Sandia’s Systems Approach to Photovoltaic Reliability PV ...energy.sandia.gov › wp-content ›...
Transcript of 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.)
Diagnostic Tools –
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.)
Non-Technical Blocks
User Satisfaction
& Performance Perception
Maintenance
Revenues & Data for Energy Markets
CO2Avoidance &
Green Certificates
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.)
Diagnostic Tools –
Components
Data Format and
TransmissionData Pooling
Data Precision &
Limiting Uncertainty
MetrologyData
Precision & Limiting
Uncertainty
Metrology
Cross Cutting Issues
Diagnostic Tools –
Systems (incl. system parameters -
PR, yield, etc.)
Diagnostic Tools –
Components
Tend
ency
of r
isin
g co
mpl
exity
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.)
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.)
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
Presentation Outline
• 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
Tracker Controller and Drive
Physical/Mechanical Interactions
DC BOSArray
Inverter
AC BOS
Distribution grid
Electrical Interactions
EnvironmentSub-Structure
DC BOSArray
Inverter
AC BOS
Data AcquisitionSystem
Tracker Controller and Drive
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
Presentation Outline
• 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
PV GeneratingBlock 3
PV GeneratingBlock 4
PV GeneratingBlock 5
PV GeneratingBlock 6
PV GeneratingBlock 7
PV GeneratingBlock 8
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