Solar PV Technologies: A Few Key Issues
CONFIDENTIALADITYA BIRLA GROUP
Amitabh Verma, Ph.D
Vice President- Technology
Solar Power Business
Outline
• Technology as profitability driver
• Importance of meteorological data
• Selection of modules for power plants
• Pros and cons of thin film modules
• Inverter choices for power plants
• Module layout options
• Concentration photovoltaic
CONFIDENTIALADITYA BIRLA GROUP
• Concentration photovoltaic
• Operation and Maintenance
1
Profitability Through Design
CONFIDENTIALADITYA BIRLA GROUP 2
Comparison of Satellite and Ground Based Data
CONFIDENTIALADITYA BIRLA GROUP
MOS is Model Output Statistics which takes into account
ground based data and satellite data
3
How Predictable is Year-To-Year Radiation Data?
CONFIDENTIALADITYA BIRLA GROUP
Brightening Dimming Brightening
Very few weather stations in the world have 40-50 years radiation data
4
Measurement Uncertainty
CONFIDENTIALADITYA BIRLA GROUP 5
3
4
5
6
7
8G
HI
(kW
h/m
2/d
ay
)
MNRE NASA Meteonorm
Global Horizontal Incidence (GHI) for Jodhpur
CONFIDENTIALADITYA BIRLA GROUP
0
1
2
Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec
Months
� NASA data is based on Satellite Imaging
� Meteonorm is based on Satellite Imaging and weather station� MNRE is based on weather station� NASA was different than MNRE and Meteonorm in ANOVA analysis
6
20
25
30
35
40
Am
bie
nt
tem
p.
(oC
)
MNRE 24 h average MNRE day time average NASA 24 h average Meteonorm 24 h average
Ambient Temperature for Jodhpur
CONFIDENTIALADITYA BIRLA GROUP
0
5
10
15
20
Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec
Am
bie
nt
tem
p.
(
Months
7
Variation of Plant Output with GHI and Ambient Temp.
CONFIDENTIALADITYA BIRLA GROUP 8
1694
1678
1709
1620
1640
1660
1680
1700
1720E
ne
rgy
Ou
tpu
t (M
Wh
)
Plant Output per MWp Using Different Weather Data
CONFIDENTIALADITYA BIRLA GROUP
1593
1520
1540
1560
1580
1600
MNRE-24 h MNRE-Day Temp NASA Meteonorm
En
erg
y O
utp
ut
(MW
h)
Weather Data
9
18.5 18.3
17.2
18.7
26.4 26
23.4
26.9
20
25
30
IRR
(%
)
Project IRR (Pre-tax) Equity IRR (Pre-tax)
Project Return
CONFIDENTIALADITYA BIRLA GROUP
0
5
10
15
MNRE-24 h MNRE-Day Temp NASA Meteonorm
IRR
(%
)
Weather data
10
Uncertainty in Yield Calculation
Parameters % Uncertainties Remarks
Year-to-Year climate variability 4%
Solar resource estimation in the
horizontal plane5%
This is due to variation in irradiance and
temperature data
Radiation in the plane of array 3%
CONFIDENTIALADITYA BIRLA GROUP
Power rating of modules 3% Variability due to manufacturing
Losses due to dirt 2%
Other sources of error 5%The losses in inverter, cable and transformers
have uncertainty
Annual Uncertainty 8.4%
Monte Carlo Simulation has to be done to calculate the probabilistic generation value
11
Inverters
Grid
Schematic of MW Solar Park
CONFIDENTIALADITYA BIRLA GROUP
AC Panel
Inverters
Combiner box with string monitoring ,
DC switch and Surge Protector
AC Switch AC Ckt BreakerStep up
Transformer
Data Monitoring
Power Socket
Sensor
12
4.00%
5.00%
6.00%
7.00%
Total Loss Pareto
CONFIDENTIALADITYA BIRLA GROUP
Near
Shading
on
Global
IAM
Factor on
Global
Non-STC
Operatio
n
Soiling
DC Array
Variabilit
y
Module
Mismatc
h
DC Cable
Loss
Inverter
Efficiency
Inverter
Limitatio
n
Inverter
Cooling
Data
Acquisiti
on & Aux
MV
Transfor
mers
AC Cable
Loss
HV
Transfor
mer
Series1 1.60% 1.90% 6.10% 3.00% 2.00% 1.50% 1.00% 1.62% 0.09% 0.41% 0.12% 1.06% 1.00% 0.47%
0.00%
1.00%
2.00%
3.00%
Total loss to the tune of 22%
13
Effect of Temperature on the Output
CONFIDENTIALADITYA BIRLA GROUP 14
Module Type Temp. Coeff.Module
Efficiency (%)Tolerance
Relative Module
Price
Mono-crystalline Silicon -0.41%/0C 14-20 3% 1.1
Multi-crystalline silicon -0.46%/0C 13-18 3% 1.0
Increasing
Area
Crystalline or Thin Film Module
CONFIDENTIALADITYA BIRLA GROUP
Cadmium Telluride Thin
Film-0.30%/0C 10-11 5% 0.90
Micromorph Silicon -0.25%/0 C 6-7 5% 0.75
Amorphous Silicon Thin
Film (Single Junction)-0.18%/0 C 5-6 5% 0.70
15
Cd-Te Module
Multi-crystalline Module
I-V Characteristic of Modules
CONFIDENTIALADITYA BIRLA GROUP
Mono-crystalline Module Amorphous Silicon Module
16
Visual Identification of Different Modules
CONFIDENTIALADITYA BIRLA GROUP
Mono-Crystalline Multi-Crystalline Amorphous Si Cadmium Telluride
17
Module Type
Module
Wattage
(Wp)
No. of
Modules
No. of
Modules in
series
No. of
Strings
Total Module
Area
Annual
Energy
(kWh)
Excess
Energy
Compared
to Multi-
crystalline Si
PR
SunPower 210 23184 18 1323 29627 8847762 0.88% 77.7
Mono-
Crystalline 210 23805 15 1587 39629 8806077 0.40% 77.4
5 MWp Plant Design with Different Modules
CONFIDENTIALADITYA BIRLA GROUP
Crystalline
silicon
210 23805 15 1587 39629 8806077 0.40% 77.4
Multi-
Crystalline
Silicon
21023805 23 1035 38025 8770840 0 77.1
Cd-Te 80 62946 9 6944 44997 8954308 2.1% 78.7
Micromorph
Silicon445 11235 3 3745 64264 9033976 3.0% 79.4
Location: Rajasthan and Central Inverter. All other parameters being same
18
� Degradation rate
� Reliability� Area available
� O&M cost
� Bankability� Installation cost
� BoS cost such as structure,
Module TypeDegradation
Rate (%/Year)Exposure Year
Mono-crystalline
Silicon0.30 8
Multi-Crystalline
Miscellaneous Considerations for Module Selection
CONFIDENTIALADITYA BIRLA GROUP
� BoS cost such as structure, cable and string monitor
� As per a study by Sandia lab
the module replacement ratefor crystalline module is 0.05%
and for thin films approx. 1%
Multi-Crystalline
Silicon0.5 9
Amorphous
Silicon0.7-1.4 7-12
Cd-Te 0.70 10
19
Around 65% of Non-Module Costs Scale with Module Efficiency
65%
35%100%
(Rs.40/Wp)
Halving efficiency
increases non-module costs by 65% and module costs must decrease correspondingly to
maintain competitiveness on a systems level
CONFIDENTIALADITYA BIRLA GROUP
systems level
Efficiency
dependent non-
module
costs
Efficiency
independent non-module
costs
Total non-
module cost
20
c-Si reference
module12.8%
Nominal module
efficiency (%)“Normalized” module
efficiency (%)Yield at
Rajasthan(kWh/kWp)
176212.8%
Thin Film Vs Crystalline Silicon
CONFIDENTIALADITYA BIRLA GROUP
Micromorph Si thin
film module 7.8% 1807 8.0%
Taking into account Rajasthan irradiation and
temperature statistics, the „normalized“ efficiency of a micromorph thin film module is 8% vs.
12.8.0% for a c-Si module
21
Entitlement of Thin Film Module Cost TakingCrystalline Silicon System Cost as Reference
Rs.120/Wp Rs.14/Wp
Rs.41.6/Wp
26x(12.8/8.0)
CONFIDENTIALADITYA BIRLA GROUP
Rs.64.4/Wp
Efficiency
dependent non-
module costs
Deduced thin
film Si module
price
Efficiency
independent
non-module
costs
Total systems
cost
Thin film module shall be20% cheaper than
Crystalline silicon moduleso as to meet the system
level cost.
22
Choice of Inverters
� Maximize efficiency and performance
� Maximize reliability and durability� Leverage scale of mass produced AC components� Easy to install� Special training not required for installation� Less risk/hazard
Inverter CTQs
+ A large number of power pointtrackers improve the yield
+ High DC system voltage range
+ Inverter integrated withmodule
+ Module level MPPT
CONFIDENTIALADITYA BIRLA GROUP
Central Inverter String Inverter Micro-inverter
and Three phase is now possible
in string inverters.
+ Reduced losses in both AC andDC cable and hence higher output
+ Dispense with junction box
+ Less skilled manpower needed
+ Inverters can be placed among the
modules+ No string monitor is needed
+ Module level MPPT
+ Module level monitoring
is possible
+ DC cabling not required+ More modular
-More expensive at present
- Bankability?
23
Central String
Less Inverters of High
Capacity
More Inverters of Less
capacity
For 1 MW:
2 No of 500kVA each
For 1 MW:
91 No of 11kVA each
++ Standard
composite Unit of
Inverter,
Transformer Station,
Switchgears as a
++ Separate Units
as Inverters, AC
Combiner Boxes,
Communication
Cables, Transformer
Comparison of Central and String for 1 MWp Plant
CONFIDENTIALADITYA BIRLA GROUP
Switchgears as a
package including
Communications
Cables, Transformer
Station, Switchgears
++ Easy
for Planning and
Installation
- - Complex
and a lot of
engineering reqd.
++ Easy
to Maintain due to less
components
++ Easy
to Maintain due to
higher reliability
24
Layout of Central Inverter Plant of 1 MWp
28 x SMA string-monitor
7 – 8 per Main Box
8 strings per combiner box8 contacts per string monitor,
Module Power: 180 Wp
26 modules per string
CONFIDENTIALADITYA BIRLA GROUP
Transformer 1000 MVA
Inverter 500 HE-11
Inverter 500 HE-11
SMA concrete
station 1000 MV
including 2 inverters
including transformer
TransferStation
2 x SMA DC main box2 per Inverter
25
Layout of String Inverter 1 MWp Plant
. . .
96 x SMC 11000TL
1 String is 20 modules
3 strings per inverter
CONFIDENTIALADITYA BIRLA GROUP
Transformer 1000 MVA
AC Combiners AC Combiners
AC Combiners for Phase
Balancing, Transformer
Station Transfer
Station
2 x SMA Web box
96 x SMC 11000TL
……………
26
Module Mounting Configuration
CONFIDENTIALADITYA BIRLA GROUP
Landscape Portrait
Which is Preferred?
27
Portrait layoutLandscape layout
Implications
CONFIDENTIALADITYA BIRLA GROUP
Shadow or Snow
Shadow or Snow
One string bypassed
Both strings bypassed
28
Concentration PV (CPV)
Salient Features of CPV� Concentration 500x� Dual axis tracking
� Uses direct radiation
� High efficiency III-V cells
Positives� Higher energy harvesting because of
• Tracking
CONFIDENTIALADITYA BIRLA GROUP
• Tracking
• Lower temp. coeff. (-0.16%/deg. C)
Negatives� Not a static system
� Reliability?
� Bankability� Cost
� Area efficiency
29
Frequency of Outages and Impact on Energy Loss
CONFIDENTIALADITYA BIRLA GROUP 31
O & M Cost (Fixed Crystalline Si as Reference)
Relative CostRelative CostRelative CostRelative Cost
O&M Activity
Fixed Crystalline
Si as the benchmark
Cd-Te Amorphous Si
Single Axis
Tracked Crystalline Si
Schedule
maintenance 1 1.2 1.4 1.5
CONFIDENTIALADITYA BIRLA GROUP
maintenance and cleaning
1 1.2 1.4 1.5
Unscheduled
maintenance1 1 1 2.5
Inverter
Maintenance1 1 1 1
Civil structure
and ground maintenance
1 1.5 2 1.5
32
Summary
• Generation over life time, capital cost and O&M are critical for economic
viability of project
• Design of Solar Power Plant is key to profitability
• Solar installation has to last for 25 years and negligence in component
selection and design may hurt in future
CONFIDENTIALADITYA BIRLA GROUP
selection and design may hurt in future
• Meteorological data are not accurately known nor can be predicted
with high certainty. Probabilistic approach to yield calculation is more prudent.
• One size does not fit all or a single design is not optimally suited for all
locations. An optimum solution for each location has to be designed.
33
CONFIDENTIALADITYA BIRLA GROUP