Power Output Analysis of Distributed Photovoltaic Systems in

Power Output Analysis of Distributed

Photovoltaic Systems in San Diego County

Mohammad Jamaly, Juan L Bosch, and Jan Kleissl

Department of Mechanical and Aerospace Engineering,

University of California, San Diego

Outline

• Objective: Analyzing aggregate ramp rates of distributed PV

systems in San Diego, CA and surrounding area (SDG&E

territory which covers an area of 10,600 Km2) in 2009

– Processing power output of the PV systems

– Comparing PV power output against ground measured and

satellite-derived irradiation (irradiation is converted to

power)

– Analyzing Ramp Rates of Aggregate measured and

modeled power output of the PV systems

Outline

• Objective: Analyzing aggregate ramp rates of distributed PV

systems in San Diego, CA and surrounding area (SDG&E

territory which covers an area of 10,600 Km2) in 2009

– Processing power output of the PV systems

– Comparing PV power output against ground measured and

satellite-derived irradiation (irradiation is converted to

power)

– Analyzing Ramp Rates of Aggregate measured and

modeled power output of the PV systems

Data

• CSI Measured Power Output:

• Provided by California Solar Initiative (CSI) RD&D program at

California Public Utilities Commission

• 15-min power output of PV systems

• SolarAnywhere (SAW): Satellite-Derived Irradiation

• Provided by Clean Power Research

• Derived from GOES visible imagery

• Global horizontal irradiation (GHI) and direct normal irradiation (DNI)

at 1 km spatial and 30-min temporal resolution

• CIMIS: Measured Irradiation

• Operated by California Irrigation Management Information System

• Measured Hourly GHI (60 minute-by-minute samples are averaged and

reported for 1 hour intervals) at 5 ground weather station

Data

• CSI Measured Power Output:

• Provided by California Solar Initiative (CSI) RD&D program at

California Public Utilities Commission

• 15-min power output of PV systems

• SolarAnywhere (SAW): Satellite-Derived Irradiation

• Provided by Clean Power Research

• Derived from GOES visible imagery

• Global horizontal irradiation (GHI) and direct normal irradiation (DNI)

at 1 km spatial and 30-min temporal resolution

• CIMIS: Measured Irradiation

• Operated by California Irrigation Management Information System

• Measured Hourly GHI (60 minute-by-minute samples are averaged and

reported for 1 hour intervals) at 5 ground weather station

Map of PV Systems and CIMIS Stations

• 86 PV systems with total 6.41 MW PTC rated capacity (kWAC), a

mean size of 74.54 kWAC and median size of 6.08 kWAC were applied

• Modeled GHI of the SAW pixels which contain the PV systems are

applied

• Measured CIMIS GHI at 5 CIMIS stations are applied

Aggregate Ramp Rates

• Differences in the aggregate PV power output (normalized by

the aggregate PV capacity) for different ramp duration

intervals; 15-min through 5-hour in 15-min increments

Largest step size of normalized

aggregate PV output versus ramp

time interval

Cumulative distribution of absolute

value of 1-hour ramp rates of

normalized aggregate PV output

Aggregate Ramp Rates

• Differences in the aggregate PV power output (normalized by

the aggregate PV capacity) for different ramp duration

intervals; 15-min through 5-hour in 15-min increments

Largest step size of normalized

aggregate PV output versus ramp

time interval

Cumulative distribution of absolute

value of 1-hour ramp rates of

normalized aggregate PV output

The Day with the Largest 1-hour Ramp Rate

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for the day with the

largest 1-hour ramp rate in 2009

• The largest aggregated 1 hour ramp for

this period was 50.6% of PV capacity

and occurred from 900 to 1000 PST

GOES Satellite Images on the day with the

largest 1-hour ramp rate in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min averaged

output to annual maximum output at that

time of day (ToD)

The Day with the Largest 1-hour Ramp Rate

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for the day with the

largest 1-hour ramp rate in 2009

• The largest aggregated 1 hour ramp for

this period was 50.6% of PV capacity

and occurred from 900 to 1000 PST

GOES Satellite Images on the day with the

largest 1-hour ramp rate in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min averaged

output to annual maximum output at that

time of day (ToD)

The Day with the Largest 1-hour Ramp Rate

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for the day with the

largest 1-hour ramp rate in 2009

• The largest aggregated 1 hour ramp for

this period was 50.6% of PV capacity

and occurred from 900 to 1000 PST

GOES Satellite Images on the day with the

largest 1-hour ramp rate in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min averaged

output to annual maximum output at that

time of day (ToD)

Largest 1-hour Ramp Rates

Aggregate modeled & measured power of all 86 PV sites and Aggregate GHI of 5 CIMIS

stations for the days with the (a) second, (b) third, and (c) fourth largest 1-hour ramp rates

Histogram of the largest 1-hour ramp rates (30% and larger) of aggregate PV output

Largest 1-hour Ramp Rates

Aggregate modeled & measured power of all 86 PV sites and Aggregate GHI of 5 CIMIS

stations for the days with the (a) second, (b) third, and (c) fourth largest 1-hour ramp rates

Histogram of the largest 1-hour ramp rates (30% and larger) of aggregate PV output

A Day with Marine Layer Breakup • Occurred when cloud evaporation coincides with an increase in solar altitude

GOES Satellite Images on a day with

marine layer breakup in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min

averaged output to annual maximum

output at that time of day (ToD)

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for a day with marine

layer breakup

• The largest aggregated 1 hour ramp for

this period was 38.1% of PV capacity

and occurred from 800 to 900 PST

(Transition from marine layer overcast

status to clear condition)

A Day with Marine Layer Breakup • Occurred when cloud evaporation coincides with an increase in solar altitude

GOES Satellite Images on a day with

marine layer breakup in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min

averaged output to annual maximum

output at that time of day (ToD)

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for a day with marine

layer breakup

• The largest aggregated 1 hour ramp for

this period was 38.1% of PV capacity

and occurred from 800 to 900 PST

(Transition from marine layer overcast

status to clear condition)

A Day with Marine Layer Breakup • Occurred when cloud evaporation coincides with an increase in solar altitude

GOES Satellite Images on a day with

marine layer breakup in 2009

• The circles represent 86 PV systems

• The area of the circles is proportional to

the power rating of the PV system (the

largest system is 939 kW)

• Color bar shows ratio of 15-min

averaged output to annual maximum

output at that time of day (ToD)

Aggregate modeled & measured power of

all 86 PV sites and Aggregate GHI of 5

CIMIS stations for a day with marine

layer breakup

• The largest aggregated 1 hour ramp for

this period was 38.1% of PV capacity

and occurred from 800 to 900 PST

(Transition from marine layer overcast

status to clear condition)

Conclusions • In 2009, the largest hourly ramp was 50.6% of PTC capacity

followed by several ramps of up to 44% of PTC capacity, which

indeed cause more challenges and additional costs for the system

operator in a very high PV penetration scenario

• The SAW was able to follow the CSI power output (measured over

86 systems) typically within 6% during the four largest ramps and

also matched the timing of the ramps accurately

• CIMIS observations were not as accurate as SAW due to smaller

number and non-representative geographical distribution with

respect to the PV sites

• The largest number of ramps occurred in the spring and summer

(many of them are caused by summer marine layer breakup)

• In April-October all large ramps were morning up-ramps which is

desirable because the load also increases during those times

Conclusions • In 2009, the largest hourly ramp was 50.6% of PTC capacity

followed by several ramps of up to 44% of PTC capacity, which

indeed cause more challenges and additional costs for the system

operator in a very high PV penetration scenario

• The SAW was able to follow the CSI power output (measured over

86 systems) typically within 6% during the four largest ramps and

also matched the timing of the ramps accurately

• CIMIS observations were not as accurate as SAW due to smaller

number and non-representative geographical distribution with

respect to the PV sites

• The largest number of ramps occurred in the spring and summer

(many of them are caused by summer marine layer breakup)

• In April-October all large ramps were morning up-ramps which is

desirable because the load also increases during those times

Conclusions • In 2009, the largest hourly ramp was 50.6% of PTC capacity

followed by several ramps of up to 44% of PTC capacity, which

indeed cause more challenges and additional costs for the system

operator in a very high PV penetration scenario

• The SAW was able to follow the CSI power output (measured over

86 systems) typically within 6% during the four largest ramps and

also matched the timing of the ramps accurately

• CIMIS observations were not as accurate as SAW due to smaller

number and non-representative geographical distribution with

respect to the PV sites

• The largest number of ramps occurred in the spring and summer

(many of them are caused by summer marine layer breakup)

• In April-October all large ramps were morning up-ramps which is

desirable because the load also increases during those times

Acknowledgements

• Funding and data from the California Solar Initiative RD&D program at

California Public Utilities Commission

• Jennifer Luoma for reading in the data

• Timothy Treadwell from the California Center for Sustainable Energy for

providing the 2009 SDG&E CSI data

• Stephan Barsun (Itron) for helpful discussions on data quality control

Thank You