BMW i ChargeForward:

PG&Es Electric Vehicle

Smart Charging Pilot

A cooperation between
Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Section 1: Project Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Section 2: System Architecture and Project Development. . . . . . . . . . . . . . . . . 8

Overview of the System Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Demand Response Technology Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
BMW Group 2nd Life Stationery Battery Storage System. . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
EV Pool Control: Back-end Automated Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
EV Pool Control: Managed Charging of a Fleet of BMW i3 EVs. . . . . . . . . . . . . . . . . . . . . . . 13
BMW i ChargeForward Smart Phone App . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Customer Enrollment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Section 3: Analysis of Charging Behavior and Driver Archetypes. . . . . . . . . 16

Driver Archetypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Comparison of Battery Electric and Range Extender Participants. . . . . . . . . . . . . . . . . . 19
Non-Home Charging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Section 4: Participation in Demand Response (DR) Events . . . . . . . . . . . . . . . 22

System Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Real Time and Day Ahead Events. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Vehicle Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Program Issues and Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Section 5: Overview of Customer Behavior Research . . . . . . . . . . . . . . . . . 30

Objectives and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
BMW i ChargeForward Program Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Motivation to Enroll and Participate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Importance of Incentives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Perception of Participation and Communication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Customer Satisfaction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Managed Charging Program Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Charging Behavior. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Charging Away from Home and Barriers to Daytime Charging . . . . . . . . . . . . . . . . . . . . 36
Comfort and Trust in Managed Charging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Thresholds and Flexibility for Managed Charging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Section 6: EV Adoption and its Potential for Further Grid Support . . . . . . . 38

Analysis of Vehicle Plug Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Section 7: BMW ChargeForward Phase 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Section 8: Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Appendix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Content Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Table of Figures
Figure 1: Project Timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 2: Overview of Project Partners and Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Figure 3: BMW i ChargeForward System Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Figure 4: Solar Production and Battery Recharge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Figure 5: Average Delay of Vehicle Pool by Event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Figure 6: BMW i ChargeForward Smart Phone App . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Figure 7: W
 eekday and Weekend Charging Demand from the BMW i3 Fleet. . . . . . . . . . . . 16

Figure 8: A
 verage kW Contribution and Vehicle Participation per Event Hour. . . . . . . . . . . . 17

Figure 9: Aggregated Power Draw by Driver Archetypes. . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Figure 10: B
 attery and Range Extender Aggregate Power Draw Comparison . . . . . . . . . . . 19

Figure 11: Alternative Charging Locations (Non-Home). . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Figure 12: Vehicle Performance from Target (100 kW). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Figure 13: D
 ay Ahead and Real Time Delivered Contribution Comparison. . . . . . . . . . . . . . 24

Figure 14: Number of Opt Outs per Event. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Figure 15: Vehicles Participating per Demand Response Event. . . . . . . . . . . . . . . . . . . . . . 26

Figure 16: Demand Response Events per Vehicle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Figure 17: Motivation to Enroll and Participate in BMW i ChargeForward. . . . . . . . . . . . . . . 32

Figure 18: L
 evels of Satisfaction with Elements of the BMW i ChargeForward Program. . . . 34

Figure 19: Influence of Renewable Energy on Managed Charging Participation. . . . . . . . . . 36

Figure 20: Forecasted EV Adoptions in PG&Es Service Territory. . . . . . . . . . . . . . . . . . . . . 38

Figure 21: Forecasted Electric Vehicles in PG&Es Service Territory. . . . . . . . . . . . . . . . . . . 39

Figure 22: California ISO Duck Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Figure 23: Weekend Percent of Vehicles Connected vs. Charging. . . . . . . . . . . . . . . . . . . . 41

Figure 24: H
 ourly BMW i ChargeForward Aggregated Load (total kWh)
Against 2020 Estimated Duck Curve Grid Needs . . . . . . . . . . . . . . . . . . . . . . . 45

Figure 25: Examples of Demand Response Event, Vehicle Pool . . . . . . . . . . . . . . . . . . . . . 48

Figure 26: E
 xample of Demand Response Event, BMW Group 2nd Life Battery System . . . 48

Figure 27: E
 xample of Demand Response Event, Combined System . . . . . . . . . . . . . . . . . 49
Executive Summary

The BMW i ChargeForward Project successfully from the California Independent System Operator.
tested the feasibility of using managed electric These grid services have the potential to result
vehicle (EV) charging as a flexible grid resource. in cost savings associated with operating and
The Project has shown the ability for electric maintaining the grid as well as owning an electric
vehicles to provide viable grid services using vehicle. For each Demand Response event, BMW
the vehicle telematics system as a basis for provided PG&E with 100kW of grid resources by
communicating grid messages to vehicles. The delaying charging for approximately 100 BMW
grid services demonstrated in this pilot included i3 vehicles in the San Francisco Bay Area and
Day Ahead and Real Time Energy, which were drawing from a BMW Group 2nd life stationary
modeled after existing proxy demand resources battery system built from reused EV batteries,
for a duration of one hour.

PAGE 4
Over the course of 18 months, from July to no customer fatigue. Results indicate
2015 to December 2016, the BMW i Charge that the electric vehicle (EV) owners have
Forward Project dispatched 209 Demand a strong interest in supporting renewable
Response events, totaling 19,500 kWh. On energy through managed charging
average 20% of the total contribution was programs. They are willing to participate
attributed to the vehicle pool and 80% from in managed charging or charge during the
the 2nd life stationary battery system. The day as long as they are not inconvenienced
amount from the vehicle share is dependent or limited in their ability to use their car.
on what time of day an event is called. If The largest barrier to day time charging
an event is called from 11 PM to 2 AM, the and managed charging is the lack of
vehicle pool contributes more significantly by workplace charging. Based on the success
increasing the share from 20% to 50% of the of the BMW i ChargeForward pilot, BMW
100 kW required. This increase in vehicle pool received a grant from the California
contribution is the result of PG&Es residential Energy Commission (CEC) to continue
EV and time-of-use rate plans which provide with a second phase of the pilot. Program
lower cost electricity prices during this time participants express an interest in having
period, thus creating an incentive for people the automakers participate and expand
to charge during these hours. their role in providing grid services; this is
seen as an added benefit to drivers who
The BMW i ChargeForward Pilot is deemed a want to optimize the positive impacts of
success both from an energy reduction and driving their electric vehicle. The second
customer satisfaction standpoint. Participants phase looks further into the grid benefits
were very satisfied with the program and were attained from greater flexible charging and
active participants in the research component more advanced charging management
as well. Based off customer research, 98% while continuing to utilize vehicle telematics
of participants indicated that they were to facilitate grid messages to the vehicle
satisfied with the program and 93% stated and driver. Overall, the size and magnitude
that they are likely to participate in a similar of this resource, and the associated cost
program in the future if offered. Since this savings, will ultimately be dependent on
program was designed to run primarily in the EV adoption and the types of grid services
background of customers lives they were that can be cost effectively offered.
able to participate at high rates and felt little

PAGE 5
1 Project Overview

The BMW i ChargeForward Project (known as and maintaining the grid as well as owning and
the Plug-in Electric Vehicle Demand Response operating a vehicle. Added grid services can
Pilot Project as described in D.12-04-045 and potentially reduce the need to increase Californias
Advice Letter 4077-E) aimed to demonstrate the electricity generation capacity and is aligned with
technical feasibility and grid value of managed the States loading order for resources, effectively
charging of electric vehicles, as a flexible and reducing energy procurement costs. The magnitude
controllable grid resource. The main goal of this of these services and their associated cost savings
project was to understand the potential of using will ultimately be dependent on EV adoption
Electric Vehicles (EV) for grid services, which can and the types of grid services that can be cost
result in cost savings associated with operating effectively offered.

FIGURE 1 Project Timeline

Pilot Launch

The RFI was released in the

third quarter of 2013 and
2013 2014 2015 2016
the RFP released one year
later in the third quarter of RFI Released RFP Released Grid Services Provided
2014. BMW was selected at
the end of 2014 and BMW
i ChargeForward officially BMW Selected
launched and began enrolling for Pilot
customers during the first half
of 2015. BMW provided grid
services from July 2015 to
December 2016.

PAGE 6
1 Project Overview

The higher cost of an EV compared to In addition to BMW, Olivine acted as the

traditional internal combustion engine (ICE) project administrator, utilizing the Olivine
vehicle is widely considered the major DER, a complete distributed energy
obstacle toward mass market EV adoption. resource management platform to manage
The strategy of this pilot was to develop the demand response program aspects of
a mechanism that partially or fully levels the the pilot. Olivinean approved Scheduling
higher costs of EVs compared to internal Coordinator (SC) with the California
combustion vehicles. The project aims to Independent System Operator (CAISO)
accomplish this goal by bundling the grid acted as the interface between PG&E and
value of an EVs demand response (DR) BMW. This included managing resource
capability over the vehicles useful life and enrollment, nominations, awards, and
beyond (as a battery 2nd life stationary grid dispatch notifications to BMW based on
storage asset). This value is paid to the EV an event schedule provided by PG&E. It is
manufacturer, who then passes on this important to note that this project did not
value to the driver. directly bid into the CAISO market; however
grid services were modeled after existing
The figure below illustrates the project CIASO proxy demand resource products.
timeline for the BMW i ChargeForward
Pilot Project. It is important to note that Olivine was also responsible for interfacing
considerable work went into this pilot prior with Whisker Labs (Whisker) to retrieve
to the launch of the program with BMW. customer-level raw meter data from their
In 2013 PG&E released a request for peel-and-stick electricity metering platform,
information (RFI) to automakers to better to perform validation, estimation and editing,
understand their interest in participating in and to calculate settlement quality meter
a grid services pilot with EVs and 2nd life data for use in event baseline, performance,
batteries. PG&E received significant interest and settlement calculations, and ultimately
in the RFI which prompted the release of a for payment of incentives directly to BMW.
request for proposal (RFP) in 2014. BMW
was selected as the vendor for this project
as a result of providing the most competitive
and comprehensive proposal.

FIGURE 2 Overview of Project Partners and Roles

The graphic provides an overview

of the partners supporting BMW Olivine PG&E
the overall project structure. (Participant) (Interface between (Program
Participant and Retail DR/ Management)
Wholesale Market)

Whisker
(Meter Data Provider)

PAGE 7
2 System Architecture & Project Development

PAGE 8
2 System Architecture & Project Development

This section includes an overview of the system by pilot participants who are PG&E residential
architecture, summary of the technology customers. Once an event is called, BMW
developed by BMW to deliver the required utilizes proprietary aggregation software to delay
demand response commitment of 100 kW, and charging of participating customers (via telematics
the customer enrollment process. This section embedded in the vehicle) in order to reduce load
also includes a description of challenges faced in on the grid. The algorithm prioritizes the reduction
project development and solutions developed to of electricity consumption from charging without
overcome these project barriers. interfering with customers mobility needs;
however drivers can opt out of event participation
at any time. To address uncontrollable fluctuations
regarding managed charging capacity, BMW
Overview of the developed a stationary battery system made up
System Architecture of eight used BMW Group batteries (100 kW/225
kWh) as backup storage to fill the gap between
The pilot requires BMW to provide 100 kilowatts available load drop from managed charging and
(kW) of capacity at any given time, regardless of the required 100 kW of DR capacity.
how many BMW i3 EVs are charging. BMW is
required to provide this capacity in the form of PG&E leverages the Whisker metering system and
either Day Ahead or Real Time Energy, which were Olivines online monitoring system to produce real
modeled after existing proxy demand resources time baseline calculations in event dispatches,
from the CAISO. An overview of this system is ensuring that the BMW systems can accurately
described in the figure below. meet the dispatch requirement.

To meet the EV managed charging component of See Figures 25-27 in the Appendix for a
the pilot, BMW has enrolled 96 BMW i3 drivers description and example of the systems used
located within the South Bay Area to participate to track the performance of each event.
in this pilot.1 All vehicles are owned and operated

FIGURE 3 BMW i ChargeForward System Architecture

PG&E initiates a DR event to

BMW (via Olivine) by sending
a signal via a standard
1
communication protocol
(OpenADR 2.0b) similar to PG&E
how PG&E communicates 3 2 BMW
with other DR providers. Once
the event has been triggered,
BMWs aggregation software
determines how much of the
100 kW load drop will be met
by managed charging and
how much by the stationary
storage resources made of
used EV batteries, or Managed
a combination of both. Charging
1
BMW enrolled 96 customers at the outset of the project.
At the end of the pilot 92 customers remained.

PAGE 9
2 System Architecture & Project Development

Demand Response BMW Group 2nd Life

Technology Stationery Battery Storage
System
Development
In January 2015, PG&E issued an
BMW was able to successfully fulfill their interconnection agreement granting
demand response commitment of 100 kW permission for BMW to operate its BMW
with the combination of BMW i3 vehicles and Group 2nd life battery system. The 2nd
a stationary battery system housed at their life battery system was an essential
Mountain View, CA campus. The stationary component to the overall system
storage asset was built to leverage used BMW architecture of this pilot. This battery
Group 2nd life EV batteries from the MINI E provided BMW with the ability to support
Cooper trial which ran from 200920012. the full required capacity at any given
In order to control the charging of the pool DR event, regardless of the vehicle pool
of BMW i3 vehicles and the stationary contribution. A description of the battery
battery storage, BMW developed a backend system is provided on next page:
system which was successful in optimizing
the contribution of the vehicle pool and
the BMW Group 2nd life battery storage
system to provide the required 100 kW. The
section below provides more detail regarding
the development of the stationary battery
storage system and the back-end system
used to manage the vehicle pool and the
battery system as well as the on-site energy
management system at the Mountain View
Campus.

PAGE 10
2 System Architecture & Project Development

Overview of 2nd Life Battery System During commissioning, several issues of

the system were discovered and a task force
Name Plate: led by BMW was put in place to resolve
225 kWh, 100 kW AC (charging and these issues. Once resolved, the BMW Group
discharging) 2nd life stationary battery storage system
Components: ran reliably throughout the pilot with minor
issues (additional details regarding issues
8x EV Grid MINI E battery packs (used)
and resolutions are described in Section 4).
in parallel
8x EV Grid MINI E BMS (battery
A micro grid system was developed at
management system)
BMW Group Technology Office USA to
1x EV Grid Super-BMS
coordinate the 2nd life battery charging with
1x Princeton Powers Systems GTIB
on-site renewable energy and grid energy,
100 inverter
supporting the main office building functions
1x Princeton Powers Systems while ensuring sufficient state of charge
site controller to cover BMWs commitment to provide
1x Princeton Powers Systems 100 kW of demand response.
isolation transformer
10ft container, fire suppression system, The remaining flexibility when recharging
battery rack, e-stop, breaker panel, the battery was leveraged to optimize the
thermal management (cooling) system, etc. usage of locally produced solar energy.
The scheduling of the recharging process
also incorporated next-day weather
forecasts to better predict the future solar
energy produced and thereby stored in
the stationary battery system.

FIGURE 4 Solar Production and Battery Recharge

This graph illustrates the local solar production during three consecutive days. Energy production from
the solar panels as well as battery recharging is displayed as negative values. The energy exported from
the battery to the building or grid is positive. The two distinct orange spikes indicate the DR events called
in the evenings where the battery dispatched power.

100

80

60

40
Battery Power (kW)
POWER (kW)

20

-20 Solar Power (kW)

-40

-60

-80
9.15.16 10:00

9.15.16 14:00

9.15.16 18:00

9.15.16 22:00

9.16.16 10:00

9.16.16 14:00

9.16.16 20:00

9.17.16 12:00

9.17.16 16:00

9.17.16 20:00
9.15.16 2:00

9.15.16 6:00

9.16.16 2:00

9.16.16 6:00

9.17.16 2:00

9.17.16 0:00

9.17.16 4:00

9.17.16 8:00

Solar and Battery Profile of the BMW Group Technology Office USA

PAGE 11
2 System Architecture & Project Development

EV Pool Control: Delay issues of several minutes (up to

Back-end Automated 10 minutes) were observed during the first
month of DR events. The root cause for
Aggregation these issues was identified by an analysis
The EV Pool Control system aggregates the of the system log files. The issue was
vehicles in the BMW i3 vehicle pool (vehicles are resolved by deploying a more advanced
owned by pilot participants) and the stationary algorithm. However, the team experienced
battery system (owned by BMW), to provide additional delays as a result of the cellular
the requested DR capacity. The back-end carrier. Issues with the communication
automated aggregation system was effective in between the vehicle back-end and the
prioritizing the participants driving needs while BMW i3 vehicle fleet during early afternoon
ensuring that BMW was able to meet their 100 hours Pacific Time were observed around
kW load requirement between the vehicles the time of project launch. Communication
and the battery system. For each DR event issues caused signals sent between the
requested there was an associated latency, back-end and vehicles to be occasionally
or delay, between when the signal is sent and delayed or lost, resulting in vehicles not
when the load drop is recorded. The latency participating in DR events when they
levels observed throughout the pilot were in line system requested they reduce their
with the acceptable levels needed to meet the charging load. This issue was not caused
majority of grid services in California. by the BMW i ChargeForward program,
but is a more fundamental problem of the
Figure 5 outlines the average time it took from GSM-based Telemetry Services. A BMW
sending the load drop signal to the vehicle until task force successfully identified issues
the load drop is visible in the meter data. For with servers of the cell phone carrier as
vehicles that are plugged in at the start of a DR the root cause, resolving the issue by
event, this delay can be negative, as the signal September 2015.
is sent 2 minutes prior to the event.

FIGURE 5 Average Delay of Vehicle Pool by Event

Average Event Delay Time
24
For each DR event there
was latency between the 22
time when the load drop
20
signal was sent and when
the vehicles picked up 18
the signal and responded, 16
thus registering the load
drop in the data read out. 14
This figure shows the 12
average time it took the
vehicles in the vehicle 10
pool to respond to each 8
event from BMWs signal.
6

2
MIINUTES

-2
7.30.15

9.6.15

10.7.15

11.9.15

11.28.15

12.28.15

1.25.16

2.16.16

3.4.16

3.23.16

4.12.16

5.10.16

5.28.16

6.16.16

7.9.16

7.31.16

8.23.16

9.14.16

9.29.16

10.25.16

11.17.16

12.15.16

DATE
2 System Architecture & Project Development

There were additional delays throughout EV Pool Control:

the pilot that were not primarily attributed Managed Charging of a
to the vehicle telematics or the vehicle pool
control algorithm. These delays were a result
Fleet of BMW i3 EVs
of various different causes, which included BMW is able to manage the charging process of
longer and unexpected processing times on the BMW i3 vehicle pool vehicle pool by leveraging
backend servers due to load sharing and cellular (GSM-based) telemetry services. The details
communication delaysthe communication of this technology are proprietary BMW intellectual
mechanisms between the different subsystems property. The communication link from the BMW
were not primarily optimized for low latency back-end system to the vehicles has shown to have
operation. However, in over 200 events called a typical latency of 10-30 seconds, a latency level
throughout the pilot, the vehicle pool had only appropriate to meet the majority of grid service in
seven events where the average delay was California. This latency reflects the capabilities of
greater than 10 minutes. Rather, the average the communication architecture used specifically
latency per event was only 2.3 minutes, far in this pilot. Other telematic approaches may result
below the required response time of 4 minutes in a lower latency. It is important to note, that this
for modeled Real Time and 24 hours for covers only the latency of the telematics system.
modeled Day Ahead events. Individual vehicles The complete technical chain of command consists
had an average reaction time of 3 minutes of a number of steps between calling a DR event
from the signal, slightly more than the event and when the effect can be measured which has
average but still within the 4 minute target. provided an average of 2.3 minute latency and was
While a faster response was not required for noted in Figure 5.
this pilot, there is potential for the vehicle pool,
by leveraging on-board telematics, to have a The back-end control software developed by BMW
shorter latency period. for this project was operational from project go-live
on July 30, 2015 to the completion of the project
at the end of December 2016. Several system
upgrades were released throughout this time to
improve the resources performance regarding
delayed activation of the resources, refining the
vehicle pool controls, and optimizing the push
notification methodology from the smart phone
app notifying customers of the event.

PAGE 13
2 System Architecture & Project Development

BMW i ChargeForward While not a component of the BMW i

Smart Phone App ChargeForward App, participants had the
ability to set a departure time with the
The BMW i ChargeForward smart phone app BMW i Remote App, the BMW Connected App
was specifically developed for this program and the in-vehicle display. This ensured that
to gather information from participants to their vehicle had enough time to fully charge
ensure that their mobility needs were always and be ready for their selected departure time.
met. This gave the participant full control over
the charging process of their vehicle. Once
enrolled in the BMW i ChargeForward program,
participants were defaulted to be selected Customer Enrollment
to participate in an event if their vehicle was
available but were always given the option to BMWs customer enrollment strategy for the
opt out. BMW i ChargeForward pilot included reaching
out to BMW i3 owners in the San Francisco
Customers had options to make sure that their Bay Area directly as well as through social
car was fully charged whenever they needed it: media outlets.
Customer could opt out pro-actively for a
single day from the BMW i ChargeForward The BMW i ChargeForward landing page
program. If an event was called that day they (BMWiChargeForward.com) was developed in
were not selected. Q4 2014 and was intended to answer frequently
Customers received push notifications if asked questions about the pilot as well as
their car was selected for an event. If the provide an online application. The pilot was
time did not fit with their needs they could included in the BMW Press Conference at the
opt out of that event and their car would 2015 International Consumer Electronics Show
resume charging. (CES) in Las Vegas, Nevada. On the same day
as the press conference, BMW and PG&E both
issued press releases. The program was well
Aside from just managing the charging times,
covered by influential news and social media
the BMW i ChargeForward app provided
sites, including: the San Francisco Chronicle,
the user with an overview of their incentive
Inside EVs, Automotive World, CleanTechnica,
balance (further discussed in Section 5) and
Green Car Congress, GreenTech Media, Utility
a FAQ section that was updated periodically
Dive and Auto Evolution.
throughout the program.

FIGURE 6 BMW i
ChargeForward
Smart Phone App
Interface of the BMW i ChargeForward
smartphone app

PAGE 14
The press coverage combined with targeted BMW Group Technology Office USA team
emails to all BMW i3 owners living in the South conducted interviews with the participants
Bay Area in January 2015 were successful who submitted the full application with the
tactics in driving traffic to the online application. following goals:
In March the BMW Group Technology Office Verify all information submitted online
USA hosted a meetup of local BMW i3 drivers Review project overview and goals
to improve the connection to the local owner Test vehicle connectivity in
network and provided additional information charging location
regarding the pilot. A second round of targeted Go over contract and sign if possible
BMW i3 owner emails was sent in April 2015 to
include all customers who had bought vehicles
During the enrollment process the BMW team
and live in the project area. Over the course
conducted 107 in-person interviews, of which 96
of the participant enrollment process, BMW
customers were subsequently enrolled in the pilot.
received significant interest in the pilot, with over
By the end of the pilot, 92 customers remained.
500 applications for only 100 available spots.
Since the project started, four customers left the
project due to a variety of reasons: two customers
In parallel the BMW team worked with BMW were unenrolled due to cellular connectivity issues
of North Americas Legal team, PG&E and its at their residence, one customer was in a car
partners to define all participant requirements accident resulting in the loss of the vehicle and the
and draft a contract to be signed by each last customer moved outside the project area.
customer and by BMW of North America. This
contract defines participant responsibilities,
On July 22, 2015, the BMW Group Technology
BMW responsibilities, and outlines criteria for
Office USA hosted a Kickoff Event for program
contract termination.
participants. The goal of the event was to
introduce the program participants to each other,
The online application was designed to quickly as well as the BMW and PG&E team. Over 50
identify interested participants and pre-screen customers attended. Following the Kickoff Event,
them based on the five most important criteria the BMW team distributed a customer handbook
for the pilot, before directing them to the full to all participants that provided extensive program
application. These criteria were: details, best practices, and a contact list in case of
Owning a BMW i3 by July 2015 questions/concerns.
PG&E customer
W illing to charge BMW i3 primarily at home Coinciding with the program launch, BMW i
during pilot ChargeForward received a second round of news
Level 2 EVSE installed by July 2015 media coverage due in part to two well-timed
Live in PGP2 Sub-LAP (determined interviews by the BMW and PG&E teams with
by ZIP code) the Atlantics CityLab website and Bloomberg
Business. Media coverage included: CityLab,
Approximately 500 people signed up as Bloomberg, Engadget, Chicago Tribune, Fortune,
interested in the program, of those that CNET, Business Insider, Forbes, SlashGear, The
indicated they were interested only 275 passed Verge, Transport Evolved, Green Car Congress,
the primary five criteria. Of those that passed and Ecomento.
the pre-screening almost 200 submitted the
full application.

PAGE 15
3 Analysis of Charging Behavior and
Driver Archetypes

FIGURE 7 Weekday and Weekend Charging Demand from the BMW i3 Fleet

The figure below displays the weekend and weekday charging of the BMW i3 vehicle pool from 8/201512/2016.
The demand curve has a similar shape for both weekday and weekend but the weekday curve is characterized
by a steeper and larger peak demand around midnight.

Weekday Power Draw (Aggregate) Weekend Power Draw (Aggregate)

180
160
140
120
100
80
60
40
20
kW

0
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0:00
1:00
2:00
3:00
4:00
5:00

HOUR

PAGE 16
3 Analysis of Charging Behavior and Driver Archetypes

This section examines the charging behavior band showing the inter-quartile range (25th75th
of the BMW i ChargeForward participants percentile). The light blue section represents the
from a variety of angels to identify trends in minimum and maximum demand measured.
charging and help influence future managed
charging programs. The weekday charging profile for the BMW i Charge
Forward fleet is relatively low from 6:00 AM to 4:00
First, the demand load curves for weekend PM, and increases beginning at 4:00 PM creating a
and weekday charging are described. Second, small peak or increase at 8:00 PM and a large peak
the section explores three possible charging at 12:00 AM. The weekend demand profile has
archetypes of the participants studied and lower maximum demand and less variability.
variations in the charging behavior of the
battery electric (BEV) and range extender (REX) The vehicle share had a wide range of load
versions of the BMW i3. Lastly, the section contribution based on when the event is called due
examines other locations, aside from home, to the charging patterns of the participants. On
that the pilot participants charged. average, 20% of the total resource was attributed
to the vehicle pool and 80% was provided by the
The demand response capacity from the battery. However, the highest vehicle contribution
vehicle pool varies significantly over the course was during events that were called within PG&Es
of a day and by day of the week. Figure 7 off-peak time-of-use periods, specifically
(on page 16) displays the aggregate weekday 11:00 PM2:00 AM. The higher share at these
and weekend charging demand from the BMW times is the result of about 60% of the BMW i
i3 Fleet between 8/201512/2016. ChargeForward participants that are on a time-
of-use rate plan, either the whole house electric
This is baseline charging data, excluding vehicle rate plan (EV-A) or the tiered, time of use
demand response event days. For these figures, plan (E-6). These rate options incentivize off-peak
a day is defined as starting at 6:00 AM and charging by offering a lower price per kWh between
ending at 5:59 AM the next day. The center the hours 11:00 PM7:00 AM on weekdays and
white line in each figure represents the median 7:00 PM2:00 PM on Weekends and Holidays.
power draw for the EV pool, with the green

FIGURE 8 Average kW Contribution and Vehicle Participation per Event Hour

kW
The time of hour the day an event is 0 5 10 15 20 25 30 35
called was strongly correlated to the
number of vehicles participating. This 23:00AM
was largely due to the high number of 67:00AM
participants on a time-of-use electric 78:00AM
rate plan that incentivize charging
after 11 PM. 89:00AM
34:00PM
45:00PM
56:00PM
67:00PM
EVENT HOUR

78:00PM
89:00PM
910:00PM
1112:00AM
0 2 4 6 8 10

NUMBER OF VEHICLES
3 Analysis of Charging Behavior and Driver Archetypes

Driver Archetypes spike in the evening regardless of weekday

or weekend. A majority of frequent drivers
are on one of PG&Es time-of-use (TOU) rates
In analyzing the charging behavior of
which would explain the significant spike
participants, three clear groups were
around 11:00 PM on weekdays. Household
identified. The first group, Frequent Drivers
drivers have the lowest spike and greatest
(42), is characterized by having a consistent
variability in charging times both on the
charging pattern and a commute of greater
weekdays and weekends.
than 30 miles a day. The second group,
Infrequent Drivers (45), had no regular
Weekends have a wider distribution but
charging schedule and drive less than 30
overall a similar load shape. Frequent
miles a day. The third and smallest group,
drivers have about half the power draw on
Household Drivers (9), had no regular
weekends, 35 kW, compared to weekdays,
schedule but drive more than 30 miles a day.
75 kW. Infrequent and household drivers
power draw is slightly lower on weekends
As expected, frequent drivers charged more
by about 3 kW which suggests there is no
often than household drivers or infrequent
significant change in driving distance or time.
drivers due to typically longer commute and
drive times. Frequent drivers had the tightest
distribution curve with the most pronounced

FIGURE 9 Aggregated Power Draw by Driver Archetypes

Frequent drivers have the largest power draw of the three driver archetypes.
Household drivers power draw curve is slightly skewed due to the small sample
size of customer that fit this driver archetype.

Weekday Power Draw at Home Weekend Power Draw at Home

80

60
POWER DRAW (kW)

40

20

0
6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR
Frequent Drivers
Household Drivers
Infrequent Drivers

PAGE 18
3 Analysis of Charging Behavior and Driver Archetypes

Comparison of Battery
Electric and Range Extender
Participants
The BMW i3 is offered in two different
versions, a battery electric vehicle (BEV) or
range extender (REX). For the 2014 BMW
i3, the BEV version is fully electric with an
EPA estimated 80 miles per full charge. The
REX version of the i3 has an EPA estimated
range of 150 miles due to the addition of a
small gas engine that charges the battery
while driving and is designed to enable the
driver to reach the next charging station and
reduce range anxiety.2 Within the BMW i
ChargeForward vehicle pool 44 participants
have a BEV and 48 have a REX BMW i3.

FIGURE 10 Battery and Range Extender Aggregate Power Draw Comparison

REX drivers have an earlier peak and a slightly larger power draw throughout the day for both weekday
and weekend. The power draw difference between REX and BEV pool at the peaks is about 10 kW with the
exception of the second weekday peak at midnight when the power draw is the same. This slight difference
in power draw may suggest that REX drivers drive slightly further between charging event. However, overall
the charging behavior is very similar between the REX and BEV.

Weekday Power Draw at Home Weekend Power Draw at Home

80

60
POWER DRAW (kW)

40

20

0
6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR
BEV
REX

Since i ChargeForward launched, BMW has increased the range for both the BEV and REX.
2

Participants for this pilot had the 2014 version of the BMW i3.

PAGE 19
3 Analysis of Charging Behavior and Driver Archetypes

Non-Home Charging While participants have been able to charge

around the Bay Area, their main source of
The BMW i3 offers a competitive range on charging is at their home. Although 81% of
a single charge but occasionally drivers will BMW i ChargeForward participants work
need an additional charge before they return full time but only 37% charge at both home
home. Although charging stations are not and work. Those who do not drive their
yet ubiquitous, participants have been able BMW i3s to work rarely, if ever, charge their
to charge around the Bay Area. Participants vehicles away from home. The largest barrier
charge in two key two key areas aside from to charging away from home, reported from
their home, at work and at public charging participants, is the availability of chargers,
stations. costs, and the risk that charging stations at
a specific location will be full.

FIGURE 11 Alternative Charging Locations (Non-Home)

The figure below indicates where participants have charged outside their homes.
Areas of purple and orange have a higher frequency of charging events.

All customers, workplace charging All customers,not

All customers, notatat work
home oror home
work

Walnut Creek Walnut Creek

Berkeley Berkeley

Oakland Oakland

San Francisco San Francisco

San Leandro San Leandro

Daly City Daly City

Hayward Hayward
Pleasanton Pleasanton

Pacifica Pacifica
Union City Union City
San Mateo
Fremont San Mateo Fremont

Palo Alto Milpitas Palo Alto Milpitas

Sunnyvale Sunnyvale
San Jose San Jose

Campbell Campbell

Los Gatos Los Gatos

PAGE 20
3 Analysis of Charging Behavior and Driver Archetypes

PAGE 21
4 Participation in Demand Response
(DR) Events

3
Prior to October 2015, BMW was required to provide 80 kW.
4
These percentages are derived from the total delivered amount, 19,500 kW.
5
Graph does not total 100% since BMW did not meet the full 100 kW in every event.

PAGE 22
4 Participation in Demand Response (DR) Events

Between the start of the project in July

2015 and end of December 2016, the BMW
System Performance
resource was called 209 times. In order to
In total, BMW has participated in 209 Demand
test the flexibility of the vehicles and battery
Response (DR) events. As stated in Section 1,
as a grid resource, events were tested in
BMW was required to provide capacity of
both Day Ahead (notifications sent 24+
100 kW over an hour-long period.3 BMW
hours before the event) and modeled Real
met the performance requirements for 90%
Time (notifications sent 4 minutes prior to
(189) of the events with an average delivered
the event). These events were modeled after
contribution of 20% for the vehicle pool and
existing proxy demand resource products
80% from the 2nd life battery system.4
developed by the California Independent
A successful event is defined as BMW
System Operator.
reaching 90% (90 kW) of the target (100 kW)
load reduction.
This section describes BMWs performance
in responding to the DR events called.
The graph below describes at what percentage
Specifically the section addresses the
of the target (100 kW per event) was attributed
overall system performance, an evaluation
to the vehicle and battery share.5
of Day Ahead and Real Time events, and
vehicle participation. In addition, this section
provides additional detail regarding the issues
that impacted system performance and the
resolutions to these issues.

FIGURE 12 Vehicle Performance from Target (100 kW)

The vehicle pool contributed an average of 20% of the target kW reduction.

100%

90%

80%

70%

60%

50%

40%

30% 80% Avg

Battery
Contribution
20%
20% Avg
10% Vehicle Pool
Contribution
kW

0%
0 20 40 60 80 100 120 140 160 180 200
EVENTS
Average of Vehicle Share Target
Average of Battery Share Target

PAGE 23
4 Participation in Demand Response (DR) Events

Real Time and per event. On average, the per vehicle capacity
during an event was 4.43 kW.8 The lower
Day Ahead Events capacity observed (compared to the maximum
capacity) is a result of a portion of vehicles that
Between July 2015 and the end of December either end or start a charging event during an
2016, the BMW resource reliably provided grid hour-long DR event.
services for both Day Ahead and Real Time
events. The results indicated that there was no During the summer when demand response
significant difference in the vehicle and battery programs are utilized frequently, customers
contribution percentage between Real Time and often feel the burden of repeated events in a row.
Day Ahead. This can place a burden on customers to perform
and often result in less successful events (lower
Across all DR events (Both Real Time and Day curtailment), poor customer satisfaction, and
Ahead), the vehicle pool of customer owned increase dropout rates. Unlike typical demand
BMW i3 vehicles has contributed, on average, response programs, the BMW i ChargeForward
approximately 18% while the BMW Group 2nd Pilot had significantly more events at a much
life battery system has contributed 77% of the higher frequency. For context, PG&Es SmartRate
targeted DR resource.6 While the vehicle share residential demand response program caps the
has varied throughout the pilot, the majority of number of events at 15 per year whereas the
events have been between 15 and 35% of the BMW i ChargeForward pilot had over 200 events
total DR resource. Each vehicle has a maximum across an 18 month period.
capacity of 6.6 kW per charge7; results from the
pilot indicate a lower average vehicle capacity

FIGURE 13 Day Ahead and Real Time Delivered Contribution Comparison

The vehicle pool contributed

to an average of 20% of the 20.1%
total resource for both Day
Ahead and Real Time events
respectively
79.9%
Day Ahead

20.2%
79.7%
Real Time

Average of Vehicle Share Average of Battery Share

6
Percentages come from the targeted amount (100 kW per event) and is slightly less than the percent from the total delivered kW.
7
Based on the maximum capacity of the BMW i3 on-board charger.
8
Average vehicle capacity was calculated by averaging the kW reduction attributed to each vehicle that participated throughout the
programs duration.

PAGE 24
4 Participation in Demand Response (DR) Events

Throughout the pilot, PG&E called an average The low opt out rate suggests that
of three to four events per week. BMW was customers were not negatively impacted
able to reliably and successfully respond by the program and didnt feel customer
to these events over 90% of the time. This fatigue from consecutive events.
indicated that both the vehicles and the Further, only two participants opted
battery had a consistent response time and out for more than two events over the
were able to meet the contribution multiple 18 months. Based on surveys of BMW
days in a row. Since this program was i ChargeForward participants, most
designed to run primarily in the background participants never felt the need to opt out.
of participants lives they were able to Of the participants surveyed, 95% noted
participate at higher rates and felt little to no that they never, or very seldom, had to
customer fatigue. change their driving or charging behavior
as a result of participating in the BMW
This is further evidenced by reviewing the i ChargeForward program.
customer opt-out rate. Each customer had
the ability to opt out of each event at any
time. However, the overall rate was low
throughout the pilot. The most opt outs for
one event was on the October 14, 2015
11 PM event with three customers opting
out. The majority of events had no opt-outs.

FIGURE 14 Number of Opt Outs per Event

Opt out rates were very low for BMW i ChargeForward in

comparison to the number of participants per event, only
6% of participants opted-out of an event.

1
OPT-OUTS

0
7.30.15

9.6.15

10.7.15

11.9.15

11.28.15

12.28.15

1.25.16

2.16.16

3.4.16

3.23.16

4.12.16

5.10.16

5.28.16

6.16.16

7.9.16

7.31.16

8.23.16

9.14.16

9.29.16

10.25.16

11.17.16

12.15.16

DATE

PAGE 25
4 Participation in Demand Response (DR) Events

Vehicle Participation The top 10% of vehicles with the highest

number of event participation share similar
driver characteristics. They are characterized
As noted in Section 3: Charging Behavior Analysis, the
as frequent drivers, who have regular charging
number of vehicles participating in an event is
patterns and are not on a TOU rate. These
correlated to the participants residential rate.
drivers habitually plug in and begin charging
In the BMW i ChargeForward group about 60%
around 8 PM in the evening and typically charge
of the participants are on a Time-of-Use (TOU)
for about 3 hours. Since a majority of the events
rate that provides an economic incentive to
were called from 89 PM, these vehicles were
charge off-peak between the hours of 11:00 PM
frequently called upon and able to participate.
and 7:00 AM.

The highest number of BMW i3s participating

at one time was 29. On average, 7 vehicles
participated in a given event with the majority
of events having between 5 and 11 of all pilot
vehicles participating. The number of vehicles
per event was strongly correlated to the
charging behavior and residential rate structure
of the customer as mentioned in Figure 8.

FIGURE 15 Vehicles Participating per Demand Response Event

Demand Response Events per Vehicle. The figure below displays the distribution of vehicles participating in events
throughout the pilot. The bars represent the number of vehicles participating in events. For example, in 27 events
five vehicles participated. The average number of vehicles participating in an event was seven. However, five events
had over 20 vehicles participating.

30

25
NUMBER OF DR EVENTS

20

15

10

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 24 25 29
NUMBER OF VEHICLES PARTICIPATING

PAGE 26
4 Participation in Demand Response (DR) Events

FIGURE 16 Demand Response Events per Vehicle

The figure below displays the distribution of event participation per vehicle. The bars represent the number of
events in which each vehicle participated. For example, four vehicles participated in 24 events. On average vehicles
participated in eight events across the pilot. However, three vehicles participated in over 50 events.

5
NUMBER OF VEHICLES

0
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72
NUMBER OF EVENTS PER VEHICLE

PAGE 27
4 Participation in Demand Response (DR) Events

Program Issues and

Solutions
Backend software updates were deployed
during the first 30 events to improve the system
response. However, there have been events
where the vehicle pool combined with the
2nd life battery or the overall system failed to
meet the required load curtailment. Problems
occurred in various areas in the technical chain
of command in the system, which was spread
among different servers at different providers
(Olivine, BMW Munich, BMW of N.A., and BMW
Technology Office in Mountain View, CA). For
each problem, the root cause was identified by
the affected project partners, and a systematic
solution was targeted to prevent the specific
error from occurring again. This strategy
improved the overall system stability.

9
The server that transmits OpenADR signals to end devices or other intermediate servers, in this case the intermediate server is BMWs
backend automated aggregation server.
10
The server that accepts the OpenADR signal, in this case the server is BMWs backend automated aggregation server.
11
The communication between the OpenADR Server and OpenADR Client noting that the OpenADR client has received the OpenADR signal.
12
Database and system that verifies the amount of load reduction from BMW, confirms that event requirements have been met and calculates
payments for modeled grid services.

PAGE 28
4 Participation in Demand Response (DR) Events

An overview of different problems which occurred is given in the following table.

DR performance Issue Resolution

OpenADR server9 taken down during a maintenance Once back online, the OpenADR
cyclea DR event was not communicated to the server again was able to communicate
OpenADR client.10 DR events.

OpenADR client did not pick up the DR event creating a Increased the timeout values at
timeout issue in the OpenADR handshake.11 The server the OpenADR client to allow the
took longer to respond to a request than the OpenADR OpenADR server to provide longer
client was waiting. processing times.

EV Pool Control failed as a result of a server The server was restarted and
maintenance cycle. automated restart-mechanism was
implemented.

Stationary 2nd Life Battery Storage System was not Restarted the battery control
operational during a DR eventcontrol program was program and implemented automated
not working correctly. Battery did not power to the grid restart mechanism.
during an event.

Whisker Labs power measurement of single participants Temporarily excluded malfunctioning

charging stations did not work properly. The power value meter from the system. Meter
was stuck at a static value, which did not reflect the true was repaired and re-connected to
charging power. This inflated the amount of capacity the system.
available from the vehicle pool, resulting in a lower overall
system contribution.

Communication errors in sending power measurement Manually reformatted and fed-in

data from the battery control system (of the stationary reliable historical raw power data into
battery) to the Olivine settlement database.12 This caused the Olivine systems.
the DR event to be flagged as failed even though the
battery did dispatch the correct amount.

Battery cell replacement took longer than expected. The issue was resolved by revising,
Battery did not power to the grid during an event. repairing and adapting the configuration
While attempting to loop-in a repaired battery pack in in the battery system components
the second live battery the battery control loop was (site controller).
non-functional.

The overall technical architecture has proven to be viable with minimal issues. With each new issue,
the project team was able to identify and correct the problem so it did not become a reoccurring issue.
All challenges encountered have been of local nature or have been related to a specific communication
link. It is important to emphasize that no overarching problem occurred that would fundamentally
question the integrity and design of the overall program architecture.

PAGE 29
5 Overview of Customer Behavior Research

PAGE 30
5 Overview of Customer Behavior Research

PG&E and BMW executed a series of surveys The objective of the first phase was
and focus groups from February 2016 to to explore the primary motivators for
December 2016 in order to gain a deeper participating in the BMW i ChargeForward
understanding of pilot participant charging Program, and participants perceptions
behavior and charging flexibilities. This research and experience with the program and the
specifically sought to better understand launch process. A total of four focus groups
participants motivation for participating in the (two in-person and two-online utilizing
pilot, perceptions and experiences with the webcams) were conducted among BMW i
program, understanding participant profiles, ChargeForward participants.13 In addition,
preferences for future managed charging an online survey distributed to the BMW i
programs (both reducing and increasing ChargeForward participants.14
charging), and the general comfort and trust
associated with allowing their charging to The objective of the second phase was
be managed by a third party. This section to understand an EV drivers general
describes the research methods, objectives comfort and trust in a managed charging
and results. program concept, that includes a third party
controlling charging by either reducing or
increasing charging based on grid conditions.
In addition, this research also explored what
Objectives and tactics would drive confidence and potential
Methodologies engagement in this program, the thresholds
for allowing a third party to manage
The BMW i ChargeForward team partnered customers battery charge, and how other
with Ipsos RDA to conduct an approximately factors impact customer confidence. Two
year-long research study looking into the needs online focus groups were conducted among
and motivators of EV drivers. The research BMW i ChargeForward participants.15 In
component was broken into two phases, addition an online survey was sent to BMW
BMW i ChargeForward program research and i ChargeForward pilot participants as well
managed charging program research. as EV owners from PG&Es Customer Voice
Panel Members.16

13
The two Customer Research In-person focus groups had 7 participants and 8 participants. The two Customer Research
online focus groups had 4 participants and 5 participants.
14
The survey was completed by 63 participants. At the time of the survey, 94 participants were enrolled resulting in a
67% response rate.
15
The first Managed Charging focus group had 7 participants, all of whom used their BMW i3 to commute to work. The second
Managed Charging focus group had 6 participants, 4 of whom were retired and 2 commuted to work by other means of
transportation.
16
PG&Es Customer Voice Panel is comprised of PG&E customers that have agreed to participate in online surveys. No incentive
was offered to these customers for taking the survey described in this report. Of the 1,054 plug-in EV Owners in the panel,
a total of 332 surveys were completedresulting in a 32% response rate. Of the 94 BMW i ChargeForward participants, a total
of 67 surveys were completedresulting in a 71% response rate.

PAGE 31
5 Overview of Customer Behavior Research

BMW i ChargeForward Results from the quantitative survey support

the findings in the focus groups but also
Program Research suggest that the incentive ranks as one of
the most important factors for enrolling in
In the first portion of the customer research, the program. The most frequently reported
PG&E and BMW explored the primary motivators reasons for deciding to participate in the
for participating in the BMW i ChargeForward program are to provide input in the next
program and the participants perceptions and generation electric vehicle and the up-front
experience in the program. This research took incentive (84% of participants stated that
place in February and April of 2016, about half these two reasons were somewhat/extremely
way through the 18 month program. important for both). See Figure 17 for the
complete responses to the motivating factors
to participate in the pilot.
Motivation to Enroll and
Participate
Importance of Incentives
As previously stated, PG&E and BMW
conducted four focus groups with participants As part of this pilot, customers receive an
to obtain qualitative customer feedback. During upfront incentive of $1,000 and an ongoing
the focus groups, most pilot participants incentive for each day they do not opt-out
identified the incentive as a key motivation for (whether an event was called or not), up to
participating (primary or secondary motivator). $540 that is distributed after the pilot has
Most participants also wanted to assist BMW ended. It is clear from the survey results
and PG&E in researching sustainability and that the incentive was a significant factor
transportation, promoting grid stabilization, reuse in participation in the pilot. The majority of
of EV batteries, and input for the next generation participants (84%) identified the up-front
electric vehicles.

FIGURE 17 Motivation to Enroll and Participate in the BMW i ChargeForward

Reasons for Participating in the

BMW i ChargeForward Pilot: % of
participants who rated somewhat
(4) or extremely (5) important
84%
Provide input in next gen elecric vehicle +
67%
Help manage load
94%
Up-front incentive +
to the following questions, Using Up-front incentive on the electric grid Process of how you wer
a five-point scale, please indicate
how important each of the
following items were in regards to
your decision to participate in the
65% Promote reuse of EV batteries 83% Ongoing participat
BMW i ChargeForward program.
(n = 69) 62% Enjoy participating in research/pilots 80% Frequency of delay
Ongoing participation incentive
78% Participation in del
61% BMWs program involvement
77% Explanation of dela
55% To save money on charging costs
75% Explanation of use
51% PG&Es program involvement
74% Coordinating the s

70% Communication fro

67% Participation requi

PAGE 32
5 Overview of Customer Behavior Research

incentive was more important compared to Perception of Participation

the ongoing incentive (62%). It is interesting and Communication
to note that participants indicated that
helping to manage load on the electrical grid Overall satisfaction was high among
and promoting the reuse of BMW Group participants surveyed. Nearly all, 92%
2nd life batteries as more important than the participants describe their participation
ongoing incentive. in the program as passive, in which the
program runs in the background of their
Participants were also surveyed to understand daily routine and does not affect them
the effectiveness of future incentives in in any significant way.
motivating participants to join a program
like the BMW i ChargeForward pilot. Results from the focus groups and online
Up-front incentives and reduction in energy survey suggest customer communication
bills were the most preferred methods, with may be improved. Although about
89% of participants rating these methods as two-thirds (68%) of participants are
extremely effective and somewhat effective. satisfied (Somewhat/Extremely) with the
Ongoing incentives were also a popular communication they are receiving, based
choice with 84% of respondents stating on verbatim comments, it appears that
this was an effective method. Not having an most participants would desire more
incentive or offering a donation to charity frequent communication about the
would have negatively impacted participation programs impacts.17 The survey indicates
in the program, with only 19% and 23% of that the majority of customers want to
customers indicating these were effective hear more about the energy saved and
strategies for participation, respectively. how the program is performing relative
to its goals and prefer to receive this
information in a monthly email.

17
As part of the pilot, participants received a quarterly newsletter detailing program performance and activity.

PAGE 33
 5 Overview of Customer Behavior Research

Customer Satisfaction workshops. Participants like the nonintrusive

nature of the program and that their driving
Overall satisfaction with the program needs were always met. Overall, participants
was very high, at the end of the program would like more communication from BMW and
98% indicated they were satisfied with PG&E regarding the programs objectives and
the BMW i ChargeForward program (a 4 if their participation is helping BMW and PG&E
or 5 rating on a 5 point scale); this is up meet those objectives.
from 92% at the first survey in February
2016. Participants were also very likely to
participate in another delayed charging
program in the future, with 93% indicating Managed Charging
this interest. In terms of advocacy for the Program Research
program, 98% would likely recommend the
program to family/friends compared to The main objective for this phase of the research
86% during the February 2016 survey. was to understand the different driving and
charging habits of the BMW i ChargeForward
Among all the items related to participating participants and what requirements or
in the program, nearly all respondents preferences they have for a future managed
(94%) were satisfied (Somewhat/Extremely) charging program, including both reducing and
with the up-front incentive and the process increasing charging. Research was conducted
of how they were invited to participate. by doing two online focus groups with the BMW
While satisfaction is lowest for workshops i ChargeForward participants and two surveys,
introducing the details of the program, one to BMW i ChargeForward Participants and
very few (7%) were dissatisfied (Somewhat the other to a customer voice panel of EV owners
Dissatisfied/Not at All Satisfied) with the in PG&Es service territory.

FIGURE 18 Levels of Satisfaction with Elements of the BMW i ChargeForward Program

%
put in next gen elecric vehicle +
67%
Level of satisfaction with elements of
the BMW i ChargeForward program:
Percent of pilot participants who
rated somewhatHelp(4) or extremely
manage load
94%
Up-front incentive +
88%
Enrolling in the
ncentive (5) satisfied to
onthe
thefollowing
electricquestion
grid Process of how you were invited to participate program
Using a five point scale, please
indicate your level of satisfaction with
mote reuse of EV batteries
each of the following items related to
the BMW i ChargeForward program.
83% Ongoing participation incentive
(n=69)
oy participating in research/pilots 80% Frequency of delayed charging events
going participation incentive
78% Participation in delayed charging events
Ws program involvement
77% Explanation of delayed charging events
save money on charging costs
75% Explanation of user agreement/App to obtain info/opt-out of delays
&Es program involvement
74% Coordinating the setup/Using the app on smartphone

70% Communication from BMW regarding program

67% Participation requirements

PAGE 34
5 Overview of Customer Behavior Research

Charging Behavior have more flexibility in their schedule or their

vehicle is at home during the day. On weekends,
Two qualitative online focus groups were the BMW i3 is typically used to run errands, go on
conducted among the BMW i ChargeForward trips into the city, or explore other areas around
participants. The first focus group consisted of the greater Bay Area.
all commuters who used the BMW i3 as their
primary vehicle to and from work. The second Range-anxiety, or the feeling of not having
focus group consisted of retirees and household enough range to be able to meet your destination,
drivers or those who commute to work by other generally exists among those who do not have
means of transportation. Participants who use a range-extenderespecially among those who
their BMW i3s to commute to work have vastly drive their BMW i3s to work. Range-anxiety is
different opinions regarding their charging one of the primary reasons participants want to
behavior than those who do not. ensure they receive a full charge at night. This
lowers the stress around not being able to make
The utilization of the BMW i3 varied from their commute or run errands when they have a
weekday to weekend. During the week, most full battery. While those with range extenders may
participants with full-time jobs primarily use have less range anxiety, they generally plan their
their BMW i3 vehicles to commute to work, trips and charging schedule to avoid using the
and some of these participants also run various gas option of the extender.
errands throughout the week after work. Unless
workplace charging is reliable and available this Almost all BMW i3 owners charge their vehicle
makes daytime charging hard to accomplish. every night at their home. Charging away from
There are BMW i3 owners who are retired or the home varies widely. Some participants charge
use another form of transportation for their daily at work, others charge periodically when
work commute and use their BMW i3s sparingly running around town, and some never charge
during the weekprimarily for errands close anywhere but at their home. Participants like to
to their homes. They have a greater ability and leave the house in the morning with the freedom
willingness to charge during the day as they and comfort that comes with a full charge.

PAGE 35
5 Overview of Customer Behavior Research

Charging Away from Home and The participants would be more likely to charge
Barriers to Daytime Charging during the day if charging stations were reliable
and readily available, but also feel there will be
While workplace charging stations may be less of a need to charge during the day as the
provided, participants indicated that charging range of EV batteries increases.
at work is increasingly becoming more difficult.
The most frequent barrier to daytime charging,
and charging away from home, is the availability Comfort and Trust in
of charging stations, cost of charging stations, Managed Charging
and vehicles parked at a charging station but no
longer charging. As mentioned above in Section Participants are very interested in managed
3: Analysis of Charging Behavior and Driver Archetypes, charging and have a high degree of confidence
81% of BMW i ChargeForward participants work in both PG&E and BMW to effectively manage
full time but only 37% charge at both home and the charging of their electric vehicles. Despite
work. On average 85% of charging is done at high interest in the program and confidence in
home with occasional charging away from home PG&E and BMW, participants still want to retain
when it is convenient or necessary. some control of their EV charging. Specifically,
participants want to have the ability to opt-out
Those who regularly drive the BMW i3 to work of events and knowledge of the exact times and
feel charging during the day is quite tedious length for events. Overall, participants rated the
mostly due to the lack of charging stations and ability to set an exact time their EV needs to
charging ethics, such as moving your car once be at a desired charge level as one of the most
it has finished charging. Participants prefer to important program feature.
charge at night, regardless of the amount of
charge left in their battery. Participants indicated
that charging at home during the night is more
convenient and their preferred option.

FIGURE 19 Influence of Renewable Energy on Managed Charging Participation

The figure displays how renewable energy impacts participation in a potential managed charging
program. Participants indicate a strong interest in charging with renewable energy, but require an
incentive to charge during the middle of the day.

63%
70% BMW i ChargeForward
PG&E's Customer Voice Panel

Likelihood to participate at home and work

if EV is charged with renewable energy

42% 83%
41% 83%
Likelihood to participate at home Likelihood to participate at home and work
and work (between 9 AM4 PM) in order if offered additional monetary incentive
to charge EV with solar energy

PAGE 36
5 Overview of Customer Behavior Research

Thresholds and Flexibility for number of opt-outs. They feel about 20 opt-outs
Managed Charging are reasonable. These participants are familiar
and confident with the program design due to
Among four potential program features tested, their participation in the BMW i ChargeForward
participants rank the ability to set a time their car program. If the number of opt-outs is limited
would need to be at the desired charge level as to 5 times in a 12-month period at home and
most important, as participants want assurance work, about half (52%) of the PG&E Customer
their EVs will be sufficiently charged for their Voice panelists would be less likely to participate
commute. An additional important feature is while far fewer (17%) BMW i ChargeForward
the ability to set a minimum level of battery participants would be deterred by this number
charge before managed charging would begin, of acceptable opt-outs.
as participants want assurance they will not be
stranded away from their homes. Participants When presented with the option of charging EVs
are more likely to participate in a program if with renewable energy (solar or wind), roughly
they know the exact times and length of the two-thirds (68%) are more likely to participate
upcoming delayed charging events. in a managed charging program. This number
drops significantly to less than half (41%) when
Participants want to retain flexibility and control participants are told they will need to adjust their
through the ability to opt-out of managed charging to the core daytime hours (9:00 AM
charging events as they see fit. For PG&Es to 4:00 PM) to take advantage of solar energy.
Customer Voice panelists, they felt that However, the presence of an additional monetary
approximately 40 opt-outs are reasonable incentive significantly increases the likelihood of
before having an incentive negatively impacted.18 participating from 41% to 83%. Thus, there is
This relatively high number could be attributed clearly a high interest in charging with renewable
to a lack of familiarity and confidence in the energy provided the participants receive an
program since they have not participated in a additional monetary incentive for changing their
similar program before. BMW i ChargeForward charging behavior.
participants are much more lenient with the

18
Important to note that PG&Es Customer Voice panelists were not participants of the BMW i ChargeForward pilot thus do not have
previous experience with a managed charging program.

PAGE 37
6 EV Adoption and its Potential for
Further Grid Support

FIGURE 20 Forecasted EV Adoptions in PG&Es Service Territory

PG&Es 2016 Expected 1,200

Forecast for EV adoption in
PG&E service territory. Years 1,000
2017-2023 were included in
NUMBER OF EVs (thousands)

the 2016 Joint IOU Electric 800

Vehicle Load Research Report
filed on December 30, 2016 in
CPUC R. 13-11-007. 600

400

200

0
2016

2017

2018

2019

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

PAGE 38
Over the next 15 years electric vehicle the potential if this pilot were to scale to a
adoption is projected to rise dramatically larger vehicle population. Aggregating EVs
as vehicle price drops, range increases, en masse creates the potential for a large
and a wider range of models becomes demand response resource.
available. By 2020 almost every major
vehicle manufacturer is expected to have a The BMW i ChargeForward pilot provides
long-range electric vehicle on the market. some insight into the magnitude of the future
However, consumers do not have to wait load reduction of this population. The table
until 2020 as manufacturers are beginning below outlines this potential by assuming
to release these vehicles now. At the end and enrollment of 20% of customers with
of 2016 the first long-range battery electric similar behavior of BMW i ChargeForward
vehicle with a competitive, mass market price participants (8% participation rate and a
point and 238 mile range was released. contribution of 4.4 kW per vehicle). Based on
these assumptions, the potential load drop
As EV adoption grows, the potential for EVs of a single event in 2030 is about 77.6 MW,
as a grid resource becomes more significant. which is enough to power approximately
Throughout the course of this pilot, an 58,000 homes in California.23 Thus, on
average of 7 out of 92 customers participated a larger scale, a similar program has the
in each event representing approximately potential to provide a significant resource.
8% of the total vehicle pool. On average, the
vehicle contribution per event is 4.43 kW. The
average contribution is attributed to vehicles
that join the event after the start, unplug
during the event or are near the end of their
charge cycle. While this may seem like a
small contribution, it is important to recognize

FIGURE 21 Forecasted Electric Vehicles in PG&Es Service Territory

By 2020 EV adoption is expected to Year Total Projected Customers Load Drop

increase from about 100,000 vehicles in
Vehicles19 Enrollment20 Participating (MW)22
2016 to over 200,000, and by 2030 PG&E
is forecasting over 1.2 million electric
in an Event21
vehicles within their service territory.
2020 226,000 45,200 3,164 14.0 MW
2025 513,000 102,600 7,182 31.8 MW
2030 1,251,000 250,200 17,514 77. 6 MW

19
Forecasted Total number of Electric Vehicles in PG&Es service territory
20
Predicted Enrollment: 20% of the forecasted EVs in PG&Es service territory for the specified year. An average was taken since the actual
enrollment in a program is unknown.
21
Customers Participating: The average number of participants in a given event assuming that 8% of customers participate in each event.
22
Load Drop (kW): The approximate load drop per event based on the number of customers participating in the event and the average
contributed load drop (4.43 kW) per vehicle per event.
23
Glossary of Energy Terms. California Commissionwww.energy.ca.gov/glossary/ISO_GLOSSARY.PDF

PAGE 39
6 EV Adoption and its Potential for Further Grid Support

Analysis of Vehicle As discussed above, electric vehicle adoption is

projected to have substantial growth. This growth
Plug Time could provide a flexible resource if it can be called
upon during the day to absorb the excess solar
In 2013, the California Independent System generation and reduce the ramp in the evening.
Operator published research on a significant While the BMW i ChargeForward pilot did not test
change in energy balance across California. the ability to absorb excess generation, PG&E
This research coined the term, the Duck Curve and BMW did evaluate participants plug times
(Figure 22), in which the net energy load drops analyzing when a customer is connected to a
significantly throughout the day as a result of charging station. Results of this analysis suggest a
significant solar energy production and then high propensity to plug in at home (during evening
ramps up sharply later in the day as solar hours), regardless of whether the vehicle began
production drops.24 charging immediately or not. Phase 2 of BMW
ChargeForward will test the ability to absorb excess
While the research predicted the Duck Curve, solar generation as noted below in Section 7.
it has now become a reality. In fact in 2015 the
lowest mid-day net load reached 2017 predicted As displayed in Figure 23 (on page 41), it is
values at 14,335 megawatts.25 The concern clear that there are a low percentage of vehicles
with the changing grid conditions is that the charging during the weekend, however, there are
sharp decrease in net energy load during the a much higher percentage of customers that are
day coupled with a rapid increase in the evening plugged in. When we compare the data on vehicles
will result in challenging grid reliability. As a connected at home and those that are at home
result of the change in grid conditions, flexible but not charging, we see an increased potential to
resources that are able to utilize the excess leverage these vehicles charge during the day.
generation during the day and aid in curbing
the steep ramp towards the end of the day
are becoming increasingly important.

FIGURE 22 California ISO Duck Curve

Electric vehicles present Net Load: March 31

one possible solution to
help thin out Californias
26,000
increasing duck curve.
24,000

22,000
2012
20,000
2013
18,000
Increased
2014 ramp
16,000 Potential 2015
overgeneration 2016
14,000 2017
201819
MEGAWATT

12,000 2020

10,000

0
0:00

3:00

6:00

9:00

12:00

15:00

18:00

21:00

HOUR

24
California ISO, Flexible Resource Fast Facts. www.caiso.com/Documents/FlexibleResourcesHelpRenewables_FastFacts.pdf
25
St. John, Jeff. (2016, November) The California Duck Curve Is Real, and Bigger Than Expected. GreenTechMedia

PAGE 40
6 EV Adoption and its Potential for Further Grid Support

While there is limited load consumption from

BMW i ChargeForward participants during the
day, results from the customer research suggest
that there is an opportunity to harness electric
vehicle charging at the workplace. The vast
majority (81%) of participants surveyed work full
time and a slightly smaller number (79%) use
their BMW i3 to commute. The majority (79%) of
these participants also indicated that the BMW
i3 was their primary vehicle with over half (62%)
indicating they charge once a day. These survey
results suggest there is a significant opportunity
to leverage EV charging at the workplace.
However, it is important that these customers
have access to charging at their place of
employment and that programs are available
to leverage this increasingly valuable load.
While the majority of participants commute
to work, less than half of the respondents (41%)
have access to charging at their workplace.

FIGURE 23 Weekend Percent of Vehicles Connected vs. Charging

The two graphs display the comparison between the percentage of vehicles charging and the
percentage of vehicles plugged in but not charging.

Percent of Vehicles Connected at Home: Weekends Percent of Vehicles Charging at Home: Weekends
100%
90%
80%
70%
60%
50%
40%
Vehicles

30%
20%
10%
0
6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR

PAGE 41
7 BMW ChargeForward Phase 2

PAGE 42
7 BMW ChargeForward Phase 2

BMW and PG&E will collaborate to build Near-term and long-term advances in electric
on the learnings described in this study vehicle technology necessitate taking a new,
through the implementation of a second broader view when managing vehicle charging.
phase of the BMW ChargeForward project. Vehicle battery size is increasing, which will
The second phase of the pilot will explore reduce the need for vehicles to charge every
two primary themestesting advanced night for most vehicles. The average commuter,
smart charging use cases that promise driving 3040 miles per day, may no longer find
additional value to the grid; and evaluating it necessary to charge every day or even every
customer engagement strategies that other day. With this range, drivers may be able
incentivize drivers to provide additional to charge only two or three times per week to
flexibility in their charging behavior. With the fully meet their mobility needs. While current
support of a California Energy Commission policies have tried to address range anxiety,
grant, BMW will work with PG&E to explore new policies will need to be in place to capture
how charging can be moved throughout the grid opportunity of range freedom. The
the day and across the geographic areas higher range vehicles increase the flexibility in
that a vehicle travels within. In 2017 and charging vehicles, which can be better adapted
2018, BMW will work with a vehicle pool to meet the grids needs. The average vehicle
of over 250 vehicles, now including BMW is parked for most of the day and night, which
iPerformance PHEVs as well as BMW i3 provides numerous opportunities to move
and BMW i8 vehicles, to explore the grid charging so to avoid charging during hours
benefits of increasing charging flexibility as when the grid is strained or power is expensive,
well as customer engagement in advanced and increase charging when it is advantageous
charge management. to do so. At the same time, this greater range
could also reduce the predictability of electric
vehicle load.

PAGE 43
7 BMW ChargeForward Phase 2

To explore the value of this increased flexibility, BMW and Olivinewith support from PG&Ewill test
new ways to shift charging across both time and locations that respond to functionalities that will be
needed to address new challenges that the grid will face in the future. For example, the growth of
renewable generation is expected to create more dynamic conditions on the grid, that require more
options for grid operators to match load to unpredictable renewable generation. Figure 24 (on page 45)
shows the hourly aggregated load curve for the entire vehicle pool and three load requirements that
will be required to enable grid management under duck curve conditions:

Nighttime charging can Adding load in the Price signals can help to
be more beneficial if the afternoon can help address defer early-evening charging.
timer peak is eliminated the growth of solar.
and charging is allowed While not all EV drivers may be
to follow nighttime wind While the afternoon hours are sensitive to TOU price signals,
production. currently peak hours, the many are, and these signals and
steady increase in solar panels additional mechanisms such as
10,000 vehicles all following will soon require new afternoon power alerts could enhance the
timers at midnight would loads to maintain grid balance. ability of grid operators to defer
mean that 77 MW of power This need is likely to be EV charging from the early
would be instantly added to localized on specific circuits. evening when drivers arrive home
the grid. If this load is Circuits during these hours to their homes or apartments.
concentrated in urban areas, are likely to exhibit strongly Charging could be deferred to
it could increase the risk of different needs depending on late night or even to the following
grid instability. Nighttime their climate and solar day(s) depending on their battery
charging has a high degree adoptionthose circuits with charge level and travel patterns.
of flexibility, as the charging significant solar penetration These programs could be
dramatically drops around will need additional load, localized to address distribution
4 AM, hours before vehicles while those circuits without issues unique to specific
generally begin their morning solar will continue to need to neighborhoods and grid areas.
commute. reduce load.

PAGE 44
7 BMW ChargeForward Phase 2

The second phase of BMW ChargeForward In addition to testing the functional capabilities
will test functionality needed to address needed to provide advanced grid services,
these challenges. For example, BMW will BMW will also test different ways to engage
test the ability to move load from nighttime customers in ways that encourage drivers
charging to daytime charging. Doing so has to provide more charging flexibility to the
the potential to address excess solar supply grid. BMW will explore how different types of
that may occur in the future on PG&Es customer engagementincluding customer
grid. BMW will also test the ability of its incentives and customer performance data
charging control system to control charging can be used to encourage customers to
in response to price signals that reflect allow BMW greater flexibility in managing
dynamic generation prices or impacts on vehicle charging. BMW will also test how a
the distribution system. renewable energy signal, showing customers
how much renewable energy their vehicles
absorbed, can serve as an incentive to
encourage smart charging.

FIGURE 24 Hourly BMW i ChargeForward Aggregated Load (total kWh)

Against 2020 Estimated Duck Curve Grid Needs

The hourly aggregated load curve for Duck Curve Solar Trough Peak Nighttime
the entire vehicle pool and three load Grid Conditions Increase flexible Eliminate Charging
requirements that will be required to Desired charging charging to all Optimize to follow
enable grid management under duck outcome accommodate charging wind and avoid
more solar timer spike
curve conditions.

90
80
70
60
50
40
30
20
10
0
6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR
Non-modified load of i ChargeForward participants

PAGE 45
8 Conclusion

PAGE 46
8 Conclusion

The BMW i ChargeForward pilot Based on the successes in this phase of

demonstrated that vehicles can be an BMW i ChargeForward, BMW has been
effective grid resource, both from a technical awarded a California Energy Commission
perspective and customer engagement grant to continue with a phase 2 of BMW i
perspective. BMW and PG&E have been ChargeForward. As noted in Section 7, BMW
successful in responding to over 209 will be working with PG&E to explore how
demand response events over the course charging can be moved throughout the day
18 months. Load curtailment from vehicles and leverage the work completed to date
has been achieved with minimum disruption but expand the scope to identify methods
to customers. In our customer research, to manage charging across the geographic
98% of customers are satisfied in the project areas that a vehicle travels within. In 2017
(a 4 or 5 rating on a 5 point scale) and 93% and 2018, BMW will work with a vehicle pool
indicated they would participate in a similar of over 250 BMW i and BMW iPerformance
project if available in the future. vehicles to explore the grid benefits of
increasing charging flexibility as well as
The first Phase of the BMW i ChargeForward customer engagement in advanced charge
program halted vehicle charging when management.
called upon for up to one hour. By creating
opportunities for greater control over vehicle
charging, the program would increase the
benefits of smart charging. The grid can
benefit from starting vehicle charging when
the grid needs load, a condition the grid
will experience more frequently in the future
as renewables increase. This would allow
vehicles to support the grid during system
overgeneration events or absorb local
renewable generation where local circuits
might be adversely impacted.

PAGE 47
Appendix
FIGURE 25 Examples of Demand Response Event, Vehicle Pool

120
The figure displays the
performance of the
vehicle pool (Residential 100
BMW i3) customers
during a demand ~40 kW load reduction
response event on 80
(paused charging)
10/21/2015 from 8:00
9:00 PM. The colored
60
bands indicate individual
vehicles charging. The
orange column indicates 40
the demand response
event duration when
charging was delayed for 20
each customer charging
at that time.
kW

0
14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00
HOUR

FIGURE 26 Example of Demand Response Event, BMW Group 2nd Life Battery System

The figure displays the 120

performance of the microgrid
at the BMW Group technology
Office USA (including the 100
battery 2nd life system) during
a demand response event on
10/21/2015 from 8:009:00 80
PM. The orange band indicates
the demand response event
duration when power was 60
exported to meet the 100 kW
required obligation.
40

20
kW

-20
~60 kW power reduction
(2nd life battery export)
-40

-60

-80

-100
14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR

PAGE 48
 FIGURE 27 Example of Demand Response Event, Combined System

The figure below displays the performance of the combined system including the microgrid at the
BMW Group Technology Office USA (top right) as well as the vehicle pool (top left) during a demand
response event on 10/21/2015 from 8:009:00 PM. The chart at the bottom of the figure represents
the aggregated resource along with the baseline (blue line) computed using the 10-in-10 baseline
methodology. The orange band indicates the demand response event duration.

125 kW

100 kW

75 kW

50 kW

25 kW

+
0 kW

-25 kW
Vehicles + 2nd life
battery together provide
-50 kW 100kW DR Capacity Baseline
Targetload
kW

-75 kW Load
14:00

16:00

18:00

20:00

22:00

0:00

2:00

4:00

HOUR

PAGE 49
Content authors:
BMW Project Manager:
Sebastian Kaluza
[email protected]

PG&E Project Manager:

David Almeida
[email protected]

PG&E Assistant Project Manager:

Paige Mullen
[email protected]

Reporting Period:
01/201512/2016

PAGE 50