimum hourly voltage– Measured near expected low voltage point – At least 114 V + (1/2) Voltage regulation bandwidth +

secondary max allowed voltage drop

• Primary line maximum hourly voltage– Measured near expected high voltage point – Less than 126 V - (1/2) Voltage regulation bandwidth

• Conductor loading– Source hourly loading (amps) less than design normal spec

34 – Additional Slides

Simplified VO: Benchmark Method Performance Thresholds (2)

Step 1. (Identify Savings Factor)Look up VOf (%ΔkWh / %ΔV) in table• Values vary by climate, saturation of AC and ER heat • Remember: VOf only counts end-user energy savings (distribution

losses calculated separately)

Step 2. (Estimate ΔV)See next slide.

Step 3. (Estimate Energy Savings)ΔkWh (savings) = kWhANNUAL * VOf * %ΔV• kWhANNUAL based on historical data• ΔV is estimated average voltage difference between CVR-on and

CVR-off cases

35 – Additional Slides

Simplified VO: Benchmark Method Savings Method

For fixed voltage reduction, VO Protocol estimates average voltage as follows, pre and post, and takes the difference:

= Regulator set point voltage setting= Hour-i metered regulator output voltage on 120 V base = Hour-i metered EOL primary voltage on 120 V base= Average annual kW demand (from measured historical data)= Average kW demand, metered at source

(Formula for line drop compensation and automated voltage feedback control adds correction for volt rise.)

36 – Additional Slides

Simplified VO: Benchmark MethodFormula for Estimating ΔV

Some Recent DE Activity• NEEA DEI Project Final Report (NEEA, 2008)

– Load Research Project (2005-2007)– Pilot Demonstration Project (c. 2005-2007)

• Distribution Efficiency Guidebook (NEEA, 2008)• Long-Term Monitoring and Tracking DE (NEEA, 2014)• Energy Smart Utility Efficiency (ESUE) Program (BPA, ongoing)• PacifiCorp DE Pilot Study• IEEE P1885. Guide for Assessing, Measuring and Verifying Volt-Var control

Optimization on Distribution Systems (Draft - Approval expected 2017)• Puget Sound Energy currently implementing VO • Avista CVR Program Impact Evaluation (Avista, 2014) • Evaluation of CVR on a National Level (PNNL, 2010)• M&V research by PNNL and WSU researchers (2014)• Green Circuits DE Case Studies (EPRI, 2011)

37 – Additional Slides