MLPE performance modeling

Confidential | © 2016 SunPower Corporation
Chetan Chaudhari, Tamir Lance, Gregory M. Kimball, Ben Bourne | May 2019
12th PV Performance Modeling and Monitoring Workshop (May 14-15, 2019)
MLPE Performance modeling
Module Level Power Electronics

2Confidential | © 2016 SunPower Corporation |
Yo Mr. White….
Like what is this
MLPE?
(condescendingly)
It is simple. It stands for
Module Level Power
Electronics. It is just a
switched mode power
supply. It …..

MLPE - Module Level Power Electronics
• Key benefits
– Balance Of System component for NEC 690.12 (Rapid Shutdown) compliance1
– Increased shade tolerance
– Simplified PV system design
• MLPE market growing 14-18% annually 2
• Need for explicit model to study power and energy impact
1 – NEC 690.12.c 2017
2 – The Global PV Inverter and MLPE Landscape H2 2016, S. Moskowitz, GTM 2016

What is PVMismatch?
• IV & PV curve trace calculator for PV system circuits3
3 – Mark Mikofski, Bennet Meyers, Chetan Chaudhari (2018). “PVMismatch Project: https://github.com/SunPower/PVMismatch". SunPower Corporation, Richmond, CA.

What is PVMismatch?
• IV & PV curve trace calculator for PV system circuits
• Model chain
– Cell > Cell string > Module > String > System
– Module layout

What is PVMismatch?
• Model chain
– Module layout
– Bypass diodes
• Model inputs
– Cell technology characteristics
– Irradiance (suns)
– Temperature (cell temperature)
– Bypass device configuration

How to use?

Case study
• Scenarios
– String inverter
– Power optimizers + String inverter
– Microinverters
N
Cell-level POA Irradiance map

Case study
• Scenarios
– String inverter
– Microinverters
• System configuration
– 10 modules x 6 strings
– Azimuth = South
– Tilt = 10 degree
– Location = Richmond, CA
• Shade
– Obstacle (SW)
– Inter row
N
Cell-level POA Irradiance map

Case study
N• Scenarios
– String inverter
– Microinverters
• System configuration
– 10 modules x 6 strings
– Azimuth = South
– Tilt = 10 degree
– Location = Richmond, CA
• Shade
– Obstacle (SW)
– Inter row
Come see Chetan at PVSC- 46 Tuesday June 18th
"PVOPEL: A Scalable Opto-Electrical Performance
Model of PV systems using Ray Tracing and
PVMismatch"

Microinverters
• Key considerations
– Microinverter efficiency applied at module level IV curve
• Microinverter efficiency depends on
– Power level or Imp
– Input Voltage (Vmp)
– Temperature (Tamb)
– Grid conditions
• Typically
– lower power levels => lower efficiency
– lower input voltage => lower efficiency
Voltage (V)
Power (W)
Contour plot for efficiency (microinverter)

Microinverters - method
1. Find MPP of each module for given irradiance
conditions
Available Incident
Energy

conditions
2. Find efficiency loss for that power level
3. Calculate Pmpmod[MI]
Pmp map (MI)
Available Incident
Energy

conditions
2. Find efficiency loss for that power level
3. Calculate Pmpmod[MI]
4. Sum the Pmpmod[MI] across the array to get Pmpsys
Pmp map (MI)
Available Incident
Energy

Microinverters - equations
𝑃 𝑚𝑝𝑠𝑦𝑠 = ' 𝑃 𝑚𝑝𝑚𝑜𝑑 𝑛 × 𝑀𝐼𝐸𝑓𝑓(𝑃 𝑚𝑝𝑚𝑜𝑑 𝑛 , 𝑉 𝑚𝑝𝑚𝑜𝑑 𝑛 )
5
678
Where
N : number of PV modules
Pmpsys : System Power
Pmpmod : module power at max. power point
Vmpmod : module voltage at max. power point
MIEff(power, voltage) : microinverter efficiency at
given operating point

DC optimizers
• Key considerations
– Corrects for module to module mismatch
– Optimizer efficiency applied at module level IV curve
– Inverter efficiency applied at MPPT circuit level
• Optimizer efficiency depends on
– Power level or Imp
– Duty Cycle = Iin/Iout = Imp/Iout
• Where Iout = Istring @Vmpsys
– Temperature (Tamb)
– Input Voltage (Vmp)
• Typically for power converters -
– lower power levels => lower efficiency
– lower duty cycles => lower efficiency
Optimizer
String
Inverter
Power (W)
Duty Cycle (%)
Imp (A)
Contour plots for efficiency
Voltage (V)

DC optimizers - method
1
Finding operating points of
optimizers
1. Find MPP of system level PV
curve from 1
2. Assume the corresponding
Vmpsys as string voltage for all
strings since they are in
parallel
Current(A)
Power(W)
Voltage (V)
Vmpsys

1
2
optimizers
curve from 1
2. Assume the Vmpsys as string
voltage for all strings
3. Find Iout for each string by
locating Istring from 2 for Vmpsys
(dotted line)
Current(A)
Power(W)
Voltage (V)
Vmpsys

1
2
optimizers
curve from 1
(dotted line)
4. Calculate duty cycle and
efficiency for each optimizer
and adjust module Pmpmod[OPT]
Current(A)
Power(W)
Voltage (V)

1
2
optimizers
curve from 1
(dotted line)
5. Sum the Pmpmod[OPT] to get
Pmpsys[OPT]
Current(A)
Power(W)
Voltage (V)

1
2
optimizers
curve from 1
(dotted line)
5. Sum the Pmpmod[OPT] to get
Pmpsys[OPT]
6. Apply String inverter
efficiency to get power value
Current(A)
Power(W)
Voltage (V)

DC optimizers - equations
Where
Isys : PV system current (A)
Vmpsys : Voltage at max. power point - system level (V)
Imp : module current at max. power point (A)
Pmpmod , Vmpmod , Impmod : module power, voltage and current at max. power point
OptEff(duty cycle, input current) : optimizer efficiency at given duty cycle and input current (%)
StrInvEff(power, voltage) : string inverter efficiency at given input power and voltage (%)
Psys : System Power
𝐷𝑢𝑡𝑦𝐶𝑦𝑐𝑙𝑒 𝑛 = 𝐼 𝑚𝑝 𝑛
𝐼 𝑠𝑡𝑟𝑖𝑛𝑔 𝑠𝑡𝑟
… (2) j
𝑃 𝑚𝑜𝑑 𝑛 = 𝑃 𝑚𝑝𝑚𝑜𝑑[𝑛] ∗ 𝑂𝑝𝑡𝐸𝑓𝑓(𝐷𝑢𝑡𝑦𝐶𝑦𝑙𝑒 𝑛 , 𝐼 𝑚𝑝𝑚𝑜𝑑 𝑛 ) … (3)
𝐼 𝑠𝑡𝑟𝑖𝑛𝑔 𝑠𝑡𝑟 = 𝐼 𝑠𝑦𝑠 𝑉 𝑚𝑝𝑠𝑦𝑠
… (1)
𝑃 𝑠𝑦𝑠 = (' 𝑃 𝑚𝑜𝑑 𝑛 ) × 𝑆𝑡𝑟𝐼𝑛𝑣𝐸𝑓𝑓 𝑃 𝑚𝑝𝑚𝑜𝑑 𝑛 , 𝑉 𝑚𝑝𝑚𝑜𝑑 𝑛 … (4)
5
678

Results – Hourly
Typical winter day
unshaded

Results – Hourly
SW
obstacle
• Since obstacle is South-West of the array, afternoon/evening shade is expected and
thus the MLPE energy recovery is also in those hours of the day
Typical winter day Annual energy by hour of day

Results – Monthly

Results – Monthly
• Winter months can gain up to 6% energy using MI and Optimizers
• Summer months can see up to 1.5% losses in power train efficiency with lack of shading

Results – Annual

Results – Annual
• On a shade constrained PV system in this study, using MLPE devices can gain up to 1 %
more energy

Conclusions
• Using PVMismatch, MLPE based systems can be modeled in detail for variety of topologies
of MLPEs

Conclusions
of MLPEs
• Performance gains of MLPEs vary across the year for a given shading scene and annual
energy gives the most accurate measure of their performance.

Conclusions
of MLPEs
• Performance gains of MLPEs vary across the year for a given shading scene and annual
energy gives the most accurate measure of their performance.
• On a shade constrained PV system in this study, using MLPE devices can gain up to 1 %
more energy annually

Acknowledgements
Special thanks to Mark Mikofski and Bennet Meyers for creating PVMismatch!

MLPE performance modeling

More Related Content

What's hot

Similar to MLPE performance modeling

Recently uploaded

MLPE performance modeling