Peer-to-peer energy trading with dynamic operating envelope: Report

Distributed energy resources (DERs) uptake present unique opportunities for

aggregators and retailers to innovate and compete in new service offerings
which may generate additional consumer benefits, such as allowing
consumers to be paid for selling power back to the energy grid. To do this,
distribution network service providers (DNSP), have been exploring the
development of Dynamic Operating Envelopes (DOE), which would allow
DNSP to set export limits dynamically and vary them by time and location.

By applying DOE, it would allow existing network capacity to be used more

efficiently such as curtailing DER output for a shorter period of time or over a
smaller area than would otherwise be the case; or by allowing higher DER
exports when network conditions permit doing so. The application of DOE
permits the energy market to maximize DER generation while managing the
risk of power flows exceeding constraints.

1. Peer-to-peer energy trading: A brief

Peer-to-peer (P2P) energy trading allows the integration of distributed
energy resources (DERs) into a distribution network. P2P energy trading
has become the next generation of energy management strategies that
considers economically beneficial aspects for prosumers. By going
through an economic approach, P2P energy trading can be a win-win
solution for system operator and electricity customers as well. However,
due to the increasing number of DERs, the expansion of P2P energy
trading becomes inevitable.
The expansion of P2P energy trading in terms of size faces some
challenges, such as congestion issues owing to bidirectional flows and
network stability issues owing to intermittent DER. While P2P energy
trading is applied on the existing network, the trading results may
significantly affect the condition of the network, especially with the
greater trading capacity involved. Therefore, there should be interaction
between market operator and system operator to exchange important
information to perform P2P energy trading as well as maintaining the
good condition of the distribution network.
P2P energy trading considers the operational time interval of [1, T ] for the
prosumer i and consumer j interacting beneath the multiple cluster c . A
prosumer is defined as an electricity customer who not only consumes
energy but also produces electricity. In this regard, consumers only
consume electricity. However, both prosumers and consumers are
presented with strategic behaviors, which means that they are unable to
 learn from others and use that knowledge in their trading decisions. Thus,
the market operator controls the strategic behaviors of consumers and
prosumers.
Fig. 1 illustrates the communication flow framework between all agents in
the proposed P2P energy trading system. Upper-level communication
denotes the interaction between the market and grid operators in
exchanging information regarding cluster P2P energy trading results ( [P ])
and total network usage cost ( Σ λ NUC). Lower-level communication involves
information-exchange activities between prosumers and consumers
through market operators. The market operator determines the optimal
aggregate trading capacity of prosumers and consumers ( P PI 2 P , P PJ 2 P ) based
on the submitted uniform trading price of P2P from the prosumers and
consumers ( λ PI 2 P , λ PJ 2 P ¿ and the allocated λ NUC of the P2P energy-trading
participants in each cluster. Thus, the market settlement is determined
based on the market results of λ PI ,2JP and P PI ,J2 P.

Figure 1. Communication flow of proposed P2P energy trading.

2. Proposed P2P energy trading with DOE
The proposed P2P energy trading with the application of operating
envelope is constructed by two main functions: upper-level and lower-
level functions. The division into two main functions aims to impose
separate decision making processes between system operator and market
operator. System operator controls the decision making of upper-level and
market operator controls the decision making of lower-level. In this case,
in the upper-level function the command or regulation towards P2P energy
trading will be given regarding the condition of the network. Where, the
 upper-level function determines the envelope-based capacity regulation
for allowed trading capacity. In addition to this, in the lower-level function,
the market operator will gather the responses from market participants
regarding the command from upper-level. The interactions between
upper-level and lower-level functions are determined as the Stackelberg
game where to ensure the convergence of the trading computation, the
alternating direction method of multipliers (ADMM) is applied. Figure 2
shows the flowchart of the proposed P2P energy trading with operating
envelopes.

Figure 2. Flowchart of proposed P2P energy trading with operating envelope.

- Upper-level function

Consider the tree-bus topology of a radial power distribution network

denoted by a set of buses and a set of branches
. In the tree-bus topology of the distribution network,
the number of line branches is . For every bus c , let P DOE ,c
denote its DOE-based controlled capacity, and Pc represent the active power
injection from bus c . In this study, the network-based cost considers only
active power injections to control the participants’ trading activity. Thus, the
upper-level network model is formulated as follows:
 OF :min Σ|P c −PDOE ,c| (1)
s . t . : P5 % , c ≤ PDOE , c ≤ P5 % , c P10 % , c ≤ P DOE , c ≤ P 10% ,c (2)
where, P5 % , c and P5 % , c denote the minimum and maximum allowed capacity
by the operating envelope which follows 5% tolerance of corresponding bus c
, respectively. Furthermore, P10 % , c and P10 % , c are the minimum and maximum
allowed capacity by the operating envelope which follows 10% tolerance of
corresponding bus c .

In congestion management, game-theory-based methods are considered to

obtain a cooperative approach for various agents through clusters. Each
cluster aims to incur the lowest cost in P2P energy trading to achieve a high
profit. Furthermore, the market operator believes that costs can be
minimized by avoiding penalties due to network congestion. Thus, the
market operator applies the Shapley value as a cooperative game theory to
evaluate the network-based cost of each cluster. Thus, the clusters
cooperate to minimize network congestion and maximize profits. The
Shapley value for the DOE limit violation level is formulated as follows:
S
|S|! ( n−|S|−1 ) !
SV c ( V ( S ) ) = ∑ ∙(V (S ⋃ {c })−V (S)) (3)
s ⊆ N {c} n!
where, S V c (V ( S )) is the Shapley value of cluster c with respect to the
characteristic value V (S).
- Lower-level function

To participate in P2P energy trading as sellers and buyers, the pricing

mechanisms of prosumers and consumers are represented as their marginal
profits derived from retail electricity prices. The trading price is determined
by the marginal buying and selling prices of buyers and sellers, respectively.
The marginal buying price indicates the maximum threshold that determines
whether the consumer will participate in P2P energy trading.
Max ¿ (4)
where x i , j and w i , j denote possible coalitions and weight of SV c between seller
i and buyer j , respectively. λ Pi , 2j P denotes the P2P price of seller i and buyer j .
λ i and λ j represent the marginal price of seller i and buyer j .
- Distributed optimization using ADMM

The integration of both systems was formulated using ADMM by considering

the objective functions of lower-level trading, f OF ,lo , and upper-level trading,
f OF ,up .Vector matrix of ADMM variables x , z , and u, scalar penalty parameter α ,
and Lagrange multiplier ρ . The overall ADMM model of cluster-based P2P
distributed control is formulated as follows:
OF :max { f OF ,up ( x ) + f OF ,lo ( z ) } s . t .x−z=0 x ∈ Pc ( V diff , c ) z ∈ Pc ( λUP , c ) (5)
Evidently, it is in the standard form of the ADMM, as presented in [7]. We
define the following augmented Lagrangian equation:
ρ 2
T
L ρ ( x , z , u )=f OF , upp ( x )+ f OF ,low ( z )+ u ( x+ z )+ ||x−z +u||2 (6)
2

where u denotes the scaled Lagrange dual multiplier and ρ represents the
scalar penalty parameter. These two variables play a significant role in
determining the performance of the update of the ADMM variables at each
iteration k , which can be formulated as follows:

{ 1
}
x k+1 ≔argmi n x Lρ ( x k , z ,u k ) + ( z−z k ) α ( z−z k )
2
(7)

≔argmi n {L ( x , z , u ) + ( x−x ) α ( x−x ) }

1
z k+1 z ρ
k k k k
(8)
2
uk +1 ≔u k + ρ ( x k+1−z k+ 1) (9)
(1
ε pri=10−3 ∙ √ C+10−3 ∙ max {‖x k+1‖2 ,‖z k +1‖2 }
0)
(1
ε dual=10−3 ∙ √ C+ ϵ rel ‖u k+1‖2
1)

3. Results

Figure 3. P2P trading results: load capacity

Figure 3 shows P2P energy trading results using operating envelope with
overall load capacity. Load 1 represents ‘flour factory’. Load 2 represents
‘joogwang’. Load 3 represents ‘gio’. Load 4 represents ‘lm’. Load 5
represents ‘tk’. The trading simulation is done for 24 hour with 5 minutes
interval for each trading.

Figure 4. P2P trading results: generation capacity

Figure 4 illustrates the generation capacity that is matched for selling in P2P
energy trading.

Figure 5. Comparison of P2P results and generation input: PV 1

Figure 5 presents the comparison between the original generation capacity

input and allowed selling capacity for P2P energy trading. As presented in
the figure, the selling capacity from PV 1 is significantly lower than the
original generation capacity input. This is due to the enforcement of
operating envelope limit. In addition, Figure 6 and Figure 7 present the line
graphs which compare the generation input and P2P energy trading with
envelope results.

Figure 6. Comparison of P2P results and generation input: PV 2

Figure 7. Comparison of P2P results and generation input: PV 3

 Figure 8. Comparison of P2P results and load input: Load 1

Figure 8 and figure 9 show the selling capacity of Load 1 as compared with
the original load capacity input and operating envelope limits, respectively.
The operating envelope limits the selling capacity of Load 1 significantly.
Where the operating envelope limit allocated to Load 1 has 2 ranges: 5% and
10% tolerances.

Figure 9. Comparison of P2P results and envelope limits: Load 1

 Figure 10. Comparison of P2P results and load input: Load 2

Similar conditions also happen to Load 2, due to the significantly low

operating envelope limit, the selling capacity of Load 2 is limited to very low
trading results. Figure 10 and figure 11 capture the effect of operating
envelope application towards selling capacity by comparing with the original
load capacity input and operating envelope, respectively.

Figure 11. Comparison of P2P results and envelope limits: Load 2

 Figure 12. Comparison of P2P results and load input: Load 3

Figure 12 shows the load capacity compared with the P2P with applying
operating envelope limit. The applied operating envelope limits enforced to
Load 3 is presented in Figure 13.

Figure 13. Comparison of P2P results and envelope limits: Load 3

 Figure 14. Comparison of P2P results and load input: Load 4

Figure 15. Comparison of P2P results and envelope limits: Load 4

Figure 14 compares the Load 4 possible buying capacity with the resulted
buying capacity based on P2P energy trading with operating envelope
application. According to the P2P energy trading results, Load 4 maximizes
trading capacity towards allowed capacity with respect to operating envelope
and thus resulting P2P results compared to operating envelopes as presented
in Figure 15.
 Figure 16. Comparison of P2P results and load input: Load 5

Figure 17. Comparison of P2P results and envelope limits: Load 5

Figure 16 shows the comparison of Load 5 with P2P energy trading with
operating envelope. According to Figure 17, Load 5 may have lower
economic benefits for sellers compared to Load 4 as it only buying capacity
most at lower 10% tolerance limit capacity.