ArticlePDF Available

Quantification of Demand-Supply Balancing Capacity among Prosumers and Consumers: Community Self-Sufficiency Assessment for Energy Trading

Authors:

Abstract and Figures

With the increased adoption of distributed energy resources (DERs) and renewables, such as solar panels at the building level, consumers turn into prosumers with generation capability to supply their on-site demand. The temporal complementarity between supply and demand at the building level provides opportunities for energy exchange between prosumers and consumers towards community-level self-sufficiency. Investigating different aspects of community-level energy exchange in cyber and physical layers has received attention in recent years with the increase in renewables adoption. In this study, we have presented an in-depth investigation into the impact of energy exchange through the quantification of temporal energy deficit–surplus complementarity and its associated self-sufficiency capacities by considering the impact of variations in community infrastructure configurations, variations in household energy use patterns, and the potential for user adaptation for load flexibility. To this end, we have adopted a data-driven simulation using real-world data from a case-study neighborhood consisting of ~250 residential buildings in Austin, TX with a mix of prosumers and consumers and detailed data on decentralized DERs. By accounting for the uncertainties in energy consumption patterns across households, different levels of PV and energy storage integration, and different modalities of user adaptation, various scenarios of operations were simulated. The analysis showed that with PV integration of more than 75%, energy exchange could result in self-sufficiency for the entire community during peak generation hours from 11 a.m. to 3 p.m. However, there are limited opportunities for energy exchange during later times with PV-standalone systems. As a potential solution, leveraging building-level storage or user adaptation for load shedding/shifting during the 2-h low-generation timeframe (i.e., 5–7 p.m.) was shown to increase community self-sufficiency during generation hours by 17% and 5–10%, respectively, to 83% and 71–76%.
Content may be subject to copyright.
energies
Article
Quantification of Demand-Supply Balancing Capacity among
Prosumers and Consumers: Community Self-Sufficiency
Assessment for Energy Trading
Milad Afzalan 1,2 and Farrokh Jazizadeh 1, *


Citation: Afzalan, M.; Jazizadeh, F.
Quantification of Demand-Supply
Balancing Capacity among
Prosumers and Consumers:
Community Self-Sufficiency
Assessment for Energy Trading.
Energies 2021,14, 4318. https://
doi.org/10.3390/en14144318
Academic Editor: David Borge-Diez
Received: 24 May 2021
Accepted: 10 July 2021
Published: 17 July 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1INFORM Lab, Department of Civil and Environmental Engineering, Virginia Tech,
Blacksburg, VA 24061, USA; afzalan@vt.edu or milad.afzalan@engie.com
2ENGIE Impact, Boston, MA 02210, USA
*Correspondence: jazizade@vt.edu
Abstract:
With the increased adoption of distributed energy resources (DERs) and renewables, such
as solar panels at the building level, consumers turn into prosumers with generation capability
to supply their on-site demand. The temporal complementarity between supply and demand at
the building level provides opportunities for energy exchange between prosumers and consumers
towards community-level self-sufficiency. Investigating different aspects of community-level energy
exchange in cyber and physical layers has received attention in recent years with the increase in
renewables adoption. In this study, we have presented an in-depth investigation into the impact of
energy exchange through the quantification of temporal energy deficit–surplus complementarity
and its associated self-sufficiency capacities by considering the impact of variations in community
infrastructure configurations, variations in household energy use patterns, and the potential for user
adaptation for load flexibility. To this end, we have adopted a data-driven simulation using real-world
data from a case-study neighborhood consisting of ~250 residential buildings in Austin, TX with a
mix of prosumers and consumers and detailed data on decentralized DERs. By accounting for the
uncertainties in energy consumption patterns across households, different levels of PV and energy
storage integration, and different modalities of user adaptation, various scenarios of operations
were simulated. The analysis showed that with PV integration of more than 75%, energy exchange
could result in self-sufficiency for the entire community during peak generation hours from 11 a.m.
to 3 p.m. However, there are limited opportunities for energy exchange during later times with
PV-standalone systems. As a potential solution, leveraging building-level storage or user adaptation
for load shedding/shifting during the 2-h low-generation timeframe (i.e., 5–7 p.m.) was shown to
increase community self-sufficiency during generation hours by 17% and 5–10%, respectively, to 83%
and 71–76%.
Keywords:
net-zero community; demand–supply balance; self-sufficiency; PV; battery; decentralized
distribution; energy exchange; peer-2-peer (P2P)
1. Introduction
Managing distributed energy resources (DERs) in microgrids draws on opportunities
that are provided by elements such as communication technologies, metering devices,
controllable loads, renewable energy sources (e.g., solar panels), storage systems, and
human–building interaction for the improved operation of a power system. These elements
have paved the way for decentralized energy distribution and reduce the drawbacks of re-
lying on centralized systems [
1
]. The integration of small-scale distributed energy resources
at the household level has increased over recent years, enabling neighborhoods to operate
as microgrids. Consumers who adopt renewables such as solar panels become prosumers,
which can supply their own energy demand. Given the imbalanced and intermittent nature
Energies 2021,14, 4318. https://doi.org/10.3390/en14144318 https://www.mdpi.com/journal/energies
Energies 2021,14, 4318 2 of 21
of renewable energy generation and the uncertainty in consumer behavior, efficient man-
agement of the prosumers’ energy surplus is an important factor in renewable integration.
The surplus energy management could take different forms including supplying surplus
to a central grid, using energy storage systems, load shifting, or energy exchange. As
such, Peer-to-Peer (P2P) energy exchange is the next-generation approach for supply and
demand balancing, in which the collective participation of prosumers and consumers at
the community level could contribute to the success of DERs’ integration and decentralized
distribution [
2
]. Different factors in cyber, physical, and behavioral layers could play a
role in driving self-sufficiency (the ratio of self-consumed generated energy to the total
demand) of the community through energy exchange. Therefore, physical and behavioral
dynamics could affect the temporal balance between generation and consumption in the
form of surplus–deficit complementarity.
Achieving self-consumption is desired as reflected in policies such as decreased feed-
in-tariff rates [
3
5
]. As a general tendency, the match of on-site generation and demand is
inherently promoted [
6
]. In community-level local energy exchange, the goal is to cover the
demand of local consumers using the surplus from prosumers and small-scale DERs. How-
ever, balancing the surplus–deficit could be affected by the inherent temporal mismatch
between solar generation and demands that are driven by uncertain and volatile energy
consumption and generation patterns observed across different households [
7
10
]. The re-
search efforts on community-level energy exchange have investigated different dimensions
including energy-flow optimization [
11
], the impact of network constraints [
12
], and its
financial aspects [
13
]. However, in previous studies, little research has been conducted on
the impact of human factors in driving complementarity capacity. In this context, human
factors are represented by realistic and volatile demand-side energy profiles, as well as
load flexibility from user-adaptive behaviors that affect self-sufficiency capacities.
In this study, we have investigated how these human factors could affect the temporal
complementarity under the constraints of infrastructure configurations, specifically the
penetration level of small-scale DER assets—i.e., solar panels and storage systems at the
prosumer level. To this end, we have examined how the reshaping of energy profiles
through load flexibility/user adaptation impacts load-balancing capacities. Furthermore,
we have investigated how diverse and realistic household load profiles (representing energy
use behaviors) affect the temporal complementarity for self-sufficiency in communities
with different mixes of prosumers and consumers. These investigations have been guided
by defining varying levels of solar panels and energy storage penetration as distributed
DERs to provide an insight into the infrastructure impact. In answering these questions,
we have adopted a data-driven quantitative analysis of the load-balancing capacities
among prosumers and consumers at a neighborhood scale. To account for realistic profiles
of consumption and generation, data from a community of ~250 households in Austin,
TX—i.e., the Pecan Street project community and the Dataport dataset—has been used
as a case-study neighborhood to create simulated communities of different scales using
statistical sampling. The impact of the study variables has been assessed by measuring the
surplus–deficit complementarity and community self-sufficiency.
The contribution of this study includes: (1) Quantifying the temporal deficit–surplus
balancing capacity for community self-sufficiency assessment between prosumers and
consumers considering the variation of demand and supply using realistic load pro-
files, as well as varied infrastructure configurations, and (2) studying community self-
sufficiency by considering human adaptation for load flexibility. In turn, the findings
could help infrastructure planners evaluate the required share of DERs and the impact of
consumer-focused programs in future networks that rely on energy trading for decentral-
ized energy management.
2. Research Background
The energy exchange paradigm enables prosumers to have better control over energy
trading and contribute to a sustainable energy market [
2
,
14
]. Energy exchange increases the
Energies 2021,14, 4318 3 of 21
utilization of renewable assets in a decentralized way and reduces power loss due to shorter
transmission distances [
15
]. In recent years, several trials and real-world implementations
have adopted energy exchange prototypes through P2P trading (see [
16
,
17
] for a compre-
hensive list of projects). Several recent studies have investigated the economic aspects
of energy exchange through coordinating load balancing across
households [1820]
. For
instance, studies have focused on approaches that help participating households improve
their electricity cost [
19
]. Furthermore, maximizing profit based on the energy consumption
level of prosumers and consumers has been investigated [20].
In one direction of the research efforts, the impact of user preferences was studied to
improve users’ willingness in contributing to the community-level load balancing [
21
,
22
].
For example, in one study [
21
], an online platform was designed in which households could
engage in energy exchange and trading based on pricing incentives and households’ energy
storage rate. It was shown that households could be classified as participants desiring
economic benefits and participants desiring individual independence and state of storage.
Their investigation of more than 300 households showed that 77% of the community was
willing to participate in energy exchange.
In another class of studies, the impact of energy storage using batteries on load
balancing has been investigated [
19
,
23
,
24
]. Zepter et al. [
23
] have shown that the effect of
combining energy exchange with storage may result in a 60% cost reduction on electricity
bills. Alam et al. [
19
] have simulated the microgrid energy and trading values during
energy exchange. They have shown that combined PV–battery systems could result in
up to 74% cost savings. However, with the increase in PV–battery adoption, the saving
in the microgrid declined after a saturating point, since the extra generation could not
be used or stored at a given moment. Nguyen et al. [
24
] estimated the cost-saving for
energy exchange and showed that 28% of maximal saving can be achieved with PV–battery
systems. However, their findings also showed that the benefit of adopting a battery
is associated with the PV panel size, and the benefit is justified if the PV panel size is
sufficiently large.
Using data-driven methods, the concept of local demand-supply balancing has been
also studied (e.g., [
25
,
26
]). The study in [
25
] suggested the optimal number of PV systems
and the capacity of combined heat and power (CHP) to maximize the local demand
and supply balancing. In [
26
], a cooperative community was suggested, in which all
the buildings shift the operating times of controllable appliances (dishwasher, washing
machine, and dryer) to maximize solar energy utilization. A case study on five buildings
showed a 15% increase in the on-site generation that is covered by buildings. In contrast
to previous efforts, in this study we have investigated how the temporal load-balancing
capacities are affected by human factors including diversity in electricity use patterns
and flexibility in operating loads. In doing so, we have accounted for the impact of
DERs’ integration levels. Furthermore, we have evaluated these capacities by considering
community diversity through statistical sampling.
3. Methodology
Our methodology centered around simulating case-study communities by sampling
from a large-scale community of residential buildings with real energy use (demand) and
solar generation profiles. Surplus–deficit complementarity and load-balancing capacities
were quantified using the data-driven analysis of operations for different community real-
izations. To form diverse communities, groups of prosumers and consumers were sampled
using the bootstrapping technique to account for the uncertainty in daily energy use pro-
files. Furthermore, in cases that involved load flexibility by considering user adaptation, the
impact of load deferral or partial load shedding was simulated for individual households,
and daily load profiles were reconstructed through simulation to reflect user adaptation.
Differences in infrastructure configurations were reflected in the number of prosumers
and consumers with PV panels and storage systems. Upon quantifying the impact of each
variable of interest on the response from individual households, we analyzed the aggregate
Energies 2021,14, 4318 4 of 21
impact of simulated energy exchange across the community. In characterizing the energy
exchange capacity of a community, we assumed that all the instantaneous available surplus
energy could be used to supply the community deficit. The following sub-sections provide
more details on the variables, definitions, and methods of simulation and quantification.
3.1. Basic Definitions
Prosumers
(H
P
)
:
Electricity consumers with at least one type of renewable source of
energy are called prosumers—they can generate energy for their use. In this study, we
considered PV panels as the renewable source of energy. A household in the prosumer
group is shown as n HP.
Consumers
(H
C
)
:
Consumers are not equipped with local sources of energy and rely
on either a central market or their peer prosumers to buy their surplus generated energy. A
household in the consumer group is shown as n HC.
Surplus energy:
Surplus energy is available when PV generation at each time instance
exceeds the instantaneous demand for prosumers. Surplus energy could be transferred
in a trading market, stored in a battery for later consumption, or fed back to a central
aggregator. Given the surplus energy of a prosumer as s
n
, n
H
P
, the total surplus energy
of the community is SC=sn.
Deficit energy:
Deficit energy is the demand that is supplied either by central aggre-
gators or peers with surplus energy. Both consumers and prosumers could face a deficit
during a day. For prosumers, the deficit happens when demand exceeds their generation
or available storage. Given the unit deficit energy as d
n
, n
{H
C
, H
P
}, the community
deficit is DC=dn.
Net demand:
The net demand is the power drawn from the central aggregator to
cover the community deficit. The net demand for each household, L(t), at each time t is:
L(t)=P(t)G(t)B(t)(1)
in which P(t) is the power demand, G(t) is the PV generation, and B(t) is the battery power.
Here, P(t) > 0, G(t) > 0, and B(t) < 0 while charging and B(t) > 0 while discharging. Also,
L(t) < 0 contributes to the accumulation of surplus energy, while L(t) > 0 contributes to the
accumulation of deficit energy. The line losses were considered to be negligible and not
considered in this equation.
Complementarity factor:
The objective of energy exchange is to reduce the aggregate
energy deficit by using community energy surplus. We defined the complementarity
factor (CF) as the ratio of instantaneous energy deficit that can be covered by prosumers in
the community:
CF(%)=100
SC
DC
(2)
in which S
C
and D
C
are the community energy surplus and deficit, respectively. S
C
is
provided by prosumers with a negative net demand, while D
C
is needed by consumers
and/or prosumers with a positive net demand. A CF = 100% is ideal because it indicates
complete independence from a central aggregator and that demand is being met merely
through energy exchange. A value of CF > 100% indicates surplus beyond community
need, which can be stored or be fed to the central aggregator.
Self-sufficiency:
We extend the definition of energy self-sufficiency for individual
households [
27
] to a community to reflect solar energy utilization in the presence of energy
exchange. Self-sufficiency is defined as:
ϕss =Rt2
t1nHpM(t)dt
Rt2
t1n∈{Hc,Hp}P(t)dt (3)
Energies 2021,14, 4318 5 of 21
in which M(t) is the power generation utilized on-site as follows:
M(t)=min{P(t), G(t)+B(t)}(4)
[t
1
, t
2
] is the timeframe for measuring self-sufficiency. In this study, assuming energy
exchange during PV output, [t1, t2] is considered to cover all hours of PV generation. The
numerator in Equation (3) is the on-site generation by PV or storage saving and is measured
for the prosumer set (H
P
), and the denominator is the community energy demand that
is measured for both prosumer (H
P
) and consumer (H
C
) sets. Simply put, community
self-sufficiency is the share of PV generation that is directly consumed by the community
during generation hours.
3.2. Dynamic Energy Use Behavior at the Household Level
Electricity daily load profiles for residential buildings are known to have high varia-
tions across households and different days [
28
,
29
]. Although certain pre-defined (a limited
number) [30] or synthetic load profiles [31] can be employed to evaluate the impact of PV
adoption in different capacities, it could limit the generalization of findings by overlooking
the stochastic nature of household load profiles. Therefore, to present a reasonable estima-
tion of community load-balancing potential, it is essential to account for such uncertainties
and include diverse possibilities in load profiles, driven by occupants’ energy behavior.
Here, for simulating energy exchange scenarios, we use the daily profiles of real demand
and generation across two months for each household. Bootstrapping sampling was used
to form communities by selecting different combinations of households from the case-study
community. Therefore, to account for the variation of load profiles, we sampled a differ-
ent number of households ranging from 20 to 100 to form different communities. Each
community realization and its associated experiments were repeated 100 times to account
for the stochastic nature of load profiles in quantifying the energy exchange capacities. In
calculating energy from a daily load profile P(t), numerical integration in the timeframe of
interest [t1, t2] was used:
E=
t2
Z
t1
P(t)dt (5)
3.3. Dynamic Energy Use Behavior and Load Profile Change at the Household Level
User adaptation could play a significant role in creating complementarity capacity in
communities for the energy exchange. In this context, users can change their load profiles
and reschedule load operations to leverage the demand elasticity of flexible loads under
incentives such as economic benefits, which in turn could improve the self-sufficiency
of the community. The users’ control over energy trading in a local market provides
incentives for them to adapt for energy management, specifically if this adaptation is
coupled with automation. In Section 4.5, we have presented a discussion on how market
design and economic behavior could motivate user participation in energy exchange.
In the smart grid context, load flexibility can be achieved by user adaptation through
shedding load or shifting it by using smart loads or appliances that rely on automated
scheduling or predictive control capabilities. Therefore, flexibility through reshaping
demand profiles could result in increased complementarity. Specific loads with high
flexibility potentials [
8
,
32
,
33
] include air conditioning systems (ACs), electric vehicles
(EVs), and wet appliances (i.e., washing machine, dryer, and dishwasher). AC flexibility
can be achieved by leveraging the thermal storage capacity in buildings and changing
the temperature setpoint to reduce demand, while in deferrable appliances (EVs and wet
appliances) flexibility could be achieved by shifting the operation/charging time.
To account for the challenges of solar energy management during the off-peak gen-
eration period, we have investigated load flexibility and profile change in driving com-
plementarity capacity in the timeframes in which solar generation is declining. In doing
so, user adaptation and profile change was simulated using appliance-level energy data
Energies 2021,14, 4318 6 of 21
associated with each daily load profile in the dataset. The load profiles for EV charging
(if any) and wet appliances (if any) were represented as hourly profiles. For AC systems,
we simulated different strategies for load reduction when solar generation is declining.
These strategies reflect different control and user preference configurations. To this end, we
adopted an AC load peak reduction model based on the work by Hu et al. [
34
,
35
], in which
the energy variations from demand response (DR) operations, associated with increasing
thermostat setpoint by 1
C, 1
C (with pre-cooling), and 2
C (with pre-cooling), were
reflected on the actual AC load profiles during a 2 h timeframe when the solar generation
was declining. This DR approach is based on a grey-box RC thermal model that leverages
several parameters from building and the environment, including temperature variations,
solar gain, thermal mass, and building envelope characteristics. The RC model is character-
ized through a system identification process that uses different analytical and data-driven
optimization techniques for the objective of minimizing the error for indoor temperature
based on example temperature measurements. A simplified empirical energy consumption
model with an on/off control has been coupled with the thermal model to estimate AC
performance [
34
]. Using this model, Hu et al. [
34
] showed that AC power reductions of
25%, 31%, and 68% could be achieved during the DR timeframe for an increase in setpoint
of 1
C, 1
C (with pre-cooling), and 2
C (with pre-cooling), respectively. We measured
the aggregate variation of the complementarity using metrics described in Section 3.1 to
account for varying levels of engagement by considering the participation of a subset of
prosumers/consumers who were willing to take adaptive actions for flexible operations.
3.4. Battery Modeling
Given the temporal mismatch of generation and demand, storage systems for pro-
sumers will create capacities for self-sufficiency and complementarity. Therefore, we also
integrated storage systems in our simulations using the specifications of the commercial
Tesla Powerwall [
36
] as the most widely installed system in the US, with a capacity of
13.5 kW h and 7 kW peak power. Taking into account the average daily solar energy
generation per prosumer in our case-study community is around 15 kW h, we assumed
one battery per house in our simulations. Considering the goal of capacity quantification,
we adopted a battery charging schedule as shown in Figure 1for each prosumer unit. The
batteries are charged at the time of surplus energy according to their physical constraints
and are discharged during the deficit. Prosumers’ surplus energy that exceeded the battery
capacity or was stored at the battery (available at later times after covering the prosumer’s
demand) was used in energy exchange. We have adopted this intuitive model for battery
scheduling to show the impact of storage in surplus–deficit complementarity and to offer
the extra surplus stored in the battery during peak demand when solar generation is
declining. However, more sophisticated battery scheduling schemes under the constraint
of dynamic pricing, cycle life, and losses could be used as well [37].
Energies 2021,14, 4318 7 of 21
Energies 2021, 14, x FOR PEER REVIEW 7 of 21
Figure 1. Battery scheduling flowchart.
4. Results and Discussion
4.1. Case-Study Community Characteristics
We used the historical energy dataset of 244 residential households located in Austin,
TX for our experiments. The dataset is available through the Pecan Street Project [38]. The
community included 119 prosumers with PV systems and 125 consumers. Since our case
study focused on the PV rooftop solar panels as the renewable source, we used the data
for July and August of 2015, as the representative summer months. Similar to the resolu-
tion of smart meters, 15-min resolution data for both demand and generation was used.
Among the generation time series, less than 0.001% of profiles (8 out of 7152) had missing
information and were eliminated from the dataset, resulting in 7144 and 7492 daily pro-
files for prosumers and consumers, respectively. Each daily profile included 96 data
points, which were annotated with a ‘household ID’ and ‘day of the year’ index for infor-
mation retrieval. Figure 2 shows the daily load profiles of net demand for an example
consumer (Figure 2a) and an example prosumer (Figure 2b) for 20 successive days as a
case of complementarity potential in the case-study community. Furthermore, Figure 3
shows the characterization of solar generation for all the prosumers in the community—
Figure 3a, all solar generation profiles; and Figure 3b, the energy generation distribution
from 9 am to 7 pm, which is the timeframe of our study. The median PV peak power was
3.95 kW (5th and 95th quantile of 1.8 kW, 6.5 kW), with the highest PV peak of 11.0 kW.
Figure 1. Battery scheduling flowchart.
4. Results and Discussion
4.1. Case-Study Community Characteristics
We used the historical energy dataset of 244 residential households located in Austin,
TX for our experiments. The dataset is available through the Pecan Street Project [
38
].
The community included 119 prosumers with PV systems and 125 consumers. Since
our case study focused on the PV rooftop solar panels as the renewable source, we used
the data for July and August of 2015, as the representative summer months. Similar to
the resolution of smart meters, 15-min resolution data for both demand and generation
was used. Among the generation time series, less than 0.001% of profiles (8 out of 7152)
had missing information and were eliminated from the dataset, resulting in 7144 and
7492 daily profiles for prosumers and consumers, respectively. Each daily profile included
96 data points, which were annotated with a ‘household ID’ and ‘day of the year’ index for
information retrieval. Figure 2shows the daily load profiles of net demand for an example
consumer (Figure 2a) and an example prosumer (Figure 2b) for 20 successive days as a case
of complementarity potential in the case-study community. Furthermore, Figure 3shows
the characterization of solar generation for all the prosumers in the community—Figure 3a,
all solar generation profiles; and Figure 3b, the energy generation distribution from 9 a.m.
to 7 p.m., which is the timeframe of our study. The median PV peak power was 3.95 kW
(5th and 95th quantile of 1.8 kW, 6.5 kW), with the highest PV peak of 11.0 kW.
Energies 2021,14, 4318 8 of 21
Energies 2021, 14, x FOR PEER REVIEW 8 of 21
Figure 2. Net demand complementarity for energy trading: (a) a consumer’s daily profiles with def-
icit energy (+ net values), (b) a prosumer’s daily profiles with surplus energy (− net values).
Figure 3. Solar generation patterns in the case-study community: (a) PV power profiles, (b) daily
averaged PV-generated energy.
To evaluate prosumers’ capacity to offer surplus for energy exchange, Figure 4 com-
pares prosumers and consumers’ energy demand during the generation timeframe. In
Figure 4a, the median net energy demand for prosumers is −0.6 kW h (5th, 25th, 75th, and
95th percentile of −18 kW h, −8 kW h, 8 kW h, and 27 kW h). This shows that prosumers
in general have the capacity for surplus exchange. In Figure 4b, it is shown that the
prosumers’ energy demand (without considering PV generation), is on average higher
than that of consumers. The median energy demand is 25 kW h (5th, 25th, 75th, and 95th
percentile of 8 kW h, 18 kW h, 33 kW h, and 52 kW h) and 14 kW h (5th, 25th, 75th, and
95th percentile of 3 kW h, 7 kW h, 24 kW h, and 42 kW h) for prosumers and consumers,
respectively. The higher energy demand of prosumers could be associated with factors
such as household size, plug-in EV ownership, or behavioral tendencies due to the avail-
ability of free energy sources. Figure 5 presents the variation of total daily energy con-
sumption across households using two months of daily profiles. In Figure 5a, 53% of
prosumers have a daily negative net energy demand (i.e., surplus to offer for exchange).
Figure 5b shows net energy for consumers with considerable variations in net demand
across households.
ab
Energy demand (kWh)
Time (hr)
Power (kW)
Time (hr)
Figure 2.
Net demand complementarity for energy trading: (
a
) a consumer’s daily profiles with
deficit energy (+ net values), (b) a prosumer’s daily profiles with surplus energy (net values).
Energies 2021, 14, x FOR PEER REVIEW 8 of 21
Figure 2. Net demand complementarity for energy trading: (a) a consumer’s daily profiles with def-
icit energy (+ net values), (b) a prosumer’s daily profiles with surplus energy (− net values).
Figure 3. Solar generation patterns in the case-study community: (a) PV power profiles, (b) daily
averaged PV-generated energy.
To evaluate prosumers’ capacity to offer surplus for energy exchange, Figure 4 com-
pares prosumers and consumers’ energy demand during the generation timeframe. In
Figure 4a, the median net energy demand for prosumers is −0.6 kW h (5th, 25th, 75th, and
95th percentile of −18 kW h, −8 kW h, 8 kW h, and 27 kW h). This shows that prosumers
in general have the capacity for surplus exchange. In Figure 4b, it is shown that the
prosumers’ energy demand (without considering PV generation), is on average higher
than that of consumers. The median energy demand is 25 kW h (5th, 25th, 75th, and 95th
percentile of 8 kW h, 18 kW h, 33 kW h, and 52 kW h) and 14 kW h (5th, 25th, 75th, and
95th percentile of 3 kW h, 7 kW h, 24 kW h, and 42 kW h) for prosumers and consumers,
respectively. The higher energy demand of prosumers could be associated with factors
such as household size, plug-in EV ownership, or behavioral tendencies due to the avail-
ability of free energy sources. Figure 5 presents the variation of total daily energy con-
sumption across households using two months of daily profiles. In Figure 5a, 53% of
prosumers have a daily negative net energy demand (i.e., surplus to offer for exchange).
Figure 5b shows net energy for consumers with considerable variations in net demand
across households.
ab
Energy demand (kWh)
Time (hr)
Power (kW)
Time (hr)
Figure 3.
Solar generation patterns in the case-study community: (
a
) PV power profiles, (
b
) daily
averaged PV-generated energy.
To evaluate prosumers’ capacity to offer surplus for energy exchange, Figure 4com-
pares prosumers and consumers’ energy demand during the generation timeframe. In
Figure 4a, the median net energy demand for prosumers is
0.6 kW h (5th, 25th, 75th,
and 95th percentile of
18 kW h,
8 kW h, 8 kW h, and 27 kW h). This shows that
prosumers in general have the capacity for surplus exchange. In Figure 4b, it is shown
that the prosumers’ energy demand (without considering PV generation), is on average
higher than that of consumers. The median energy demand is 25 kW h (5th, 25th, 75th,
and 95th percentile of 8 kW h, 18 kW h, 33 kW h, and 52 kW h) and 14 kW h (5th, 25th,
75th, and 95th percentile of 3 kW h, 7 kW h, 24 kW h, and 42 kW h) for prosumers and
consumers, respectively. The higher energy demand of prosumers could be associated with
factors such as household size, plug-in EV ownership, or behavioral tendencies due to the
availability of free energy sources. Figure 5presents the variation of total daily energy
consumption across households using two months of daily profiles. In Figure 5a, 53% of
prosumers have a daily negative net energy demand (i.e., surplus to offer for exchange).
Figure 5b shows net energy for consumers with considerable variations in net demand
across households.
Energies 2021,14, 4318 9 of 21
Energies 2021, 14, x FOR PEER REVIEW 9 of 21
(a) (b)
Figure 4. Distribution of (a) net energy demand and (b) energy demand between 9 am and 7 pm for the entire community.
(a)
(b)
Figure 5. Distribution of net energy for (a) prosumers and (b) consumers from 9 am to 7 pm for the entire community.
According to the U.S. Energy Information Administration, the daily average con-
sumption for houses in Texas is 39 kW h with an average hourly power demand of ~1.6
kW, which is higher than the average demand in the United States (that is, 1.22 kW). The
average power demand for the prosumers and consumers in the case-study community
during the solar generation hours (9 am to 7 pm) were 2.5 kW and 1.4 kW, respectively.
In the absence of other representative metadata such as building size, occupancy infor-
mation, or other socio-economic factors for the state of Texas or the sampled case-study
community, the energy efficiency of the community appears to follow the trends in the
state of Texas.
Energy use patterns’ variability and uncertainty in the case-study community: In-
dividual households had varied energy use profiles on different days and these variations
could affect the reliability of energy exchange. Using K-mean clustering, in Figure 6a,b we
have presented the clusters of net energy profiles for prosumers and consumers. The
power profiles were normalized based on their daily maximum values as follows:
p
(
t
)
=
(
)

(
(
)
)
,
t
{
1
,
.
.
.
,
96
}
(6)
where p(t) is the power at timestamp t, max(p(t)) is the maximum power observed over
a day, and p(t)
is the normalized power.
Energy demand (kWh)
Figure 4.
Distribution of (
a
) net energy demand and (
b
) energy demand between 9 a.m. and 7 p.m.
for the entire community.
Energies 2021, 14, x FOR PEER REVIEW 9 of 21
(a) (b)
Figure 4. Distribution of (a) net energy demand and (b) energy demand between 9 am and 7 pm for the entire community.
(a)
(b)
Figure 5. Distribution of net energy for (a) prosumers and (b) consumers from 9 am to 7 pm for the entire community.
According to the U.S. Energy Information Administration, the daily average con-
sumption for houses in Texas is 39 kW h with an average hourly power demand of ~1.6
kW, which is higher than the average demand in the United States (that is, 1.22 kW). The
average power demand for the prosumers and consumers in the case-study community
during the solar generation hours (9 am to 7 pm) were 2.5 kW and 1.4 kW, respectively.
In the absence of other representative metadata such as building size, occupancy infor-
mation, or other socio-economic factors for the state of Texas or the sampled case-study
community, the energy efficiency of the community appears to follow the trends in the
state of Texas.
Energy use patterns’ variability and uncertainty in the case-study community: In-
dividual households had varied energy use profiles on different days and these variations
could affect the reliability of energy exchange. Using K-mean clustering, in Figure 6a,b we
have presented the clusters of net energy profiles for prosumers and consumers. The
power profiles were normalized based on their daily maximum values as follows:
p
(
t
)
=
(
)

(
(
)
)
,
t
{
1
,
.
.
.
,
96
}
(6)
where p(t) is the power at timestamp t, max(p(t)) is the maximum power observed over
a day, and p(t)
is the normalized power.
Energy demand (kWh)
Figure 5. Distribution of net energy for (a) prosumers and (b) consumers from 9 a.m. to 7 p.m. for the entire community.
According to the U.S. Energy Information Administration, the daily average consump-
tion for houses in Texas is 39 kW h with an average hourly power demand of
~1.6 kW
,
which is higher than the average demand in the United States (that is, 1.22 kW). The average
power demand for the prosumers and consumers in the case-study community during
the solar generation hours (9 a.m. to 7 p.m.) were 2.5 kW and 1.4 kW, respectively. In the
absence of other representative metadata such as building size, occupancy information, or
other socio-economic factors for the state of Texas or the sampled case-study community,
the energy efficiency of the community appears to follow the trends in the state of Texas.
Energy use patterns’ variability and uncertainty in the case-study community
: In-
dividual households had varied energy use profiles on different days and these variations
could affect the reliability of energy exchange. Using K-mean clustering, in Figure 6a,b
we have presented the clusters of net energy profiles for prosumers and consumers. The
power profiles were normalized based on their daily maximum values as follows:
p(t)=p(t)
max(p(t)) , t {1, . . . , 96}(6)
where p(t) is the power at timestamp t, max(p(t)) is the maximum power observed over a
day, and p(t)is the normalized power.
Energies 2021,14, 4318 10 of 21
Energies 2021, 14, x FOR PEER REVIEW 10 of 21
Figure 6. Clustered net energy profiles of prosumers and consumers and their entropy distribution: (a) clusters of prosum-
ers’ daily profiles, (b) clusters of consumers’ daily profiles, (c) entropy of prosumer households, and (d) entropy of con-
sumer households. Values in the legend for subplot (a) and (b) show the frequencies of clusters in the community.
As shown in Figure 6, except for cluster two with 20% frequency and no surplus en-
ergy, in the prosumer group all clusters showed some surplus energy with different peak
levels. Cluster 3 had the highest potential for energy exchange due to its sharp surplus
peak during PV generation in addition to its moderately low demand peak when PV gen-
eration diminished. On the consumer side (Figure 6b), cluster two was a suitable candi-
date for energy exchange due to its high demand during the peak of PV generation, while
cluster three (with uniform consumption), and clusters one and five (with peak demand
around 6:00 pm) offered some potential for energy exchange. Cluster four had a peak
around 11 pm and fit better for storage-based energy exchange.
To quantify the uncertainty associated with consistent patterns of energy use for each
household, the entropy of a household, En, was calculated as:
E
=
P
(
C
)
log
P
(
C
)
(7)
in which P(Ci) is the probability of observing cluster i, and K is the total number of clusters.
A low value of En indicates the higher stability of households’ load profiles across different
days, while a higher En denotes low predictability. The lowest value for En is zero when
all daily load profiles of a household belong to one cluster. Figure 6c,d shows the distri-
bution of entropy versus average daily net energy for prosumers and consumers, and each
data point represents one household. The horizontal/vertical dashed lines reflect the 25th
and 75th percentile of the values. For each of the nine areas, the case-study community
included a variety of households with low to high predictability and energy demand,
which further reflects the varied range of households with different energy use behaviors.
4.2. Baseline Surplus–Deficit Temporal Complementarity Quantification
As the baseline case, we investigated the temporal complementarity capacity in com-
munities of different sizes and with varied energy use patterns without considering stor-
age systems or load flexibility.
Equal prosumer/consumer size: We simulated sampled communities of N = {20, 40,
60, 80, 100} households with the same number of prosumers and consumers—N⁄2 for each
set. Using the complementarity factor (Equation (2)), we measured the extent of temporal
energy balancing capacity at the community level. For each community size and each hour
Figure 6. Clustered net energy profiles of prosumers and consumers and their entropy distribution:
(
a
) clusters of prosumers’ daily profiles, (
b
) clusters of consumers’ daily profiles, (
c
) entropy of
prosumer households, and (
d
) entropy of consumer households. Values in the legend for subplot
(a,b) show the frequencies of clusters in the community.
As shown in Figure 6, except for cluster two with 20% frequency and no surplus
energy, in the prosumer group all clusters showed some surplus energy with different peak
levels. Cluster 3 had the highest potential for energy exchange due to its sharp surplus peak
during PV generation in addition to its moderately low demand peak when PV generation
diminished. On the consumer side (Figure 6b), cluster two was a suitable candidate for
energy exchange due to its high demand during the peak of PV generation, while cluster
three (with uniform consumption), and clusters one and five (with peak demand around
6:00 p.m.) offered some potential for energy exchange. Cluster four had a peak around
11 p.m. and fit better for storage-based energy exchange.
To quantify the uncertainty associated with consistent patterns of energy use for each
household, the entropy of a household, En, was calculated as:
En=
K
i=1
P(Ci)log(P(Ci)) (7)
in which P(C
i
) is the probability of observing cluster i, and K is the total number of clusters.
A low value of E
n
indicates the higher stability of households’ load profiles across different
days, while a higher E
n
denotes low predictability. The lowest value for E
n
is zero when all
daily load profiles of a household belong to one cluster. Figure 6c,d shows the distribution
of entropy versus average daily net energy for prosumers and consumers, and each data
point represents one household. The horizontal/vertical dashed lines reflect the 25th and
75th percentile of the values. For each of the nine areas, the case-study community included
a variety of households with low to high predictability and energy demand, which further
reflects the varied range of households with different energy use behaviors.
4.2. Baseline Surplus–Deficit Temporal Complementarity Quantification
As the baseline case, we investigated the temporal complementarity capacity in
communities of different sizes and with varied energy use patterns without considering
storage systems or load flexibility.
Equal prosumer/consumer size:
We simulated sampled communities of N = {20, 40,
60, 80, 100} households with the same number of prosumers and consumers—N
2 for each
set. Using the complementarity factor (Equation (2)), we measured the extent of temporal
Energies 2021,14, 4318 11 of 21
energy balancing capacity at the community level. For each community size and each
hour of PV generation, 100 experiments, representing 100 simulated communities, were
performed. Figure 7presents the variation of prosumers’ net energy, consumers’ demand,
and the complementarity factor for three successive hours (2–5 p.m.) as examples. This
figure shows that increasing the community size resulted in a linear change in net energy
and demand with R
2
> 0.98 for all cases. Therefore, the complementarity factor (CF) for
various community sizes remains almost constant with an std < 1.5 for all three subplots.
Moving from 2–3 p.m. (the highest PV generation timeframe) to 3–4 p.m. and 4–5 p.m., CF
reduced from an average of ~ 50.0% to 31.8% and 15.5%, respectively. This reduction is
associated with both the decline of PV generation and the considerable increase in demand
from prosumers and consumers at the hours stretching into the evening. As illustrated,
for 4–5 p.m., average prosumers’ net energy is positive. Therefore, unlike 2–4 p.m., the
aggregate surplus energy offered by a subset of prosumers is not sufficient to cancel out
the deficit of other households.
Energies 2021, 14, x FOR PEER REVIEW 11 of 21
of PV generation, 100 experiments, representing 100 simulated communities, were per-
formed. Figure 7 presents the variation of prosumers’ net energy, consumers’ demand,
and the complementarity factor for three successive hours (2–5 pm) as examples. This fig-
ure shows that increasing the community size resulted in a linear change in net energy
and demand with R2 > 0.98 for all cases. Therefore, the complementarity factor (CF) for
various community sizes remains almost constant with an std < 1.5 for all three subplots.
Moving from 2–3 pm (the highest PV generation timeframe) to 3–4 pm and 4–5 pm, CF
reduced from an average of ~ 50.0% to 31.8% and 15.5%, respectively. This reduction is
associated with both the decline of PV generation and the considerable increase in de-
mand from prosumers and consumers at the hours stretching into the evening. As illus-
trated, for 4–5 pm, average prosumers’ net energy is positive. Therefore, unlike 2–4 pm,
the aggregate surplus energy offered by a subset of prosumers is not sufficient to cancel
out the deficit of other households.
Figure 7. Comparison of surplus energy, deficit energy, and complementarity factor (CF) for various community sizes
with an equal number of prosumers and consumers at (a) 2–3 pm, (b) 3–4 pm, and (c) 4–5 pm. The left axis represents the
net energy values, and the right axis represents CF values.
Table 1 shows the outcome of simulated load-balancing capacity for each hour of the
entire solar energy generation timeframe. The highest CF happens at 12–1 pm, in which,
on average, prosumers’ surplus is 72.5% of the entire community’s demand. However,
before 10 am and after 4 pm, net energy complementarity decreases considerably. For 5–
6 pm and 6–7 pm, the CF is less than 5% and 1%, respectively, indicating almost no base-
line potential for energy exchange. Apart from these hours, the CF varies between 15.5%
(3–4 pm) and 62.9% (11–12 pm) with a standard deviation of 24%. Therefore, in commu-
nities with an equal distribution of prosumers and consumers and without any storage or
user adaptation capacities, during the maximum generation hours, ~70% of demand could
be covered by community surplus. Energy exchange could be used to maximize commu-
nity self-consumption and avoid trade-ins.
Varied prosumers’ ratios (PR): Given the linear dependence of load-balancing ca-
pacities on the community size (N), we investigated the impact of varied prosumers’ ratios
(PRs) for a community of N = 20 with 100 repetitions. Figure 8 presents CF from 2 to 5 pm
and for PRs (i.e., PV integration ratio) of {25%, 50%, 75%, 100%}. From 2 to 3 pm, with 75%
and 100% PRs, the CF exceeded 100% (109% and 274%, respectively), indicating full com-
plementarity capacities plus excess surplus for storage or trade-in. For 25% and 50% PRs,
the median of CF is 19% and 56%, respectively. From 3 to 4 pm, the CF exceeds 100%
(149%) only with 100% PR, while for 25%, 50%, and 75% PRs, 13%, 33%, and 74% of the
community demand can be supplied by prosumers. However, from 4 to 5 pm, the CF only
reaches 50% with 100% PR, while lowering PV integration further limits the complemen-
tarity in the community (CF of only 7% with 25% PR).
20 40 50 60 100
Community size
-50
0
50
100
Net energy (kWh)
0
20
40
60
80
100
Complementarity factor (%)
Prosumers
Consumers
20 40 50 60 100
Community size
0
40
80
120
Net energy (kWh)
0
20
40
60
80
100
Complementarity factor (%)
Prosumers
Consumers
20 40 50 60 100
Community size
0
40
80
120
Net energy (kWh)
0
20
40
60
80
100
Complementarity factor (%)
Prosumers
Consumers
abc
Figure 7.
Comparison of surplus energy, deficit energy, and complementarity factor (CF) for various community sizes with
an equal number of prosumers and consumers at (
a
) 2–3 p.m., (
b
) 3–4 p.m., and (
c
) 4–5 p.m. The left axis represents the net
energy values, and the right axis represents CF values.
Table 1shows the outcome of simulated load-balancing capacity for each hour of the
entire solar energy generation timeframe. The highest CF happens at 12–1 p.m., in which,
on average, prosumers’ surplus is 72.5% of the entire community’s demand. However,
before 10 a.m. and after 4 p.m., net energy complementarity decreases considerably. For
5–6 p.m. and 6–7 p.m., the CF is less than 5% and 1%, respectively, indicating almost no
baseline potential for energy exchange. Apart from these hours, the CF varies between
15.5% (3–4 p.m.) and 62.9% (11–12 p.m.) with a standard deviation of 24%. Therefore, in
communities with an equal distribution of prosumers and consumers and without any
storage or user adaptation capacities, during the maximum generation hours, ~70% of
demand could be covered by community surplus. Energy exchange could be used to
maximize community self-consumption and avoid trade-ins.
Varied prosumers’ ratios (PR):
Given the linear dependence of load-balancing capac-
ities on the community size (N), we investigated the impact of varied prosumers’ ratios
(PRs) for a community of N = 20 with 100 repetitions. Figure 8presents CF from 2 to 5 p.m.
and for PRs (i.e., PV integration ratio) of {25%, 50%, 75%, 100%}. From 2 to 3 p.m., with
75% and 100% PRs, the CF exceeded 100% (109% and 274%, respectively), indicating full
complementarity capacities plus excess surplus for storage or trade-in. For 25% and 50%
PRs, the median of CF is 19% and 56%, respectively. From 3 to 4 p.m., the CF exceeds
100% (149%) only with 100% PR, while for 25%, 50%, and 75% PRs, 13%, 33%, and 74%
of the community demand can be supplied by prosumers. However, from 4 to 5 p.m.,
the CF only reaches 50% with 100% PR, while lowering PV integration further limits the
complementarity in the community (CF of only 7% with 25% PR).
Energies 2021,14, 4318 12 of 21
Table 1.
Energy surplus, energy deficit, and complementarity factor for various community sizes with equal numbers of
prosumers and consumers.
Time of Day Community Size Prosumer Net Energy (kW h) *,** Consumer Net Energy (kW h) * CF (%) *
9 a.m.–10 a.m.
20 3.9 10.1 17.2
40 8.6 22.1 14.8
60 12.0 33.0 15.3
80 16.4 42.6 15.6
100 20.0 54.3 15.1
10 a.m.–11
a.m.
20 2.6 13.0 45.7
40 3.8 23.5 43.1
60 6.1 36.0 41.5
80 9.4 48.4 42.5
100 10.1 60.3 41.2
11 a.m.–12
p.m.
20 6.0 13.7 62.2
40 13.6 27.5 63.9
60 21.5 41.5 64.6
80 28.2 54.8 63.7
100 34.1 71.6 59.8
12 p.m.–1 p.m.
20 10.2 15.3 79.3
40 19.6 33.2 69.6
60 29.4 47.1 71.7
80 39.8 62.6 72.1
100 48.9 78.9 69.9
1 p.m.–2 p.m.
20 8.5 18.8 61.4
40 17.3 35.2 61.8
60 25.0 55.2 57.8
80 32.6 72.1 57.3
100 42.3 89.4 58.4
2 p.m.–3 p.m.
20 6.6 18.9 52.2
40 12.6 37.1 49.9
60 21.1 58.2 50.2
80 27.0 76.4 49.6
100 32.7 95.6 48.3
3 p.m.–4 p.m.
20 1.5 20.2 32.7
40 3.8 41.5 31.9
60 5.3 63.2 30.8
80 7.6 81.2 32.1
100 10.0 103.2 31.6
4 p.m.–5 p.m.
20 5.6 20.7 16.2
40 9.0 44.8 16.1
60 17.5 64.8 15.0
80 22.4 86.9 15.2
100 27.4 109.1 15.2
5 p.m.–6 p.m.
20 14.3 23.3 4.9
40 29.4 48.1 4.5
60 42.9 70.7 4.6
80 57.8 94.9 4.6
100 71.1 117.2 4.8
6 p.m.–7 p.m.
20 23.8 25.2 0.8
40 48.0 51.4 0.8
60 70.0 75.6 0.9
80 94.1 101.0 0.9
100 117.9 125.3 0.9
Average 35.6
* Values for each cell in these columns reflect the average of 100 experiments. ** Values for each cell in these columns reflect the aggregation
of negative net energy (the numerator of Equation (2)) and positive net energy (prosumers’ deficit contributor in Equation (2)) from
individual prosumers.
Energies 2021,14, 4318 13 of 21
Energies 2021, 14, x FOR PEER REVIEW 13 of 21
Figure 8. Comparison of complementarity factor (CF) for different of prosumers’ ratios (PR) in communities of N = 20 at
(a) 2–3 pm, (b) 3–4 pm, and (c) 4–5 pm.
Figure 9 shows a sample distribution of CF from 2 to 3 pm for a PR of 0.25—first case
in Figure 8a. A two-sample Kolmogorov–Smirnov (KS) test indicated a normal distribu-
tion (p-value = 0.89). The KS test for all possible scenarios showed p-values of higher than
0.05 except for one case. Therefore, in presenting the results, we have used the average of
the 100 repeated experiments.
Figure 9. Histogram of complementarity factor at 2–3 pm for prosumers’ ratio of 0.25.
Figure 10 compares the average CF values and their 95% confidence interval for all
PRs during the hours of PV generation. Due to the high variation across different PRs and
hours of the day, the y-axis is shown on a logarithmic scale. For all the scenarios, it was
observed that from 10 am to 4 pm, the aggregate net energy of prosumers was negative
(i.e., indicating surplus energy), with a high potential for energy exchange. For 25% and
50% PRs, the highest CF values were 27% and 75% from 12 to 1 pm. The 75% and 100%
PRs are the only conditions that provide capacities for community self-consumption with
CF values of more than 100% from around 10 am to 3 pm, with the highest values of 165%
and 420% from 12 to 1 pm, respectively. Furthermore, from 11 am to 3 pm, with high PRs,
energy storage could provide opportunities for increased complementarity and self-suffi-
ciency after 3 pm—on average the remaining surplus was 30% (with 75% PR) and 217%
(with 100% PR) of the community demand from 11 am to 3 pm. However, for PRs lower
than 75%, the complementarity of the loads could ideally enable renewables’ self-con-
sumption without the need for storage or supply to the grid. After 4 pm, the aggregate
prosumers’ net energy is positive, which considerably limits the complementarity. There-
fore, even with the 100% PR, without energy storage only 12% and 2% of the community
deficit could be supplied by PV generation. Figure 10 summarizes the community capac-
ities for temporal self-sufficiency considering different PV penetration ratios. It must be
noted that simulations were carried out by integrating real load profiles to report these
25 50 75 100
Prosumer ratio (%)
0
100
200
400
600
Complementarity factor (%)
Experiment
Median
25 50 75 100
Prosumer ratio (%)
0
50
100
150
200
250
300
350
Complementarity factor (%)
Experiment
Median
25 50 75 100
Prosumer ratio (%)
0
20
40
60
80
100
Complementarity factor (%)
Experiment
Median
bc
a
Figure 8.
Comparison of complementarity factor (CF) for different of prosumers’ ratios (PR) in communities of N = 20 at (
a
)
2–3 p.m., (b) 3–4 p.m., and (c) 4–5 p.m.
Figure 9shows a sample distribution of CF from 2 to 3 p.m. for a PR of 0.25—first case
in Figure 8a. A two-sample Kolmogorov–Smirnov (KS) test indicated a normal distribution
(p-value = 0.89). The KS test for all possible scenarios showed p-values of higher than 0.05
except for one case. Therefore, in presenting the results, we have used the average of the
100 repeated experiments.
Energies 2021, 14, x FOR PEER REVIEW 13 of 21
Figure 8. Comparison of complementarity factor (CF) for different of prosumers’ ratios (PR) in communities of N = 20 at
(a) 2–3 pm, (b) 3–4 pm, and (c) 4–5 pm.
Figure 9 shows a sample distribution of CF from 2 to 3 pm for a PR of 0.25—first case
in Figure 8a. A two-sample Kolmogorov–Smirnov (KS) test indicated a normal distribu-
tion (p-value = 0.89). The KS test for all possible scenarios showed p-values of higher than
0.05 except for one case. Therefore, in presenting the results, we have used the average of
the 100 repeated experiments.
Figure 9. Histogram of complementarity factor at 2–3 pm for prosumers’ ratio of 0.25.
Figure 10 compares the average CF values and their 95% confidence interval for all
PRs during the hours of PV generation. Due to the high variation across different PRs and
hours of the day, the y-axis is shown on a logarithmic scale. For all the scenarios, it was
observed that from 10 am to 4 pm, the aggregate net energy of prosumers was negative
(i.e., indicating surplus energy), with a high potential for energy exchange. For 25% and
50% PRs, the highest CF values were 27% and 75% from 12 to 1 pm. The 75% and 100%
PRs are the only conditions that provide capacities for community self-consumption with
CF values of more than 100% from around 10 am to 3 pm, with the highest values of 165%
and 420% from 12 to 1 pm, respectively. Furthermore, from 11 am to 3 pm, with high PRs,
energy storage could provide opportunities for increased complementarity and self-suffi-
ciency after 3 pm—on average the remaining surplus was 30% (with 75% PR) and 217%
(with 100% PR) of the community demand from 11 am to 3 pm. However, for PRs lower
than 75%, the complementarity of the loads could ideally enable renewables’ self-con-
sumption without the need for storage or supply to the grid. After 4 pm, the aggregate
prosumers’ net energy is positive, which considerably limits the complementarity. There-
fore, even with the 100% PR, without energy storage only 12% and 2% of the community
deficit could be supplied by PV generation. Figure 10 summarizes the community capac-
ities for temporal self-sufficiency considering different PV penetration ratios. It must be
noted that simulations were carried out by integrating real load profiles to report these
25 50 75 100
Prosumer ratio (%)
0
100
200
400
600
Complementarity factor (%)
Experiment
Median
25 50 75 100
Prosumer ratio (%)
0
50
100
150
200
250
300
350
Complementarity factor (%)
Experiment
Median
25 50 75 100
Prosumer ratio (%)
0
20
40
60
80
100
Complementarity factor (%)
Experiment
Median
bc
a
Figure 9. Histogram of complementarity factor at 2–3 p.m. for prosumers’ ratio of 0.25.
Figure 10 compares the average CF values and their 95% confidence interval for all PRs
during the hours of PV generation. Due to the high variation across different PRs and hours
of the day, the y-axis is shown on a logarithmic scale. For all the scenarios, it was observed
that from 10 a.m. to 4 p.m., the aggregate net energy of prosumers was negative (i.e.,
indicating surplus energy), with a high potential for energy exchange. For 25% and 50%
PRs, the highest CF values were 27% and 75% from 12 to 1 p.m. The 75% and 100% PRs are
the only conditions that provide capacities for community self-consumption with CF values
of more than 100% from around 10 a.m. to 3 p.m., with the highest values of 165% and 420%
from 12 to 1 p.m., respectively. Furthermore, from 11 a.m. to 3 p.m., with high PRs, energy
storage could provide opportunities for increased complementarity and self-sufficiency
after 3 p.m.—on average the remaining surplus was 30% (with 75% PR) and 217% (with
100% PR) of the community demand from 11 a.m. to 3 p.m. However, for PRs lower than
75%, the complementarity of the loads could ideally enable renewables’ self-consumption
Energies 2021,14, 4318 14 of 21
without the need for storage or supply to the grid. After 4 p.m., the aggregate prosumers’
net energy is positive, which considerably limits the complementarity. Therefore, even
with the 100% PR, without energy storage only 12% and 2% of the community deficit
could be supplied by PV generation. Figure 10 summarizes the community capacities for
temporal self-sufficiency considering different PV penetration ratios. It must be noted
that simulations were carried out by integrating real load profiles to report these values.
Nonetheless, either the integration of less than 75% PR or expanding the time span of
11 a.m.–3 p.m. to get the same demand/supply matching could be achieved by leveraging
demand elasticity for load profile change enabled by different economic or environmental
incentives, such as higher revenue or the desire for clean energy use (more discussion have
been presented in Section 4.5).
Energies 2021, 14, x FOR PEER REVIEW 14 of 21
values. Nonetheless, either the integration of less than 75% PR or expanding the time span
of 11 am–3 pm to get the same demand/supply matching could be achieved by leveraging
demand elasticity for load profile change enabled by different economic or environmental
incentives, such as higher revenue or the desire for clean energy use (more discussion
have been presented in Section 4.5).
Figure 10. Impact of varying prosumer ratios on complementarity factor of the community during
PV generation hours for 25%, 50%, 75%, and 100% prosumers’ ratio.
In what follows, we have investigated how energy storage and load flexibility due to
user adaptation affect these capacities.
4.3. Energy Storage Integration
To evaluate the impact of energy storage on community load balancing, we consid-
ered different levels of battery storage adoption among prosumers in the sample commu-
nities of the baseline analyses. For each prosumers’ ratio, different battery integration ra-
tios of {0.25, 0.5, 0.75, and 1} were used in simulating communities with N = 20 for 100
repetitions. For example, for the PR of 0.5 and the battery integration ratio of 0.5, ten
prosumers and ten consumers form the community, out of which five prosumers have
battery storage. The ultimate goal of quantifying complementarity is to evaluate the po-
tential for improved self-sufficiency at the community level. Therefore, Figure 11 presents
the community self-sufficiency for the solar generation timeframe (9 am to 7 pm) under
different prosumers and battery integration ratios. The battery integration ratio of 0 in
Figure 11 represents the cases presented in Figure 10. Maximum improvements of 4.8%,
11.3%, 13.0%, and 17.0% in self-sufficiency are observed with full battery integration for
25%, 50%, 75%, and 100% of PRs, respectively. With 100% PV–battery adoption, the ideal
community self-sufficiency reaches almost 83% by enabling the surplus energy stored
during the 10 am to 3 pm timeframe to be used for self-consumption and energy exchange
later in the day.
9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19
Hour of the day
1
10
100
600
25% prosumer ratio
50% prosumer ratio
75% prosumer ratio
100% prosumer ratio
Negative net energy for prosumers
Figure 10.
Impact of varying prosumer ratios on complementarity factor of the community during
PV generation hours for 25%, 50%, 75%, and 100% prosumers’ ratio.
In what follows, we have investigated how energy storage and load flexibility due to
user adaptation affect these capacities.
4.3. Energy Storage Integration
To evaluate the impact of energy storage on community load balancing, we considered
different levels of battery storage adoption among prosumers in the sample communities
of the baseline analyses. For each prosumers’ ratio, different battery integration ratios of
{0.25, 0.5, 0.75, and 1} were used in simulating communities with N = 20 for 100 repetitions.
For example, for the PR of 0.5 and the battery integration ratio of 0.5, ten prosumers and
ten consumers form the community, out of which five prosumers have battery storage. The
ultimate goal of quantifying complementarity is to evaluate the potential for improved
self-sufficiency at the community level. Therefore, Figure 11 presents the community self-
sufficiency for the solar generation timeframe (9 a.m. to 7 p.m.) under different prosumers
and battery integration ratios. The battery integration ratio of 0 in Figure 11 represents the
cases presented in Figure 10. Maximum improvements of 4.8%, 11.3%, 13.0%, and 17.0% in
self-sufficiency are observed with full battery integration for 25%, 50%, 75%, and 100% of
PRs, respectively. With 100% PV–battery adoption, the ideal community self-sufficiency
reaches almost 83% by enabling the surplus energy stored during the 10 a.m. to 3 p.m.
timeframe to be used for self-consumption and energy exchange later in the day.
Energies 2021,14, 4318 15 of 21
Energies 2021, 14, x FOR PEER REVIEW 15 of 21
Figure 11. The improved community self-sufficiency with different ratios of PV and battery integra-
tion.
4.4. User Adaptation and Load Profile Change as Complementarity Capacity
User adaptation through load flexibility and profile change has the potential to im-
prove self-sufficiency, specifically for the timeframe when solar energy generation is de-
clining (as shown in Figure 10). It is known that not all users are willing to adopt adaptive
behavior for load flexibility because of different circumstances of convenience and com-
fort. However, a portion of users will participate in smart load scheduling for load profile
change due to economic or environmental incentives. Therefore, we considered different
levels of participation as {20%, 40%, 60%, 80%, 100%} for different combinations of adap-
tation modalities and load profile change {AC 1°, AC 1° + deferrable loads, AC 1°(with
pre-cooling), AC 1°(with pre-cooling) + deferrable loads, AC 2°(pre-cooling), AC 2°(pre-
cooling) + deferrable loads}. AC a° refers to an adaptation mode, in which the thermostat
setpoint is increased (considering our summer season analysis) by a° to shed the load. Pre-
cooling enables the energy management system to benefit from the thermal storage ca-
pacity of the building to use energy available earlier in the day. On the other hand, defer-
rable loads (EV and wet appliances) pertain to rescheduling of the appliances that have
high demands and could potentially be operated at a different time. Given the low poten-
tial of energy exchange during the hours leading to evening as shown in Figure 10, the
impact of load flexibility due to user adaptation was considered during this relatively
short timeframe of needi.e., 57 pm.
Figure 12 presents community self-sufficiency for different PV integration levels and
different levels of participation and load profile change from prosumers/consumers in
taking adaptative actions. The flexibility participation ratio of 0 in Figure 12 represents
the cases presented in Figure 10. Each line represents one adaptation modality, and each
data point shows the mean value from 100 experiments in communities of 20 households.
The user adaptation could provide additional complementarity capacity on both prosum-
ers and consumers’ sides. For 25% PR, a maximum increase of 4.4% (from 22.1% to 26.5%),
for 50% PR, an increase of 5.4% (from 39.3% to 44.7%), for 75% PR, an increase of 9.0%
(from 53.2% to 62.2%), and for 100% PR, an increase of 10.4% (65.1% to 75.5%) were ob-
served. Regarding the level of participation, regardless of PRs, on average improvements
of 2.2%, 4.1%, 5.0%, and 7.5% with 25%, 50%, 75%, and 100% participation over the base-
line were observed. User adaptation has the potential to boost self-sufficiency up to about
10% for a PR of 75% and higher. Considering a more conservative adaptation modality of
AC 1°(with pre-cooling) + deferrable loads and a 50% participation, the self-sufficiency
0 25 50 75 100
Battery integration (%)
20
30
40
50
60
70
80
90 25% PV integration
50% PV integration
75% PV integration
100% PV integration
Figure 11. The improved community self-sufficiency with different ratios of PV and battery integration.
4.4. User Adaptation and Load Profile Change as Complementarity Capacity
User adaptation through load flexibility and profile change has the potential to im-
prove self-sufficiency, specifically for the timeframe when solar energy generation is de-
clining (as shown in Figure 10). It is known that not all users are willing to adopt adaptive
behavior for load flexibility because of different circumstances of convenience and comfort.
However, a portion of users will participate in smart load scheduling for load profile
change due to economic or environmental incentives. Therefore, we considered different
levels of participation as {20%, 40%, 60%, 80%, 100%} for different combinations of adap-
tation modalities and load profile change {AC 1
, AC 1
+ deferrable loads, AC 1
(with
pre-cooling), AC 1
(with pre-cooling) + deferrable loads, AC 2
(pre-cooling), AC 2
(pre-
cooling) + deferrable loads}. AC a
refers to an adaptation mode, in which the thermostat
setpoint is increased (considering our summer season analysis) by a
to shed the load.
Pre-cooling enables the energy management system to benefit from the thermal storage
capacity of the building to use energy available earlier in the day. On the other hand,
deferrable loads (EV and wet appliances) pertain to rescheduling of the appliances that
have high demands and could potentially be operated at a different time. Given the low
potential of energy exchange during the hours leading to evening as shown in Figure 10,
the impact of load flexibility due to user adaptation was considered during this relatively
short timeframe of need—i.e., 5–7 p.m.
Figure 12 presents community self-sufficiency for different PV integration levels and
different levels of participation and load profile change from prosumers/consumers in
taking adaptative actions. The flexibility participation ratio of 0 in Figure 12 represents the
cases presented in Figure 10. Each line represents one adaptation modality, and each data
point shows the mean value from 100 experiments in communities of 20 households. The
user adaptation could provide additional complementarity capacity on both prosumers
and consumers’ sides. For 25% PR, a maximum increase of 4.4% (from 22.1% to 26.5%), for
50% PR, an increase of 5.4% (from 39.3% to 44.7%), for 75% PR, an increase of 9.0% (from
53.2% to 62.2%), and for 100% PR, an increase of 10.4% (65.1% to 75.5%) were observed.
Regarding the level of participation, regardless of PRs, on average improvements of 2.2%,
4.1%, 5.0%, and 7.5% with 25%, 50%, 75%, and 100% participation over the baseline were
observed. User adaptation has the potential to boost self-sufficiency up to about 10% for a
PR of 75% and higher. Considering a more conservative adaptation modality of AC 1
(with
Energies 2021,14, 4318 16 of 21
pre-cooling) + deferrable loads and a 50% participation, the self-sufficiency improvement
is about 4 to 5%. This potential could be considerably increased by extending the user
adaptation capacities throughout the day and expanding on smart home automation.
Energies 2021, 14, x FOR PEER REVIEW 16 of 21
improvement is about 4 to 5%. This potential could be considerably increased by extend-
ing the user adaptation capacities throughout the day and expanding on smart home au-
tomation.
Figure 12. Impact of flexible (adaptive) behavior on self-sufficiency through load profile change for
(a) 25% PV integration, (b) 50% PV integration, (c) 75% PV integration, and (d) 100% PV integration.
User adaptation versus battery storage: To compare the impact of user adaptation
versus energy storage, Figure 13 shows the improvement of self-sufficiency by leveraging
load flexibility and battery storage compared to the baseline as presented in Figure 10. It
should be noted that the self-sufficiency improvement from energy storage stems from
surplus throughout the day while user adaptation pertains to adaptation during a limited
timeframe. With a low PR of 25%, the impact of user flexibility is comparable with the
battery storage, while for high PV integration ratios battery storage shows a higher im-
provement compared to user flexibility. Nonetheless, with high PV ratios of 75% and
100% in the community, user flexibility could create capacities for self-sufficiency im-
provements that are 69% and 61% of that of battery integration, respectively. Therefore,
in the case of low or no battery integration in a community, user adaptation and its result-
ant flexibility can be deemed an alternative approach for increasing complementarity and
thus self-sufficiency at the community level. The potentials from user adaptation could be
realized by using monitoring systems for user–building interactions [39,40], as well as au-
tomated demand-side Home Energy Management (HEM) systems that use the trade-off
between community self-sufficiency and user convenience. Advanced technologies, such
as smart appliances and smart thermostats, which allow for automation while accounting
for user preferences, would be an integral part of these smart home eco-systems.
(a)
AC 1
°
(pre-cooling)
AC 1
°
(pre-cooling) and deferrables
AC 2
°
(pre-cooling) and deferr ables
AC 2
°
(pre-cooling)
AC 1
°
and deferrables
AC 1
°
(b)
(d)(c)
Figure 12.
Impact of flexible (adaptive) behavior on self-sufficiency through load profile change for
(
a
) 25% PV integration, (
b
) 50% PV integration, (
c
) 75% PV integration, and (
d
) 100% PV integration.
User adaptation versus battery storage:
To compare the impact of user adaptation
versus energy storage, Figure 13 shows the improvement of self-sufficiency by leveraging
load flexibility and battery storage compared to the baseline as presented in
Figure 10
. It
should be noted that the self-sufficiency improvement from energy storage stems from
surplus throughout the day while user adaptation pertains to adaptation during a lim-
ited timeframe. With a low PR of 25%, the impact of user flexibility is comparable with
the battery storage, while for high PV integration ratios battery storage shows a higher
improvement compared to user flexibility. Nonetheless, with high PV ratios of 75% and
100% in the community, user flexibility could create capacities for self-sufficiency improve-
ments that are 69% and 61% of that of battery integration, respectively. Therefore, in the
case of low or no battery integration in a community, user adaptation and its resultant
flexibility can be deemed an alternative approach for increasing complementarity and
thus self-sufficiency at the community level. The potentials from user adaptation could
be realized by using monitoring systems for user–building interactions [
39
,
40
], as well as
automated demand-side Home Energy Management (HEM) systems that use the trade-off
between community self-sufficiency and user convenience. Advanced technologies, such
as smart appliances and smart thermostats, which allow for automation while accounting
for user preferences, would be an integral part of these smart home eco-systems.
Energies 2021,14, 4318 17 of 21
Energies 2021, 14, x FOR PEER REVIEW 17 of 21
Figure 13. Comparison of self-sufficiency improvement with different user adaptation modalities and battery integration.
4.5. Market Design and User Behavior Dimensions’ Impact on Self-Sufficiency
Market design impact: In recent years, the economic dimension of market design has
been the subject of research from theoretical [41,42] and analytical perspectives [43,44]
with recent expansions to real-world implementations for energy exchange [45,46]. In a
P2P energy trading market design, considerations are given to individual user preferences
and different trading strategies [47,48]. Specifically, the P2P market relies on the auction
mechanism [49,50], in which participants could choose their bid resulting in different op-
tions for price and quantity of energy to purchase and/or sell. Some approaches in market
design include uniform double auction [46,51], combinatorial auction [43], and time dis-
crete discriminatory double auction [45] that enable users to actively participate in pricing
and energy allocation mechanisms. Previous field studies and real-world implementa-
tions have shown the increased added value of energy trading from an economic perspec-
tive [45,46] and the amount of energy traded among peers compared to using utility-set
fixed rates [45].
In a field study in Switzerland, in a setting comparable to our case-study communi-
ties, a community of prosumers (with PV and PV–battery owners) and consumers partic-
ipated in a P2P market [45]. The double auction mechanism was adopted as the market
design for prosumers and consumers to have the option of purchasing or selling green
energy under their preferred conditions. In this market, after the bids were collected, the
market was cleared such that the highest purchase order was matched the lowest sell
price, and this procedure was continued until the entire bid list was checked, resulting in
the price for each exchange being the mean of the purchase/sell bid [45]. The remaining
demand was supplied by the central utility provider with their standard tariff. Given this
arbitrarily defined pricing market, the prosumers with excessive solar (already supplying
their own demand or feeding their battery unit, if any) could set a price lower than pur-
chasing from the utility provider and higher than the feed-in tariff, which would benefit
both prosumers and consumers. For example, in the field study of Worner et al. [45], con-
sumers were willing to pay lower rates to their peers for solar energy (0.19 CHF/kW h)
compared to the utility price (0.21 CHF/kW h), whereas prosumers were willing to set
even lower prices on average (0.13 CHF/kW h), resulting the trading process to be eco-
nomically beneficial for both parties. On average, the increased revenue for prosumers
was 32% while consumers had a 7% cost saving on their bills [45]. Moreover, the authors
showed that the variations in price were compatible with the solar energy generation level
assumed in our simulations. In other words, the price increased in hours that the solar
energy generation was low, and ramped down as we approached high-generation hours.
Thus, our simulations could represent the economic dimension of the market dynamics.
Therefore, the P2P market could potentially provide incentives through added economic
value for both sides in addition to utilizing a clean energy resource.
User-driven demand–supply balancing for DER integration: As presented, we have
considered different ratios of infrastructure configurations and showed how they affect
25% PV 50% PV 75% PV 100% PV
0
5
10
15
20
Improve in
self-sufficiency (%)
AC 1°
AC 1° and deferrables
AC 1°(pre-cooling)
AC 1°(pre-cooling) and deferrables
AC 2°(pre-cooling)
AC 2° (pre-cooling) and deferrables
Battery
Figure 13. Comparison of self-sufficiency improvement with different user adaptation modalities and battery integration.
4.5. Market Design and User Behavior Dimensions’ Impact on Self-Sufficiency
Market design impact
: In recent years, the economic dimension of market design
has been the subject of research from theoretical [
41
,
42
] and analytical perspectives [
43
,
44
]
with recent expansions to real-world implementations for energy exchange [
45
,
46
]. In a
P2P energy trading market design, considerations are given to individual user preferences
and different trading strategies [
47
,
48
]. Specifically, the P2P market relies on the auction
mechanism [
49
,
50
], in which participants could choose their bid resulting in different
options for price and quantity of energy to purchase and/or sell. Some approaches in
market design include uniform double auction [
46
,
51
], combinatorial auction [
43
], and
time discrete discriminatory double auction [
45
] that enable users to actively participate in
pricing and energy allocation mechanisms. Previous field studies and real-world imple-
mentations have shown the increased added value of energy trading from an economic
perspective [
45
,
46
] and the amount of energy traded among peers compared to using
utility-set fixed rates [45].
In a field study in Switzerland, in a setting comparable to our case-study communities,
a community of prosumers (with PV and PV–battery owners) and consumers participated
in a P2P market [
45
]. The double auction mechanism was adopted as the market design for
prosumers and consumers to have the option of purchasing or selling green energy under
their preferred conditions. In this market, after the bids were collected, the market was
cleared such that the highest purchase order was matched the lowest sell price, and this
procedure was continued until the entire bid list was checked, resulting in the price for
each exchange being the mean of the purchase/sell bid [
45
]. The remaining demand was
supplied by the central utility provider with their standard tariff. Given this arbitrarily
defined pricing market, the prosumers with excessive solar (already supplying their own
demand or feeding their battery unit, if any) could set a price lower than purchasing from
the utility provider and higher than the feed-in tariff, which would benefit both prosumers
and consumers. For example, in the field study of Worner et al. [
45
], consumers were
willing to pay lower rates to their peers for solar energy (0.19 CHF/kW h) compared to the
utility price (0.21 CHF/kW h), whereas prosumers were willing to set even lower prices on
average (0.13 CHF/kW h), resulting the trading process to be economically beneficial for
both parties. On average, the increased revenue for prosumers was 32% while consumers
had a 7% cost saving on their bills [
45
]. Moreover, the authors showed that the variations in
price were compatible with the solar energy generation level assumed in our simulations.
In other words, the price increased in hours that the solar energy generation was low,
and ramped down as we approached high-generation hours. Thus, our simulations could
represent the economic dimension of the market dynamics. Therefore, the P2P market
could potentially provide incentives through added economic value for both sides in
addition to utilizing a clean energy resource.
User-driven demand–supply balancing for DER integration
: As presented, we have
considered different ratios of infrastructure configurations and showed how they affect
Energies 2021,14, 4318 18 of 21
the self-sufficiency of the communities, as well as strategies that could be adopted to
improve self-sufficiency when the solar energy generation declines moving towards the
late afternoon hours. We considered the impact of human adaptive behaviors for demand
elasticity to improve self-sufficiency in those hours (Section 4.4). However, the impact of
adaptive behavior could be expanded into earlier hours to improve self-sufficiency for
communities with less PV penetration. Economic incentives have been traditionally used
for shifting energy use behavior so that the cost (both economic and environmental) of
energy generation could be minimized. Adaptive behaviors could include load shifting to
a different time of the day and load shedding for appliances/devices that could be used
more efficiently. For example, AC systems could be operated using human-centered control
strategies to reduce energy use. Therefore, users could either reduce their energy use or
shift the operation to a different time of the day. Considering the energy source during
the high solar generation hours is low-cost green energy, it is preferred that the energy
consumption is shifted to those hours. However, there are strategies to reduce energy use
even for those hours such as the efficient use of AC systems specifically for house types that
could benefit from more refined thermal energy management compared to the conventional
single thermostat controller system. As previous research has shown [
14
], the behavioral
incentives for energy trading are categorized into three classes of ‘energy matching,’ which
is the willingness to use green local energy; ‘preference satisfaction,’ which is setting
preferences for what clean resources to use and what price to set for purchasing or selling
energy; and ‘uncertainty reduction,’ which is associated with obtaining economic profit
through energy exchange and increasing revenue. Accordingly, the DER integration level
(for PV or battery adoption) could be reduced or self-sufficiency values increased by
selecting one of these methods, depending on the user’s preferences, values/beliefs, and
lifestyle [52].
4.6. Limitations
There are a number of limitations associated with this study: (1) Like any data-driven
study, the findings presented here were extracted from a dataset. Therefore, the general-
izability of findings is associated with similarity in the energy profiles of prosumers and
consumers. Nonetheless, we used the data from the Pecan Street Project [
38
], which is
currently one of the largest campaigns for energy studies, and selected ~250 households
from the ERCOT grid. Furthermore, the majority of the analyses presented here include
sub-hourly resolution demand and generation data. Demand data was extracted from
smart meters, which are a ubiquitous and available metering infrastructure in most regions.
Generation data, in the absence of real PV data, can also be estimated from weather and ge-
ographical location information with open-source solutions (e.g., PV Watts [
53
]). Therefore,
similar types of analyses can be carried out on other datasets. (2) The findings for demand–
supply balancing potential for the community were obtained through the aggregation of
energy surplus and deficit of individual consumers, assuming energy trading without
including transmission loss, network constraints, and other limiting factors in the economic
operation of the market, such as exercise of market power in price-setting strategies [
54
].
Studying these factors could shed further light on the findings of this work. (3) For the
scenarios with battery integration, since we were interested in quantifying the commu-
nity capacity for energy exchange, we considered flat pricing rates. Therefore, batteries
were discharged at the earliest time that demand exceeded generation. In the presence of
dynamic pricing that reflects the dynamics of demand, optimization techniques could be
explored for maximizing the benefit from energy exchange. (4) We relied on the presence
of PV systems that were installed in prosumers’ households, with the same capacity as
specified in the dataset. Therefore, the impact of different PV capacities or battery sizing
was not taken into account. (5) Different climate conditions could be considered to expand
the analyses for air conditioning load flexibility. (6) We presented the results of the case
study for one summer season. Therefore, integrating the results over a yearly period could
show the impact of seasonality.
Energies 2021,14, 4318 19 of 21
5. Conclusions
A comprehensive analysis of surplus–deficit complementarity capacity in commu-
nities with different ratios of prosumers and consumers was conducted to quantify the
load-balancing capacity of communities through self-consumption and energy exchange.
In doing so, we quantified the temporal variation of surplus–deficit complementarity
capacity by considering varied infrastructure configurations—i.e., the ratio of prosumers to
consumers. Improved complementarity led to increased community-level self-sufficiency,
which was quantified by considering the impact of varying levels of energy storage inte-
grations compared to user adaptation for load flexibility. In communities with at least 75%
prosumers, without any energy storage capacity, and no economical behavior adaptation
for demand elasticity, community surplus–deficit balancing could be achieved through en-
ergy exchange during times of high solar generation (from 11 a.m. to 2 p.m.) and additional
surplus energy was available for storage. With an equal distribution of prosumers and
consumers, ~75% of deficit energy in the community could be canceled out by prosumers
at the peak PV generation hours. Therefore, energy exchange could ideally be used to
avoid the need for considerable energy storage infrastructure. With 100% PV integration,
the community self-sufficiency could be extended to 3 p.m. However, for the timeframe
leading to the evening (from 4 p.m. to 7 p.m.), when PV generation diminishes regardless
of the PV integration ratio, the community could no longer be self-sufficient through energy
exchange. As a result, the impact of alternative measures such as battery storage integration
or creating flexibility potential from user adaptation was studied. It was observed that
with 100% battery integration or 100% participation of users in adjusting flexible loads
during critical timeframes, the self-sufficiency of the community could improve by 17%
and 10.4% up to 83% or 76%, respectively. User flexibility was considered to be thermostat
adjustment with pre-cooling, as well as rescheduling the operations of deferrable loads
for the time that solar generation is declining—i.e., 5–7 p.m. A more moderate case of
user adaptation which included a 1
increase in setpoint plus load rescheduling and for
50% user participation resulted in ~5% self-sufficiency improvement. Furthermore, in the
best cases, it was observed that in the absence of energy storage systems, user adaptation
for load flexibility showed an improvement in self-sufficiency by ~60% of what could be
offered by commercial conventional storage systems.
Future directions of this research include: (1) the study of automated smart home
technologies that optimize user adaptation under the constraints of comfort and con-
venience; (2) assessing the impact of seasonality and community diversity; (3) inves-
tigating the impact of different DER configurations, network constraints, and control
algorithms for the realization of energy exchange using human adaptation; and (4) inves-
tigating different economic and social factors that drive and enhance community level
demand/supply matching.
Author Contributions:
Conceptualization, M.A. and F.J.; methodology, M.A.; investigation, M.A.;
visualization, M.A.; writing—original draft preparation, M.A.; writing—review and editing, F.J. and
M.A.; supervision, F.J.; funding acquisition, F.J. All authors have read and agreed to the published
version of the manuscript.
Funding:
This work was supported in part by Virginia Tech’s Open Access Subvention
Fund (VTOASF).
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Chopra, A.; Kundra, V. A Policy Framework for the 21st Century Grid: Enabling Our Secure Energy Future; CiteSeerX: University Park,
PA, USA, 2011.
2.
Tushar, W.; Saha, T.K.; Yuen, C.; Smith, D.; Poor, H.V. Peer-to-peer trading in electricity networks: An overview. IEEE Trans. Smart
Grid 2020,11, 3185–3200. [CrossRef]
3.
German Renewable Energy Sources Act. Available online: https://en.wikipedia.org/wiki/German_Renewable_Energy_Sources_
Act (accessed on 13 July 2021).
Energies 2021,14, 4318 20 of 21
4.
Nowak, S. Trends 2013 in Photovoltaic Applications: Survey Report of Selected IEA Countries between 1992 and 2012; Technical Report
IEA-PVPS T1-23; Int. Energy Agency (IEA): Paris, France, 2013; Volume 2013.
5.
News, I.C. As Rooftop Solar Grows, What Should the Future of Net Metering Look Like? Available online: https:
//insideclimatenews.org/news/11062019/rooftop-solar-net-metering-rates-renewable-energy-homeowners- utility-state- law-
changes-map (accessed on 13 July 2021).
6.
Salom, J.; Widén, J.; Candanedo, J.; Lindberg, K.B. Analysis of grid interaction indicators in net zero-energy buildings with
sub-hourly collected data. Adv. Build. Energy Res. 2015,9, 89–106. [CrossRef]
7.
Wang, Y.; Chen, Q.; Hong, T.; Kang, C. Review of smart meter data analytics: Applications, methodologies, and challenges. IEEE
Trans. Smart Grid 2018,10, 3125–3148. [CrossRef]
8.
Afzalan, M.; Jazizadeh, F. Residential loads flexibility potential for demand response using energy consumption patterns and
user segments. Appl. Energy 2019,254, 113693. [CrossRef]
9.
Afzalan, M.; Jazizadeh, F. A Machine Learning Framework to Infer Time-of-Use of Flexible Loads: Resident Behavior Learning
for Demand Response. IEEE Access 2020,8, 111718–111730. [CrossRef]
10.
Afzalan, M.; Jazizadeh, F. Efficient integration of smart appliances for demand response programs. In Proceedings of the 5th
Conference on Systems for Built Environments, Shenzhen, China, 7 November 2018; pp. 29–32.
11.
Zhang, Z.; Li, R.; Li, F. A novel peer-to-peer local electricity market for joint trading of energy and uncertainty. IEEE Trans. Smart
Grid 2019,11, 1205–1215. [CrossRef]
12.
Liu, T.; Tan, X.; Sun, B.; Wu, Y.; Guan, X.; Tsang, D.H. Energy management of cooperative microgrids with p2p energy
sharing in distribution networks. In Proceedings of the 2015 IEEE International Conference on Smart Grid Communications
(SmartGridComm), Miami, FL, USA, 2–5 November 2015; pp. 410–415.
13.
Cui, S.; Wang, Y.-W.; Xiao, J.-W. Peer-to-peer energy sharing among smart energy buildings by distributed transaction. IEEE
Trans. Smart Grid 2019,10, 6491–6501. [CrossRef]
14.
Morstyn, T.; Farrell, N.; Darby, S.J.; McCulloch, M.D. Using peer-to-peer energy-trading platforms to incentivize prosumers to
form federated power plants. Nat. Energy 2018,3, 94–101. [CrossRef]
15.
Andoni, M.; Robu, V.; Flynn, D.; Abram, S.; Geach, D.; Jenkins, D.; Peacock, A. Blockchain technology in the energy sector: A
systematic review of challenges and opportunities. Renew. Sustain. Energy Rev. 2019,100, 143–174. [CrossRef]
16.
Sousa, T.; Soares, T.; Pinson, P.; Moret, F.; Baroche, T.; Sorin, E. Peer-to-peer and community-based markets: A comprehensive
review. Renew. Sustain. Energy Rev. 2019,104, 367–378. [CrossRef]
17.
Zhang, C.; Wu, J.; Long, C.; Cheng, M. Review of existing peer-to-peer energy trading projects. Energy Proc.
2017
,105, 2563–2568.
[CrossRef]
18.
Zhou, Y.; Wu, J.; Long, C. Evaluation of peer-to-peer energy sharing mechanisms based on a multiagent simulation framework.
Appl. Energy 2018,222, 993–1022. [CrossRef]
19.
Alam, M.R.; St-Hilaire, M.; Kunz, T. Peer-to-peer energy trading among smart homes. Appl. Energy
2019
,238, 1434–1443.
[CrossRef]
20.
An, J.; Lee, M.; Yeom, S.; Hong, T. Determining the Peer-to-Peer electricity trading price and strategy for energy prosumers and
consumers within a microgrid. Appl. Energy 2020,261, 114335. [CrossRef]
21.
Hahnel, U.J.; Herberz, M.; Pena-Bello, A.; Parra, D.; Brosch, T. Becoming prosumer: Revealing trading preferences and decision-
making strategies in peer-to-peer energy communities. Energy Policy 2020,137, 111098. [CrossRef]
22.
Morstyn, T.; McCulloch, M.D. Multiclass energy management for peer-to-peer energy trading driven by prosumer preferences.
IEEE Trans. Power Syst. 2018,34, 4005–4014. [CrossRef]
23.
Zepter, J.M.; Lüth, A.; del Granado, P.C.; Egging, R. Prosumer integration in wholesale electricity markets: Synergies of
peer-to-peer trade and residential storage. Energy Build. 2019,184, 163–176. [CrossRef]
24.
Nguyen, S.; Peng, W.; Sokolowski, P.; Alahakoon, D.; Yu, X. Optimizing rooftop photovoltaic distributed generation with battery
storage for peer-to-peer energy trading. Appl. Energy 2018,228, 2567–2580. [CrossRef]
25.
Long, C.; Wu, J.; Zhang, C.; Cheng, M.; Al-Wakeel, A. Feasibility of peer-to-peer energy trading in low voltage electrical
distribution networks. Energy Proc. 2017,105, 2227–2232. [CrossRef]
26.
Lopes, R.A.; Martins, J.; Aelenei, D.; Lima, C.P. A cooperative net zero energy community to improve load matching. Renew.
Energy 2016,93, 1–13. [CrossRef]
27.
Luthander, R.; Widén, J.; Nilsson, D.; Palm, J. Photovoltaic self-consumption in buildings: A review. Appl. Energy
2015
,142, 80–94.
[CrossRef]
28.
Kwac, J.; Flora, J.; Rajagopal, R. Household energy consumption segmentation using hourly data. IEEE Trans. Smart Grid
2014
,
5, 420–430. [CrossRef]
29.
Kong, W.; Dong, Z.Y.; Jia, Y.; Hill, D.J.; Xu, Y.; Zhang, Y. Short-term residential load forecasting based on LSTM recurrent neural
network. IEEE Trans. Smart Grid 2017,10, 841–851. [CrossRef]
30.
Linssen, J.; Stenzel, P.; Fleer, J. Techno-economic analysis of photovoltaic battery systems and the influence of different consumer
load profiles. Appl. Energy 2017,185, 2019–2025. [CrossRef]
31.
Quoilin, S.; Kavvadias, K.; Mercier, A.; Pappone, I.; Zucker, A. Quantifying self-consumption linked to solar home battery
systems: Statistical analysis and economic assessment. Appl. Energy 2016,182, 58–67. [CrossRef]
Energies 2021,14, 4318 21 of 21
32.
Pipattanasomporn, M.; Kuzlu, M.; Rahman, S.; Teklu, Y. Load profiles of selected major household appliances and their demand
response opportunities. IEEE Trans. Smart Grid 2013,5, 742–750. [CrossRef]
33.
Knezovi´c, K.; Marinelli, M.; Codani, P.; Perez, Y. Distribution grid services and flexibility provision by electric vehicles: A review
of options. In Proceedings of the 2015 50th International Universities Power Engineering Conference (UPEC), Stoke on Trent, UK,
1–4 September 2015; pp. 1–6.
34.
Hu, M.; Xiao, F.; Wang, L. Investigation of demand response potentials of residential air conditioners in smart grids using
grey-box room thermal model. Appl. Energy 2017,207, 324–335. [CrossRef]
35.
Hu, M.; Xiao, F. Investigation of the demand response potentials of residential air conditioners using grey-box room thermal
model. Energy Proc. 2017,105, 2759–2765. [CrossRef]
36.
Tesla Powerwall. Available online: https://www.tesla.com/support/energy/powerwall/learn/how-powerwall- works (accessed
on 13 July 2021).
37. Barbour, E.; González, M.C. Projecting battery adoption in the prosumer era. Appl. Energy 2018,215, 356–370. [CrossRef]
38. Source: Pecan Street Inc., Dataport. Available online: https://www.pecanstreet.org/dataport/ (accessed on 13 July 2021).
39.
Afzalan, M.; Jazizadeh, F.; Wang, J. Self-configuring event detection in electricity monitoring for human-building interaction.
Energy Build. 2019,187, 95–109. [CrossRef]
40.
Jazizadeh, F.; Afzalan, M.; Becerik-Gerber, B.; Soibelman, L. EMBED: A dataset for energy monitoring through building electricity
disaggregation. In Proceedings of the Ninth International Conference on Future Energy Systems, Karlsruhe, Germany, 12–15 June
2018; pp. 230–235.
41. Klemperer, P. What really matters in auction design. J. Econ. Perspect. 2002,16, 169–189. [CrossRef]
42. Rosen, C.; Reinhard, M. An auction design for local reserve energy markets. Decis. Support Syst. 2013,56, 168–179. [CrossRef]
43.
Block, C.; Dirk, N.; Christof, W. A market mechanism for energy allocation in micro-chp grids. In Proceedings of the 41st Annual
Hawaii International Conference on System Sciences (HICSS 2008), Waikoloa, Big Island, HI, USA, 7–10 January 2008; p. 172.
44.
Griego, D.; Schopfer, S.; Henze, G.; Fleisch, E.; Tiefenbeck, V. Aggregation effects for microgrid communities at varying sizes and
prosumer-consumer ratios. Energy Proc. 2019,159, 346–351. [CrossRef]
45.
Wörner, A.M.; Ableitner, L.; Arne, M.; Felix, W.; Tiefenbeck, V. Peer-to-peer energy trading in the real world: Market design and
evaluation of the user value proposition. In Proceedings of the 4th International Conference on Information Systems (ICIS 2019),
Munich, Germany, 15–18 December 2019.
46.
Mengelkamp, E.; Johannes, G.; Kerstin, R.; Kessler, S.; Orsini, L.; Weinhardt, C. Designing microgrid energy markets: A case
study: The Brooklyn Microgrid. Appl. Energy 2018,210, 870–880. [CrossRef]
47.
Bichler, M.; Alok, G.; Wolfgang, K. Research commentary—Designing smart markets. Inf. Syst. Res.
2010
,21, 688–699. [CrossRef]
48.
Wolfgang, K.; Collins, J.; Reddy, P. Power TAC: A competitive economic simulation of the smart grid. Energy Econ.
2013
,
39, 262–270.
49.
Koolen, D.; Liangfei, Q.; Wolfgang, K.; Alok, G. The sustainability tipping point in electricity markets. In Proceedings of the 38th
International Conference on Information Systems: Transforming Society with Digital Innovation, ICIS 2017, Seoul, Korea, 10–13
December 2018.
50.
Mengelkamp, E.; Johannes, G.; Weinhardt, C. Decentralizing energy systems through local energy markets: The LAMP-Project.
In Multikonferenz Wirtschaftsinformatik; Verlag: Berlin, Germany, 2018; pp. 924–930.
51. Lacity, M.C. Addressing key challenges to making enterprise blockchain applications a reality. MIS Q. Exec. 2018,17, 201–222.
52.
Heydarian, A.; McIlvennie, C.; Arpan, L.; Yousefi, S.; Syndicus, M.; Schweiker, M.; Jazizadeh, F.; Rissetto, R.; Pisello, A.L.;
Piselli, C.; et al.
What drives our behaviors in buildings? A review on occupant interactions with building systems from the lens
of behavioral theories. Build. Environ. 2020,179, 106928. [CrossRef]
53.
National Renewable Energy Laboratory (NREL). PV Watts. Available online: https://pvwatts.nrel.gov/ (accessed on 13
July 2021).
54.
Bigerna, S.; Bollino, C.A.; Polinori, P. Renewable energy and market power in the Italian electricity market. Energy J.
2016
, 37.
[CrossRef]
... In another class of distributed energy management appli-cations, load profiling could facilitate the integration of renewable resources such as solar generation. Accordingly, the increased integration of renewables is a key element to move towards decarbonization and distributed energy management in the smart grid [43], [44]. In particular, in Figure 9, clusters 9, 16, 26, and 35, whose demand patterns in the noon timeframe coincide with the solar generation, could benefit more from photovoltaic (PV) integration. ...
... Additionally, clusters 20 and 29 have almost zero demand at noon, implying they could benefit from storage if they have PV-battery systems, and could self-consume the solar energy at later times when electricity price rate is higher. Alternatively, with the rise of new technologies such as peer-to-peer energy trading between prosumers and consumers [44], these clusters, for PV-equipped houses, could also offer their high amount of on-site generation to their consumer neighbors. Finally, clusters 14, 15, 26, and 27 have sharp peaks after midnight, presumably due to EV charging and wet appliances (e.g., dishwasher) operation. ...
Article
Full-text available
With the widespread adoption of smart meters in buildings, an unprecedented amount of high-resolution energy data is released, which provides opportunities to understand building consumption patterns. Accordingly, research efforts have employed data analytics and machine learning methods for the segmentation of customers based on their load profiles, which help utilities and energy providers to promote customized/personalized targeting for energy programs. Existing energy consumption segmentation techniques use assumptions that could result in reduced quality of clusters in representing their members. Therefore, in this paper, we investigated a two-stage clustering method for capturing more representative load shape temporal patterns and peak demands through a cluster merging approach. In the first stage, load shapes are clustered (using classical clustering algorithms such as self-organizing map) by allowing a large number of clusters to accurately capture variations in energy use patterns, and cluster centroids are extracted by accounting for limited shape misalignment within the range of DR timeframes - i.e., ~2 hours. In the second stage, clusters with similar centroids and power magnitude ranges are merged by using Complexity-invariant Dynamic Time Warping. We used three datasets consisting of ~250 households (~15000 profiles) to demonstrate the efficacy of the framework, compared to baseline methods, and to discuss the impact on energy management. The proposed investigated merging-based clustering resulted in an increase in correlation (between cluster centroids and the corresponding members) by 8.2%, 8.9%, and 2.6% for datasets 1 to 3, respectively.
... Independence of the energy from the main grid can be achieved via interdependence between developed and developing MGs when the G2G-based energy trade is formed [11]. In [12], temporal complementarity between supply and demand in energy trade has been proposed for the entire community of energy self-sufficiency. For an efficient ETS, fluent switching between consumer and seller can be a crucial factor for variant operating situations. ...
Article
Full-text available
Inefficiencies in energy trading systems of microgrids are mainly caused by uncertainty in non-stationary operating environments. The problem of uncertainty can be mitigated by analyzing patterns of primary operation parameters and their corresponding actions. In this paper, a novel energy trading system based on a double deep Q-networks (DDQN) algorithm and a double Kelly strategy is proposed for improving profits while reducing dependence on the main grid in the microgrid systems. The DDQN algorithm is proposed in order to select optimized action for improving energy transactions. Additionally, the double Kelly strategy is employed to control the microgrid’s energy trading quantity for producing long-term profits. From the simulation results, it is confirmed that the proposed strategies can achieve a significant improvement in the total profits and independence from the main grid via optimized energy transactions.
Article
Buildings with solar photovoltaic (PV) generation and a stationary battery energy storage system (BESS) may self-sustain an uninterrupted full-level electricity supply during power outages. The duration of off-grid operation is dependent on the time of the power fault and the capabilities of the home energy management system (HEMS). In this paper, building resilience is quantified by analyzing the self-sustainment duration for all possible power outages throughout an entire year. An evaluation method is proposed and exercised on a reference house in California climate zone 9 for which the detailed electricity usage is simulated using the EnergyPlus software. The influence of factors such as energy use behavioral patterns, energy storage capacity from the BESS, and an electric vehicle (EV) battery on the building resilience is evaluated. Varying combinations of energy storage and controllable loads are studied for optimally improved resilience based on user preferences. It is shown that for the target home and region with a solar PV system of 7.2 kW, a BESS with a capacity of 11 kWh, and an EV with a battery of 80 kWh permanently connected to the home, off-grid self-sustained full operation is guaranteed for at least 72 h.
Article
Full-text available
Load shapes obtained from smart meter data are commonly utilized to understand daily energy use patterns for adaptive operations in applications such as Demand Response (DR). However, they do not provide information on the underlying causes of specific energy use patterns – i.e., inference on appliances’ time-of-use (ToU) as actionable information. In this paper, we investigated a scalable machine learning framework to infer the appliances’ ToU from energy load shapes in a collection of residential buildings. A scalable and generalized inference model obviates the need for model training in each building to facilitate its adoption by relying on training data from a set of previously observed buildings with available appliance-level data. To this end, we demonstrated the feasibility of using load shape segmentation to boost ToU inference in buildings by learning from their nearest matches that share similar energy use patterns. To infer an appliance ToU for a building, classification models are trained for inference on subintervals of load shapes from matched buildings with known ToU. The framework was evaluated using real-world energy data from Pecan Street Dataport. The results for a case study on electric vehicles (EV) and dryers showed promising performance by using 15-min smart meter load shape data with 83% and 71% F-score values, respectively, and without in-situ training.
Article
Full-text available
Peer-to-peer trading is a next-generation energy management technique that economically benefits proactive consumers (prosumers) transacting their energy as goods and services. At the same time, peer-to-peer energy trading is also expected to help the grid by reducing peak demand, lowering reserve requirements, and curtailing network loss. However, large-scale deployment of peer-to-peer trading in electricity networks poses a number of challenges in modeling transactions in both the virtual and physical layers of the network. As such, this article provides a comprehensive review of the state-of-the-art in research on peer-to-peer energy trading techniques. By doing so, we provide an overview of the key features of peer-to-peer trading and its benefits of relevance to the grid and prosumers. Then, we systematically classify the existing research in terms of the challenges that the studies address in the virtual and the physical layers. We then further identify and discuss those technical approaches that have been extensively used to address the challenges in peer-to-peer transactions. Finally, the paper is concluded with potential future research directions.
Article
Occupant behavior has a significant impact on building systems’ operations and efficiency. As a result, several innovative approaches have been introduced to quantify the dynamics of occupants within indoor environments, such as interactions with different building systems and the impact of various feedback and interventions to reduce the building energy consumption. To achieve this, researchers have highlighted the importance of reducing energy consumption without impacting occupant comfort. As a result, there is an increasing body of research evaluating how different theories of behavior across a variety of disciplines can explain occupant interactions with building systems. Future progress in this area calls for an in-depth understanding of behavioral theories in explaining occupant interactions with different building systems. In this paper, we have used a structured literature review approach to investigate how different psychological, sociological, and economic theories have been applied to explain occupant interactions with heating and cooling (HVAC systems), opening windows and ventilation, lighting and shading, electronic appliances, domestic hot water, as well as energy conservation behaviors. Throughout the paper, we identify the most common theories and methodologies applied within the existing research, general findings related to how occupants interact with different building systems, as well as a number of identified gaps within the literature. Finally, we provide a discussion on directions for future research studies in this area under each building system.
Article
Increasing distributed renewable energy generation renders balancing of energy supply and demand more challenging. Peer-to-peer (P2P) exchange of decentrally generated energy is a promising means to optimize renewable energy flows. While previous research has primarily focused on the technological and economic benefits of P2P energy trading, little research has investigated customer preferences in this context so far. It is thus still unclear to what extent private actors such as homeowners are actually willing to participate in P2P energy communities and if so, under which conditions. Here, we address this research gap by analyzing homeowners' trading decisions in simulated P2P electricity trading scenarios. Findings based on a sample of 301 German homeowners show that community electricity prices and state of charge of private energy storages are key determinants of homeowners' trading behavior. We moreover identified four target groups that systematically differed in their decision-making strategies ranging from price-focused prosumers (38.9%) to classic non-trading consumers (22.6%). Our findings provide promising insights for the design of P2P communities and allow policy makers to more accurately predict risks and benefits associated with P2P projects. Finally, our findings contribute to the development of tailored policy instruments aiming to increase P2P trading in decentralized energy communities.
Article
Demand response (DR) is considered an effective approach in mitigating the ever-growing concerns for supplying the electricity peak demand. Recent attempts have shown that the contribution from the aggregate impact of flexible individual residential loads can add flexibility to the power grid as ancillary services. However, current DR schemes do not systematically distinguish the varying potential of user contribution due to the highly-varied usage behaviors. Thus, this paper proposes a data-driven approach for quantifying the potential of individual flexible load users for participation in DR. We introduced a metric to capture the predictability of usage in a future DR event using the historical consumption data for different load types. The metric helps to sort the users with flexible loads in a community according to their potential for load shifting scenarios. We then evaluated the applicability of the metric in the DR context to assess the extent of energy reduction for different segments of users. In our analysis, we included electric vehicle, wet appliances (dryer, washing machine, dishwasher), and air conditioning. The analysis of real-world data shows that the approach is effective in identifying suitable user segments with higher predictive potential for demand reduction. We also presented a cross-appliance comparison for assessing the flexibility potential of different user segments. As a case study based on Pecan Street Project, the findings suggest that potentially ~160 MWh reduction might be achieved in Austin, TX through only 20% participation of the selected flexible loads for the residential sector during a 2-h event.
Article
In the future power system, an increasing number of distributed energy resources will be integrated including intermittent generation like photovoltaic (PV) and flexible demand like electric vehicles (EVs). It has been long thought to utilize the flexible demand to absorb the PV output locally with the technical solution proposed while an effective commercial arrangement is yet to be developed due to the significant uncertainty associated with local generation. This paper proposes a peer-to-peer (P2P) local electricity market model incorporating both energy trading and uncertainty trading simultaneously. The novelty is to match the forecast power with demand having time flexibility and the uncertain power with demand having power flexibility. This market enables more PV uncertainty to be balanced locally rather than propagating to the upper layer system. In the test case, 55.3% of PV forecast error can be balanced locally in the proposed joint market. In comparison, 43.6% of PV forecast error is balanced locally when the forecast power and uncertain power are traded separately in a day-ahead market and a real-time market. The proposed P2P market can also motivate PV owners to improve forecast accuracy.
Article
The advent of more proactive consumers, the so-called “prosumers”, with production and storage capabilities, is empowering the consumers and bringing new opportunities and challenges to the operation of power systems in a market environment. Recently, a novel proposal for the design and operation of electricity markets has emerged: these so-called peer-to-peer (P2P) electricity markets conceptually allow the prosumers to directly share their electrical energy and investment. Such P2P markets rely on a consumer-centric and bottom-up perspective by giving the opportunity to consumers to freely choose the way they buy their electric energy. A community can also be formed by prosumers who want to collaborate, or in terms of operational energy management. This paper contributes with an overview of these new P2P markets that starts with the motivation, challenges, market designs moving to the potential future developments in this field, providing recommendations while considering a test-case.
Article
Efficient building energy management is very essential for energy saving and green society. This paper investigates a sustainable energy management for an energy building cluster with distributed transaction. The building cluster consists of several types of energy buildings, e.g., office, industrial, and commercial buildings. We firstly formulate utility functions for the buildings of consuming energy based on the characteristics of their controllable loads. Then a two-stage energy sharing strategy is presented. In the first stage the total social energy cost is minimized through finding the optimal energy sharing profiles in a distributed way. In the second stage the clearing for mutual energy sharing is modeled as a non-cooperative game, and the existence of the equilibrium of the game is illustrated and a relaxation-based algorithm is introduced to search for the equilibrium. Moreover, a real-time model for each building to overcome real-time uncertainties such as renewable energy generation and base loads is provided. The simulation results show that the proposed energy sharing strategy is economically beneficial for the energy buildings, computationally efficient, and is promising to facilitate a sustainable regional building cluster.
Article
Self-consumption is an increasingly economically attractive solution for electricity generated by distributed energy resources, which also reduces transmission losses, power line overload and grid instability. Yet, at single-building scale, production and demand do often not coincide. Microgrids are proposed as a means to trade surplus energy produced with local prosumers and consumers, allowing to take advantage of complementary load profiles. Here we define a consumer as a single metered residential unit, a prosumer as a consumer with a photovoltaic system, and the microgrid as a collection of both prosumers and consumers. In this article, we systematically analyze to what extent varying sizes and prosumer-consumer ratios of microgrids affect local selfconsumption and self-sufficiency rates. To that end, we developed a simulation model that uses real-world load profiles from 4190 residential buildings as input data. We find that the prosumer-to-consumer ratio is more important than the absolute microgrid size, for microgrids sizes greater than ten. The results also indicate that prosumer-to- consumer ratios in the range of 40%-60% have the best performance for the various production to demand ratios analyzed. Each simulation is also compared to the baseline scenario of a stand-alone prosumer, which shows significantly better self-consumption ratios and self-sufficiency ratios for microgrids due to aggregation effects. Finally, this work may also be used as a reference to design residential microgrid communities for various prosumer-consumer compositions and various production-to-demand ratio.