Sandia’s Systems Approach to Photovoltaic Reliability PV ...energy.sandia.gov › wp-content ›...

Sandia’s Systems Approach to Photovoltaic Reliability

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company,for the United States Department of Energy’s National Nuclear Security Administration

under contract DE-AC04-94AL85000.

PV Systems Integrator WorkshopClarion Hotel, San Jose

Wednesday, March 31 – Thursday April 1, 2010

Michael Quintana and Jennifer Granata*Sandia National Laboratories

Presentation Outline

• Why System Reliability?• Some of the basics• Increase detail and complexity• Summary

Diverse Users/Stakeholders Pose Challenges

How do you align these?

Some Considerations

• Reliability is the probability of simultaneously satisfying:– The performance requirement– In a specified environment– At a particular time

• Reliability takes into account service life (expected lifetime)

• Quality is defined as creating a product suitable for the intended purpose, and doing it consistently.

• Failure definition can be inconsistent--user dependent--causing difficult to attain reliability objectives

• System: all components used to convert sunlight to electricity and deliver it to the grid in a usable form; adhering to all safety and grid quality requirements.

• Identify system reliability requirements; apply these to development of components as early as possible

• Comprehensive reliability plan requires:– Data– Methodologies– Tools– Models

Sandia’s Systems Approach to PV System Reliability

Why is a Systems Approach Needed?• Reliable components alone do not deliver reliable power

• Reliable systems can deliver reliable power in well understoodapplications and environments

• Supply chain is getting increasingly diverse; e.g. CEP

• Increasingly sophisticated components/systems; e.g. CPV

• Designs, materials, and technologies are getting increasingly diverse

• Increasingly numerous stake-holders making system reliability a much more complex target

– Utilities

– States

– System owners

– PPA brokers

– Underwriters

– Etc.

Complex PV Systems Deployed in Complex Markets

Sandia’s PV Program Takes a Comprehensive Approach

Non-Technical Blocks

User Satisfaction

& Performance Perception

Maintenance

Revenues & Data for Energy Markets

CO2Avoidance &

Green Certificates

Data Format and

TransmissionData Pooling

Data Precision &

Limiting Uncertainty

Metrology

Cross Cutting Issues

Diagnostic Tools –

Systems (incl. system parameters -

PR, yield, etc.)


Components

Tend

ency

of r

isin

g co

mpl

exity

Minimum Monitoring Specification

Technical Blocks

Building Integration /

Building Energy

Balances

Specific changes and additions for Concentration / Tracking

Additional Generator /

Hybrid Systems

StorageInverterArray

Power Quality

Multi-User Systems

ClimateIssues of Remote

Monitoring& Web Based

Systems

Advanced Climatic

Inputs (GIS, satellite,

etc.)


User Satisfaction


Maintenance


CO2Avoidance &

Green Certificates


User Satisfaction


Maintenance


CO2Avoidance &

Green Certificates

Data Format and


Data Precision &


Metrology




PR, yield, etc.)


Components

Data Format and


Data Precision &


MetrologyData

Precision & Limiting

Uncertainty

Metrology




PR, yield, etc.)


Components

Tend

ency

of r

isin

g co

mpl

exity

Tend

ency

of r

isin

g co

mpl

exity


Technical Blocks


Building Energy

Balances



Hybrid Systems


Power Quality

Multi-User Systems



Systems

Advanced Climatic


etc.)


Technical Blocks


Building Energy

Balances



Hybrid Systems


Power Quality

Multi-User Systems



Systems

Advanced Climatic


etc.)

PPA’s

Financing

• Reliability– Predictive Modeling– Real-Time Studies– Accelerated Testing and Failure Modes Effects and

Analysis– Technology Transfer and Codes & Standards

• System Modeling and Analysis• System Grid Integration• Test & Evaluation (System, Inverter, BOS, Modules)• Market Transformation

Sandia Program Elements

Adapt Common Tools --Focus on Adoption--Minimize Cost


• Why System Reliability?• Some of the basics

Define system and its functional operation; define components and their functional operationDefine failure modes

• Increase detail and complexity

Boundary Diagrams

Codes & Standards

FMEA’s

Baseline Performance

RBD’s

Real-time Data

Long-term exposure

Failure Analysis

Mitigation

Accelerated Tests

Predictive Modeling

PV System Reliability Looking Out

Climatic Environment

Grid

Physical site

Arrays

Inverter

DC BOS

AC BOSSystem Monitor

Resource

PV System Reliability Looking In

We divide factors into three categories:

– Performance is the primary factor

– Economics enters the picture as we consider the cost of reliability measures vs. cost of unreliability

– Social factors are driven primarily by bureaucratic and/or aesthetics factors

Factors that influence how reliability is addressed

Environment

DC BOSArray

Inverter

Sub-Structure

AC BOS

Data AcquisitionSystem

Tracker Controller and Drive

Distribution grid

Met System

Environmental Interactions

We use Boundary Diagrams to Define Types of Inter-Relationships that Affect Reliability

Environment

Array

Sub-Structure


Physical/Mechanical Interactions

DC BOSArray

Inverter

AC BOS

Distribution grid

Electrical Interactions

EnvironmentSub-Structure

DC BOSArray

Inverter

AC BOS

Data AcquisitionSystem


Distribution grid

Met System

Data, Control, and Communication Interactions

Develop Generic FMEAs

(Optional) (Xfmea)

Program A

Develop Program-Specific FMEAs(Xfmea)

Develop Program-Specific FMEAs(Xfmea)

Program B

Develop &ExecuteFMEA

Strategic Plan

Develop &Execute FMEA Resource Plan

ManagementReview

Test & Field

Failures

EFFECTIVE FMEA PROCESS

1B

2 3

3

9

41A

10

8

Linkage to Other

Processes

Execute Actionsto Reduce/

Eliminate RiskFMEA Quality

Audits

5

SupplierFMEAs

6

7

Planning Stage ImplementationStage

Performing FMEAs Stage Review Stage

Integrated Software Support

2.1 - Doesn't open

1.1 - Doesn't disconnect

1.1.1 - Dirty AC power

1.2.1 - Inverter stopsoperating

1.1.1 - Dirty AC power

1.1 - Allow high voltagespike through

1.4 - Arcing (maybeshould be moved to

lower level)

1.3 - Short Circuit

1.2 - Intermittentcontact

1.1 - Open circuit

2.1.1 - Fire, electricalfailure

1.1.1 - Personnel safetyconcern

1.1 - Safety - directcontact with grid

3.33.34 - grid sensors

1.2 - Catestrophic failure

1.1 - Loss of outputpower

3.32.38 - Inductors

3.32.37 - Capacitors

1.1.1 - damage to othercomponents

1.1.2 - blow fuse

1.4.1 - Heating

1.4.2 - Damage toelectrical components

1.3.1 - Trip BranchBreaker

1.3.2 - Heating (highimpedance short).Melt conductor.

1.2.1 - Temporary powerinterruption

1.1.1 - No output power

3.39 - Fuse block

3.36 - Grid isolationswitch (automatic)

3.35 - Disconnect(isolation) switch

3.33 - Control circuitry

3.32 - Filter components

3.31 - Surge suppressioin

3 - AC output

1.1 - Open circuit

1.2 - Open circuit

1.1 - Short Circuit

1.2.1.2 - broken wire

1.2.1.1 - loose screw

1.1.1.4 - Corrosion

1.1.1.3 - failed solderjoint

1.1.1.2 - broken wire

1.1.1.1 - loose screw

3.30 - Connectionterminal block

Cause-effect diagram shows possible failure modes

Inverter FMEA

Cau

seEf

fect

Failure

System Component

CauseEffectFailureSystem

c-Si FMEA

OxS (occurrence X severity) plots (pareto) help to focus attention on critical aspects of the system

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 250

22

2

4

6

8

10

12

14

16

18

2020 20

15

12 12 12 12

10 10

9 9 9

8 8 8 8 8 8 8 8 8 8 8 8

6

Causes Ranked by Initial Occ x Sev (1 - 25)

Cause

Initi

al O

cc x

Sev

Initia

Project: Crystalline Silicon Device

Causes Ranked by Initial Occ x Sev (1 - 25)1: Oi x Si = 20 (4 x 5) - corrosion (Item: 14 - Frame)2: Oi x Si = 20 (4 x 5) - improper installation (wrong metals, poor processes)

(Item: 14 - Frame)3: Oi x Si = 15 (5 x 3) - One or more cracked cells (Item: 3 - Cell Strings)4: Oi x Si = 12 (4 x 3) - Increased series resistance due to solder joint

degradation & or failure at gridline interface (Item: 3 - Cell Strings)5: Oi x Si = 12 (4 x 3) - Fatigue due to thermal cycling (Item: 3 - Cell

Strings)6: Oi x Si = 12 (3 x 4) - improper use / installation (Item: 9 - Junction Box)7: Oi x Si = 12 (3 x 4) - improper installation (Item: 14 - Frame)8: Oi x Si = 10 (2 x 5) - Cracked cell (Item: 3.7 - Solar Cell)9: Oi x Si = 10 (2 x 5) - Solder bond failure (Item: 3 - Cell Strings)10: Oi x Si = 9 (3 x 3) - Decreased power in a single cell (Item: 3 - Cell

Strings)11: Oi x Si = 9 (3 x 3) - open circuit (Item: 9.10 - Bypass Diodes)12: Oi x Si = 9 (3 x 3) - Delamination from glass (loss of optical coupling)

(Item: 2 - EVA (Front))13: Oi x Si = 8 (2 x 4) - moisture uptake by EVA (Item: 2 - EVA (Front))

1 2 3 4 5 6 7 8 9.................................................................................................................................................................................0

22

2

4

6

8

10

12

14

16

18

202020

15

12121212

1010

9 9 9

8 8 8 8 8 8 8 8 8 8 8 8

6

5 5 5 5 5 5

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

3 3 3 3 3 3 3

2 2 2 2 2

1

Causes Ranked by Initial Occ x Sev (1 - 68)

Cause

Init

ial O

cc x

Sev

Initia


• Why System Reliability?• Some of the basics• Increase detail and complexity

Develop data resources Develop methodologiesShare information/data, tools and processes

• Summary

Codes & Standards

FMEA’s

Baseline Performance

RBD’s

Real-time Data

Long-term exposure

Failure Analysis

Mitigation

Accelerated Tests

Predictive Modeling

Some Current Activities/Recommendations

1. Develop a reliability model of the system2. Use reliability block diagrams (RBD’s) to

describe relationships (series, parallel, etc)3. Collect component and subsystem field data4. Collect component and subsystem lab based

(ALT) data5. Adjust data to fit environmental/operational

constraints and conditions (natural and man-made environment)

6. Use stochastic and deterministic methodologies to predict reliability

7. Verify and/or adjust prediction through field data

• Database used as a repository for field data: PVROM

• Standardized method for collecting and maintaining O&M data

• Web-based: user friendly• Secure/Proprietary• Data directly exported to

predictive model tools • Initially populated with

legacy TEP data; others are adopting

• More partners are being sought

PV Reliability O&M Database

System Long Term Exposure Tests

Objective: Determine system degradation rates through controlled exposure tests in varied environments

• System configurations assure real life effects• Tests being initiated in hot/dry, hot/humid, and cold

climates• Subset of modules subjected to lab level baseline tests• Exposure tests subjected to quarterly inspections and

semi-annual performance tests• Maintain control modules indoors• Minimize measurement errors

System Level Model/RBD

PV System Level

34.5KVTransmisssion

Line

FuseDisconnect

Fuse

Fuse

Fuse

Fuse

Fuse

Fuse

Fuse

Fuse

PV GeneratingBlock 1

PVGenenerating

Block 2







Blocks have lower level reliability models

Analysis of failure data rolls up for system reliability.

Analysis of repair and maintenance data rolls up for system availability.

Fuse failure

Maintenance resources and associated costs can be tracked and critical inventory anticipated for planning purposes

PV150 Inverter

Inverter FailureEvents

Controller Interlock Design Internal MatrixSolidStaterelay

Environmental

Software

GridEffects

Other

Rodents chewing insulation

Wrong voltage limits

Lightning outage

Relay failure

Weeds too high

Insects short PWB

This type of field experience is impetusfor conducting Inverter level FMEA’s

• Model predicts for any component and any level of the system: degradation vs time--reliability vs time--availability vs time

Photovoltaic Reliability and Availability Model (PVRAM)

Component

Actual Number

of Failures

5 yr Cum

Expected Number

of Failures

5 yr Cum

Expected Number

of Failures

10 yr Cum

Expected Number

of Failures

20 yr Cum

PV 150 Inverter (26 cSi arrays) 125 132 231* 429*

PV Module 29 26 31 38 AC Disconnect 22 17 23 31

Lightning 16 10 20 41 208/480

Transformer 5 3 3 3

Row Box 34 25 35 50 Marshalling Box 2 4 7 11 480VAC/34.5KV

Xformer 5 4 5 9

• Model prediction is accurate at 5 years

• Predictions for 10 and 20 years need additional data

• Model is being exercised by running sensitivity analyses

Model Results for our initial run of PVRAM

How would system reliability change by doubling inverter lifetime?

• Real field data (limited amounts,always more data needed)• Viewing reliability of each component within a system shows

“weakest links” – opportunities for improvements and R&D efforts• Model the changes in availability or cost if improvements are made

to one of the weak links – how LCOE is affected? Do changes produce ROI?

• Allows for trade-offs: Accept O&M costs for replacing less reliable inverter vs. cost of more reliable inverter and reduced O&M

ReliaSoft BlockSim 7 - www.ReliaSoft.com

Block Reliability vs Time

Time, (t)

Rel

iabi

lity,

R(t

)=F(

t)

0 4000800 1600 2400 32000.000

1.000

0.200

0.400

0.600

0.800

Reliability

PV ArraySystem208/480 TransformerAC DisconnectPower MeterPV 150 InverterPV StringRow Box

Mike MundtSandia Labs11/19/20082:48:20 PM

Time [hours]

PV System Reliability vs Time

Expected reliability of components and the system based

on utility PV data

The system reliability is driven by the inverter reliability!

ReliaSoft BlockSim 7 - www.ReliaSoft.com

Block Reliability vs Time

Time, (t)

Rel

iabi

lity,

R(t

)=F(

t)

0 4000800 1600 2400 32000.000

1.000

0.200

0.400

0.600

0.800

Reliability

PV Array Exp Inverter I...System208/480 TransformerAC DisconnectPower MeterPV 150 InverterPV StringPV String

Mike MundtSandia Labs11/19/20085:00:18 PM

Time [hours]

PV System Reliability vs Time

What if: Double the lifetime of the inverter, screen out early

failures

The system reliability would improve drastically!

How do I improve my module design?

• Design of Experiment methods drive Accelerated Lifetime Testing• ALT allows lifetime determination; reduces uncertainty• R&D or process improvement opportunities identified• Module manufacturers benefit from module improvement, warranty

predictions• More data needed!

ReliaSoft Weibull++ 7 - www.ReliaSoft.comReliability vs Time Plot

Folio1\Data 1: β=2.9632, η=11.8702, ρ=0.9859

Time, (t)Rel

iabi

lity,

R(t

)=1-

F(t)

0.000 25.0005.000 10.000 15.000 20.0000.000

1.000

0.200

0.400

0.600

0.800

R

FWRF

MWMF

MWMF

MS12

Time [years]

Module Reliability vs Time

If all modules in one population, half fail after 10 years

Rel

iabi

lity:

Pro

babi

lity

of p

rodu

cing

> 8

0% o

f orig

inal

pow

er

Reliability of worse modules

Reliability of better modules

ReliaSoft Weibull++ 7 - www.ReliaSoft.com

Degradation vs Time

Time, (t)

Deg

rada

tion

0.000 30.0006.000 12.000 18.000 24.000100000.000

250000.000

130000.000

160000.000

190000.000

220000.000

Linea

TF1

TF2

TF3

TF4

MikeSand11/12:24

Failure defined as 80% of initial power output

Module Degradation vs Time

Deg

rada

tion:

Pow

er o

ver t

ime

Time [years]

Four module strings

R&D to improve lifetime or tighten distribution

ALT data needed: determine lifetime, difference in populations, warranty

Accomplishments: Direct Industry Support

• Teamed with module manufacturer to demonstrate Sandia’s reliability tools

• Created FMEA in a team environment; specific issues to manufacturer’s process identified

Failure Modes and Effects Analysis

Cu-Sn Tape at 60° C, 70% RHShort Time

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

0.00 5.00 10.00 15.00 20.00

Time (hrs.)

Elec

trica

l res

ista

nce

(ohm

s)

tape # 1tape # 2tape # 3tape # 4tape # 5

15mm overlap, 30mm long

1mm overlap, 30mm long

control, 30mm long

Open

Accelerated tests developed to quantify failure risk

Sorensen et al., The Effect of Metal Foil Tape Degradation on the Long Term Reliability of PV Modules, 34th IEEE PVSC

Next steps :• Long term exposure tests at array level• Baseline performance; degradation• Modules/coupons for ALT validation• Apply diagnostics as needed• Reliability model• Incorporate ALT results into RBD model

High failure

risk ID’d

Reliability Methodologies With Industry Partner—A Success Story

• Define reliability needs; some needs vary with application and customer (residential, commercial, utility), industry segment (integrators, manufacturers, financiers, etc.), technology, and stakeholder.

• Reliability database of fielded system failure modes, failure rates, degradation rates and O&M costs to be used to create predictive model(s)

– data needs to be protected from disclosure and potential misuse• Fielded system reliability and accelerated aging evaluation needed

– for predictive models and correlations between lab and field tests• Safety-related failures are high priority; risk of injury and industry

liability/reputation• Improve existing tests, increase use of best practices/methodologies for

reliability and accelerated aging tests, and expanded applications of the information derived from lab and field evaluations

Summary: Major Themes in Sandia’s Program

Sandia’s Systems Approach to Photovoltaic Reliability PV ...energy.sandia.gov › wp-content ›...

Documents

Transcript of Sandia’s Systems Approach to Photovoltaic Reliability PV ...energy.sandia.gov › wp-content ›...