Behind-the-meter PV and
Storage Adoption
Brittany Farrell, Ph.D.
Energy Systems Researcher
Clean Power Research
6/14/2023
©2023 Clean Power Research, LLC. The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Daily (January)
Average household load (kW)
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0
12
24
Hour of the day
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Seasonally
1.4
1.4
1.2
1.2
Average household load (kW)
Average household load (kW)
Daily (January)
1.0
0.8
0.6
0.4
0.2
0.0
Jan
Apr
Jul
Oct
1.0
0.8
0.6
0.4
0.2
0.0
0
12
24
0
12
24
Hour of the day
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
Big picture: understand load
Spatially | Temporally
Seasonally
1.4
1.2
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Over many years
6000
Jan
Apr
Jul
Oct
1.0
Total system peak day load (MW)
1.4
Average household load (kW)
Average household load (kW)
Daily (January)
0.8
0.6
0.4
0.2
0.0
0
12
24
0
12
5000
4000
3000
2020
2025
2030
2035
2040
2000
1000
24
0
0
12
24
Hour of the day
Images from NYSEG/RGE PV Hosting Capacity maps:
https://www.arcgis.com/apps/webappviewer/index.html?id=84de299296d649808f5a149e16f2d87c
The information contained in this presentation is confidential.
The effects of solar and storage
Individual home load effects
Home
6.0
4.0
Load (kWh)
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
-12.0
0
12
24
Hour of the day
The information contained in this presentation is confidential.
The effects of solar and storage
Individual home load effects
Home
Solar
6.0
4.0
Load (kWh)
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
-12.0
0
12
24
0
12
24
Hour of the day
The information contained in this presentation is confidential.
The effects of solar and storage
Individual home load effects
Home
Solar
Storage
6.0
4.0
Load (kWh)
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
-12.0
0
12
24
0
12
24
0
12
24
Hour of the day
The information contained in this presentation is confidential.
The effects of solar and storage
Individual home load effects
Home
Solar
Storage
Net
6.0
4.0
Load (kWh)
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
-12.0
0
12
24
0
12
24
0
12
24
0
12
24
Hour of the day
The information contained in this presentation is confidential.
The effects of solar and storage
500
450
6000
Model
Actual
5000
400
350
Peak day load (MW)
Thousands
Cumulative Residential PV Adopters
Aggregate load effects: peak day system load over time with adoption
300
250
200
150
100
4000
3000
2000
1000
50
0
2000
0
2020
2040
Year
2060
0
12
Hour
24
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
The information contained in this presentation is confidential.
The effects of solar and storage
500
450
6000
Model
Actual
5000
400
350
Peak day load (MW)
Thousands
Cumulative Residential PV Adopters
Aggregate load effects: peak day system load over time with adoption
300
250
200
150
100
4000
3000
2000
1000
50
0
2000
0
2020
2040
Year
2060
0
12
Hour
24
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
The information contained in this presentation is confidential.
The effects of solar and storage
500
450
6000
Model
Actual
5000
400
350
Peak day load (MW)
Thousands
Cumulative Residential PV Adopters
Aggregate load effects: peak day system load over time with adoption
300
250
200
150
100
4000
3000
2000
1000
50
0
2000
0
2020
2040
Year
2060
0
12
Hour
24
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
The information contained in this presentation is confidential.
Historical Adoption
• How many buildings already have
solar/storage?
• How many buildings could have
solar/storage?
• How fast will buildings adopt
solar/storage?
• How will those buildings be
distributed?
• How does that solar/storage affect
the local grid?
The information contained in this presentation is confidential.
Adoption Potential
• How many buildings already have
solar/storage?
• How many buildings could have
solar/storage?
• How fast will buildings adopt
solar/storage?
• How will those buildings be
distributed?
• How does that solar/storage affect
the local grid?
The information contained in this presentation is confidential.
Adoption Rate
• How many buildings already have
solar/storage?
• How many buildings could have
solar/storage?
• How fast will buildings adopt
solar/storage?
• How will those buildings be
distributed?
• How does that solar/storage affect
the local grid?
The information contained in this presentation is confidential.
Adoption Location
• How many buildings already have
solar/storage?
• How many buildings could have
solar/storage?
• How fast will buildings adopt
solar/storage?
• How will those buildings be
distributed?
• How does that solar/storage affect
the local grid?
The information contained in this presentation is confidential.
Adoption Load
Impact
• How many buildings already have
solar/storage?
• How many buildings could have
solar/storage?
• How fast will buildings adopt
solar/storage?
• How will those buildings be
distributed?
• How does that solar/storage affect
the local grid?
The information contained in this presentation is confidential.
Historical Adoption
Historical Adoption
Why do we need historical adoption data?
• Train the adoption model
• Rule out future adoption for premises
that have already adopted
• Model current grid conditions
The information contained in this presentation is confidential.
Historical Adoption
What data is needed?
Data type
Adoption
modeling
Load
modeling
System location (coordinates,
address)
✓
✓
Installation date
✓
System size (kW)
✓
✓
Array configuration
(tilt, azimuth)
✓
Equipment specifications
(inverters, modules)
✓
External effects (shading,
soiling)
✓
The information contained in this presentation is confidential.
Historical Adoption
Where does the data come from?
PowerClerk®
Synchronize your PowerClerk
interconnection program
Utility interconnection
records
Custom Import
Send Clean Power Research a
custom file format containing
required information​
The information contained in this presentation is confidential.
Historical Adoption
What if the data doesn’t contain enough detail?
Building footprints
CPR can match addresses to
building footprints to infer
some specifications
Spec Inference
Infer battery and PV adoption
and PV system specifications
using NET load data (AMI data)
The information contained in this presentation is confidential.
Historical Adoption
What if the data doesn’t contain enough detail?
Building footprints
CPR can match addresses to
building footprints to infer
some specifications
Spec Inference
Infer battery and PV adoption
and PV system specifications
using NET load data (AMI data)
The information contained in this presentation is confidential.
Adoption Potential
Adoption potential
What is the upper bound of adoption?
Thousands of Cumulative Residential PV
Adopters
600
500
400
300
200
Model
Actual
100
0
2010
2030
Year
2050
The information contained in this presentation is confidential.
Adoption potential
Detailed solar resource potential study process overview
3d LiDAR Data
3d Building Models
Gridded Elevation, Tilt,
and Azimuth
SolarAnywhere Typical GHI
Year Weather Data
Shading Analysis and
Hourly PV Output
Simulation for Every
Roof Surface
The information contained in this presentation is confidential.
Adoption potential
• Solar resource potential can be used to
determine an upper bound of power from
rooftop systems
• For adoption modeling, we also use the
idea of customers with “solar access” to
set the upper bound
• Data from utilities about businesses
and households
• Census data
• Owner-occupied single-family
homes often makes a good
proxy
The information contained in this presentation is confidential.
Adoption Rate
Flex Forecast: Behind-the-Meter PV Adoption Modeling
Provides utility planners with
the ability to build forecasts
based on dozens of scenarios
Utilizes CPR’s trusted
SolarAnywhere® irradiance
modeling and state of the art
solar rates database
Solar + SPSS Adoption Saturation Profile
Combines historical
adoption, technological cost,
and socioeconomic data to
model adoption
Developed in partnership
with SMUD and deployed for
Eversource Massachusetts
Connects with Integral
Analytics’ LoadSEER to
forecast adoption load
impacts into the future
400
350
Thousands of Customers
Top-down adoption
modeling at the service
territory level (optional
geographic breakdown)
300
Cumulative Solar Adopters
Cumulative SPSS Adopters
Non-adopters
No Solar Access
250
200
150
100
50
0
2021
2028
2035
Year
2042
2049
The information contained in this presentation is confidential.
Solar and Solar + Storage adoption model
Electric
rate
Solar
generation
Electric
load
Bill
calculations
PV
prices
Available
incentives
•
The logistic regression model is based on
historical adoption and results in a probabilistic
forecast of solar adopters based on a number of
predictive factors which can include:
• Market penetration
• Total system cost
• Incentives
• Payback
• Social and economic data
•
Solar + Storage adopters are then determined as
a subset of solar adopters based on an
attachment rate
System cost
Payback
Social and Economic data
Historic
adoption data
Adoption
model
Adoption Scenarios
The information contained in this presentation is confidential.
Scenarios
Running scenarios
• Scenarios are defined using an Excel
template
• Scenarios can be run on demand
through PowerClerk
Common scenarios:
• Solar or storage pricing
• Incentive introduction or
retirement
• Rate pricing or structure
Scenario results
• Typically takes a few minutes per
scenario
• Results available in Excel and CSV
formats with special formatting options
available for integration with Integral
Analytics LoadSEER
changes
• Solar tariff changes
• Customer interest in DERs
• Valuation of reliability
The information contained in this presentation is confidential.
Flex Forecast
Scenario
Comparisons
Solar installed cost
Thousands of Estimated Cumulative Adopters
Cumulative Solar Adoption by Cost Scenarios
90
80
70
60
Decreased Solar Cost
Base Solar Cost
Increased Solar Cost
50
40
30
20
10
0
Year
Modeled PV Adoption Rates for Residential Customers in the Boston Area
The information contained in this presentation is confidential.
Adoption Location
Optional Forecast
Geographic Breakdown
Flex Forecast
Results
Demographic Data by Geographic Area
Density, Housing, Land Cover, and
Economic Characteristics
Annual Predicted
Adopters by
Geographic Area
The information contained in this presentation is confidential.
Adoption Load
Impact
Bulk PV fleet power
estimations
•
•
Model individual PV systems
or aggregate systems to the
circuit, feeder, or substation
level
System specific weather data
that can be time correlated to
real net loads back to 1998,
or used with TGY data
Can also provide real time
and forecasted production
Site 1
Site 2
8
AC Power (kW)
•
10
6
4
2
0
Aggregated AC Power (kW)
10
AC Power (kW)
Adoption
Load
Impact
Per System AC Power (kW)
8
6
4
2
0
Date
The information contained in this presentation is confidential.
Adoption
Load
Impact
Eversource MA data flow
Load Forecasting
(LoadSEER)
Flexible DER
Adoption Modules
Utility Function
PowerClerk
Analytics
PowerClerk DER
System
Specifications
Integral Analytics
LoadSEER
Distribution &
Resource Planning
(LoadSEER)
PowerClerk
DER System
Specifications
Accurate PV Power
Simulation
System Operation
(DERMS, ADMS, EMS)
The information contained in this presentation is confidential.
Brittany Farrell | brittanyf@cleanpower.com
The information contained in this presentation is confidential.
Clean Power Research: Advancing the Energy Transformation
25 years working towards a clean-powered planet
Expertise
Secure, enterprise-grade cloud software
Focus: Renewable energy, DERs / EVs, Solar
data & intelligence
40+ patents
Partnered with Dr Perez @ SUNY Albany
Industries Served
70+ Electric Utilities & Energy Agencies
•
Munis/IOUs/Co-ops
100+ Solar Industry Partners
•
Independent engineers, Solar financiers,
operators, installers; Utility planners
Trusted, adaptable and efficient cloud software solutions
Utility Industry’s Leading
Business Process Automation
Software for DERs and Beyond
Team
100+ employees
• HQ: Bellevue, WA
• Research: Napa, CA
• Satellites: NY & MA
20+ people with advanced degrees:
•
Meteorology/Atmospheric
Science/Engineering/Environment/Resources/Business
Cloud-Based Services | Web Browser & API Enabled | Secure with High Availability
Customer Engagement and
Adoption Modeling Software
for Utilities
Bankable Solar Data and Intelligence
for Financial Planning, Operations &
Forecasting
The information contained in this presentation is confidential.