Applied Energy 222 (2018) 383–395

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

A cooperative game approach for coordinating multi-microgrid operation T

within distribution systems
⁎
Yan Dua, Zhiwei Wangb, Guangyi Liub, Xi Chenb, Haoyu Yuanc, Yanli Weid, Fangxing Lia,
a
Dept. of Electrical Engineering and Computer Science, University of Tennessee, Knoxville, USA
b
GEIRI North America, San Jose, CA, USA
c
Peak Reliability, Loveland, CO, USA
d
Southern California Edison, Rosemead, CA, USA

H I GH L IG H T S

• AA coalitional operation model for multiple microgrids to achieve global optimum.

• A cost allocation method from cooperative game theory to achieve local optimum.
• Thelinearized optimal power flow with voltage constraints to realize cooperation.
• economy benefits of multi-microgrid cooperation are simulated and analyzed.

A R T I C LE I N FO A B S T R A C T

Keywords: This paper focuses on simulating the potential cooperative behaviors of multiple grid-connected microgrids to
Benders decomposition achieve higher energy efficiency and operation economy. Motivated by the cooperative game theory, a group of
Coalitional operation individual microgrids is treated as one grand coalition with the aim of minimizing the total operation cost. Next,
Cooperative game given that each microgrid operator is an independent and autonomous entity with the aim of maximum self-
Cost allocation
interest, a cost allocation method based on the concept of core in the cooperative game is implemented to ensure
linearized optimal power flow for distribution
(LOPF-D)
a fair cost share among microgrid coalition members, which guarantees the economic stability of the coalition.
Multi-microgrid Considering the combinatorial explosive characteristic of the cost allocation problem, Benders Decomposition
(BD) algorithm is applied to locate the core solution with computational efficiency. In addition, since microgrid
coalition is formed at the distribution system level, network losses is not negligible. After considering network
losses, the coalition operation model of multi-microgrid becomes an optimal power flow problem. A linearized
optimal power flow for distribution (LOPF-D) model is applied instead of the conventional ACOPF model to
reduce computation burden, meanwhile maintaining adequate accuracy. Case studies on standard IEEE systems
demonstrate the advantages of multi-microgrid cooperation and the robustness of the formulated grand coali-
tion. In addition, comparisons with the conventional ACOPF model verifies the high performance of the pro-
posed LOPF-D model.

1. Introduction loads, and refrigerators, which add considerable flexibility to microgrid

The worldwide energy and environmental crisis has led to the large-
scale development of renewable energy sources (RES) and distributed
energy resources (DER), which in return has brought microgrid tech-
nology under spotlight in the power industry. A microgrid is a small-
scale electric power system which contains distributed resources and
load, and can operate in either grid-connected mode or islanded mode
[1]. Currently, emerging new types of demand-side resources have been
spotted in microgrids, including electrical vehicles, air conditioning
operation. The advantages of integrating microgrids into distribution
systems are multifold: first, the DER units that reside in a microgrid can
support the local energy demand, hence reduces its reliance on the
upper-level utility grid and enhances the reliability of power supply;
second, it follows that microgrid facilitates environmentally-friendly
energy consumption by utilizing renewable energy-fueled generators,
i.e., wind turbines, photovoltaic panels, and fuel cells; last but not least,
by supplying the energy demand via local distributed generators (DGs),
microgrid can reduce long-distance transmission loss, as well as the

⁎
Corresponding author.
E-mail address: fli6@utk.edu (F. Li).

https://doi.org/10.1016/j.apenergy.2018.03.086
Received 20 December 2017; Received in revised form 27 February 2018; Accepted 25 March 2018
Available online 24 April 2018
0306-2619/ © 2018 Elsevier Ltd. All rights reserved.
Y. Du et al.
Nomenclature
πP(t), πQ(t) active/reactive power exchange price between dis-
tribution system and transmission system at interval t,
in $/MWh
Pgrid (t) active power exchange (power flow at PCC) at time
interval t, in MW
Qgrid (t) reactive power exchange (power flow at PCC) at time
interval t, in MW
PMT,m(t) micro turbine generation of the mth microgrid at time
interval t, in kW
Cgas fuel cost of micro turbine, in $/kWh
ηMT efficiency of micro turbine
min max
PMT ,PMT lower and upper bound of micro turbine, in kW
Pbmin,Pbmax lower and upper bound of boiler, in kW
νh heat to electricity ratio of micro turbine
Sth,m(t) energy level of thermal energy storage in mth microgrid
at time interval t, in kWh
Sthcap capacity of thermal energy storage in mth microgrid at
,m (t )
time interval t, in kWh
Sthmin
,m (t ) minimum energy level of thermal energy storage in mth
microgrid at time interval t, in kWh
Pth,m (t) charge/discharge rate of thermal energy storage in mth
microgrid at time interval t, in kW
ηh charge/discharge efficiency of thermal energy storage
Pb,m(t) boiler generation of the mth microgrid at time interval t,
in kW
investment on large-scale transformers and transmission lines.
Most recently, with the increasing penetration of renewable energy
into power systems, the concept of multi-microgrid (MMG) comes up on
the stage, which refers to a cluster of microgrids connected with each
other in close electrical or spatial distance [2,3]. The aim of MMG is to
achieve resilience and stability via fast power exchange and to further
obtain a high and smooth penetration of DERs into the bulk system.
Possible architectures for multi-microgrid regarding layout and inter-
faces accompanied by cost and reliability analysis are discussed in [4].
To achieve a coordinated penetration of multi-microgrid into the bulk
power system, a hierarchical control strategy is proposed in [5,6],
which includes the primary droop-control of power electronic devices,
the secondary control for voltage/frequency restoration and synchro-
nization, and the tertiary control of real and reactive power. The last
one is in association with microgrid energy management system, and
can be formulated as an economic dispatch problem with the aim of
maximizing economic profit.
The focus of this paper lies in the tertiary control level of a multi-
microgrid system. In retrospect, existing works mainly cover two topics
related to this field: planning and operation. In terms of the former, Ref.
[7] applies the Decision-Tree (DT) method to plan the capacities of
energy storage devices within microgrids to realize local power bal-
ance; Ref. [8] includes the coupling physical and operational con-
straints of electrical and heating/cooling networks for multi-energy
microgrids in the design of the capacities of DER units; Ref. [9] com-
bines both DER sizing and placement problems into one mixed integer
linear programming, where microgrid is modeled as a multi-node
system instead of an aggregated single-node model to better consider
power flow and heat flow balances.
With regard to operation, existing research works mainly adopt two
approaches to coordinate MMG economic dispatch: the centralized
approach and the decentralized approach. The main idea behind the
centralized optimization method is to aggregate all the entities into the
system as one unity with a collective objective. In the case of multi-
microgrid coordination, a central controller is selected (i.e. distribution
system operator, DSO) to organize the operation of all the DGs and
loads regardless of their individual interests. In this aspect, Ref. [10]
384
Applied Energy 222 (2018) 383–395
PmD (t ) electrical load of the mth microgrid at time interval t, in
kW
h
pLoad ,m (t ) thermal load of the mth microgrid at time interval t, in
kW
PPV,m(t) PV generation of the mth microgrid at time interval t, in
kW
PWT,m(t) wind turbine generation of the mth microgrid at time
interval t, in kW
Psolar,m(t) thermal solar energy of the mth microgrid at time in-
terval t, in kW
PiD (t ),QiD (t ) active/reactive load at bus i in distribution system at
time interval t
PiG (t ),QiG (t ) active/reactive generation at bus i in distribution
system at time interval t
rk(rij), xk(xij) the resistance and reactance of the kth line in the dis-
tribution system
PLk (t ),Q Lk (t ) active and reactive power flow on the kth line in the
distribution system at time interval t
Lk Lk
PLoss (t ),QLoss (t ) active and reactive line loss on the kth line in the
distribution system at time interval t
Lk Lk
∂PLoss / ∂PiG,∂PLoss / ∂QiG active loss factor of the kth line to the ith bus
generation
Lk Lk
∂QLoss / ∂PiG,∂QLoss / ∂QiG rective loss factor of the kth line to the ith bus
load
Vi(t) voltage magnitude of bus i at time interval t, in p.u.
Vimin,Vimax lower and upper voltage level of bus i, in p.u.
establishes a centralized control model of a group of microgrids that can
exchange power with their neighbors, where the objective is to max-
imize the total profit of all microgrid operators. Simulation results in-
dicate that local energy exchange improves individual operation
economy by making full use of the zero-cost renewable energy. In [11],
the interactions between the upper-level distribution system and the
multi-microgrid system are further considered, and the DSO is included
as an additional independent entity in MMG coordination. To decrease
model complexity and improve computational efficiency, decentralized
dispatch methods have been applied in [12,13], where the global op-
timization model is decomposed into several independent sub-problems
using Lagrange relaxation method and solved by local entities. Model
predictive control (MPC) scheme is implemented in [14,15] in a dis-
tributed manner to address the uncertainties of load and renewable
energy within the microgrids and to maintain a steady power exchange
with the rest of the distribution system. The authors in [16,17] explore
the optimal risk-constrained bidding strategies of microgrids for pro-
viding ancillary service to the utility grids using decentralized and
centralized approaches, respectively.
There exist some challenges with the above conventional models
[18]: in the centralized method, since it requires full communication
among all entities within the entire network, it is not scalable, espe-
cially not suitable for plug-and-play DERs like electrical vehicles; in the
decentralized method, since local entities independently work on their
own optimal dispatch schedule without the information from other
entities, this complete isolation from the rest of the system usually
cannot reach global optimum. In summary, the centralized method has
a simple implementation to realize global optimum, while the decen-
tralized method focuses on local optimum. Nevertheless, there remains
some gap between the two goals, which may sabotage the coordinated
operation of the multi-microgrid. The reason is that each microgrid is a
highly independent and profit-driven entity with the goal of max-
imizing its self-interest. Thus conflicts of interests between the local
microgrid (local optimum) and the system operator (global optimum)
may drive microgrid away from coordination.
The motivation of this work is to address the above mentioned
concerns between global optimum and local optimum. In this paper, we
Y. Du et al.
propose a cooperative game approach to implement multi-microgrid
coordinated operation. Non-cooperative game and cooperative game
are the two fundamental pillars of game theory. Both games intend to
reach a globally balanced status where no player can get any further
improvement of their interests, which is referred to as a Nash
Equilibrium (NE) in the former, and a core status in the latter. However,
the difference between the two lies in that the non-cooperative game
focuses on obtaining the maximum individual payoff for each single
player without evaluating the global welfare. While in the cooperative
game, a coalitional optimization model is first developed to reach the
global optimum, then a cost or profit allocation model is established to
fairly distribute the collective benefits among all the players to guar-
antee local optimum. Hence, it can be safely concluded that the co-
operative game approach is a natural fit for multi-microgrid coopera-
tion problem to mitigate the potential interest collisions between global
and local stake holders.
In retrospect, considerable efforts have been made in implementing
cooperative game into the field of power system, from transmission cost
allocation [19–21] to revenue distribution among the portfolio of
power generators [22,23]. The Shapley value from cooperative game
theory is applied in [24] to measure the flexibility of the fast-ramping
DERs. However, application of cooperative game in multi-microgrid
coordination is still at its initial stage of research, which leaves great
potentials for further explorations. In this regard, the authors in [25,26]
treat each microgrid as an active cooperative player seeking for po-
tential coalitions with their neighbors to share power and save trans-
mission cost. A merge-and-split algorithm is developed to guide the
formation of different coalitions under environmental changes. The
advantages of direct power exchange among local DERs and consumers
are discussed in [27], and the Shapley value is proposed as the optimal
cost saving division among the players, which belongs to the set of core
solution. A cooperative generation planning model for interconnected
microgrids is proposed in [28,29], in which both the long-term in-
vestment cost and short-term operational cost are included in the fair
cost distribution model. A Nash bargaining solution is implemented as
the optimal cost share scheme. An interactive energy game matrix
(IEGM) is developed in [30] to describe the available capacity reserve
that each microgrid can provide to their neighbors and the cooperation
level.
In this paper, we also apply the cooperative game theory to realize a
coordinated operation of multiple grid-connected microgrids at the
distribution system level. The present work is an extension from the
discussion in our previous work [31]. The problem under discussion
includes two major parts: the first one is to achieve the group
24 25
23
PCC
MG8
Transmission 1 2 3 4 5 6 7
System
(ISO) MG1
19 20 21 22
MG6 MG7
Distribution system
(DNO)
Fig. 1. IEEE 33-bus distribution system with 10 microgrids.
Applied Energy 222 (2018) 383–395
rationality by conducting a coalitional economic dispatch for multi-
microgrid system with the aim of minimizing the total operation cost;
the second part is to fairly distribute the cost among group members,
under the criterion that each single player, as well as each subset of the
players, is able to receive some cost savings from the multi-microgrid
cooperation, which is in essence the local optimal solution. Therefore,
they are more willing to cooperate with their neighbors, such that the
economic stability of multi-microgrid cooperation can be enhanced.
In addition to the application of cooperative game approach to the
multi-microgrid coordination, another highlight of this work is that a
linearized optimal power flow for distribution (LOPF-D) model is im-
plemented to adequately model the network losses, since multi-micro-
grid system is connected to the bulk system at the distribution level,
where the distribution network losses are considerable as much as
5–12%. With a large-scale microgrid penetration into the power system,
the conventional ACOPF model becomes highly computationally in-
tensive. We validate that LOPF-D can be a substitute of ACOPF with
adequate accuracy and much higher computation efficiency, which
paves the way for potential real-world application of the proposed
multi-microgrid cooperation.
The rest of the paper is organized as follows: Section 2 illustrates the
multi-microgrid coalitional operation model that aims at maximizing
global economy; Section 3 presents the LOPF-D model with multi-mi-
crogrid penetration; Section 4 gives the mathematical description of the
fair cost allocation among microgrids based on the core concept in
cooperative game theory and the associated decomposition method to
obtain the optimal solution; Section 5 presents the case studies on IEEE
standard test cases; and finally, conclusions are given in Section 6.
2. Coalitional operation model of multi-microgrid system
2.1. A brief introduction of cooperative game theory
Cooperative game, or coalitional game, is the study concerned with
a group of rational players who coordinate their actions and pool their
winning, which consequently leads to the problem of how to divide the
extra earnings or total cost among the coalition members [32]. A co-
operative game consists of two essential elements: (1) a set of players N
= {1,2…,i,…,n} and (2) a characteristic function v that specifies the
value created by different subsets of the players. A coalition c refers to a
subset of the players. The grand coalition includes all the players. An
allocation x is a way to distribute the value created by grand coalition,
marked as v(1), among all the players. Several other related definitions
are listed as follows:
26 27 28 29 30 31 32 33 MG10
G2
MG2 MG9 MG5
8 9 10 11 12 13 14 15 16 17 18
MG3 MG4
PV panel Wind Micro
turbine turbine
Fuel cell Thermal Boiler
storage
385
Y. Du et al. Applied Energy 222 (2018) 383–395

(1) an allocation (x1, x2,…, xn) is individually rational if xi ≥ v(i) for all i; generator output. If the mth microgrid and its DGs does not belong to
n
(2) an allocation (x1, x2,…, xn) is efficient if ∑i = 1 xi = v(1); coalition c, cm is 0, and P{MT,b},m(t) equals 0. Otherwise we have
(3) an allocation (x1, x2,…,xn) is coalitionally rational if ∑i∈c xi ≥ v(c) for P{min max
MT ,b} ⩽ P{MT ,b},m (t ) ⩽ P{MT ,b} .
all the subsets c;
(4) an allocation (x1, x2,…,xn) is said to lie in the core of the game if it (3) Charge/discharge constraints of thermal energy storage:
satisfies all three conditions above.
Sth,m (t ) = Sth,m (t −1)−ΔtPth,m (t )/ ηth (5)
From the above definitions, it can be discovered that a core allo- Sthmin ⩽ Sth,m (t ) ⩽ Sthmax (6)
cation ensures that every player in the grand coalition benefits more
than in the case when they each work alone or form coalitions. Thus, no Pth,m (t ) ⩽ cm Pthdmax,Pth,m (t ) ⩾ −cm Pthcmax (7)
one would be willing to leave the grand coalition and its stability can be
ensured. Sth,m (NT ) ⩾ Sth,m (0) (8)
In the case of multi-microgrid coordination, the players are in-
Eq. (5) is the inter-temporal constraint of energy level in the thermal
dividual microgrid operators, and all players are assumed to auto-
storage. For simplicity, we assume that the charging and discharging
matically form a grand coalition with ex-ante binding contracts. The
efficiency of the storage are the same. Eq. (6) implies that the energy
associated characteristic function is the total operation cost. The focus
level of the thermal storage should be within a certain range. Eq. (7)
of the problem is to find the core allocation of the total operation cost
indicates the maximum charge and discharge rate. Similar to Eqs. (3)
such that every microgrid operator can receive some cost saving from
and (4), 0–1 binary variable cm indicates whether the mth microgrid
cooperation. This section will specify the objective functions and con-
operator and its energy storage belongs to the microgrid coalition c or
straints of the operation model of multi-microgrid grand coalition, and
not. Eq. (8) requires that the energy level at the end of the operation
Section 4 will center on tackling the fair cost allocation problem.
cycle should be no lower than its initial value.

2.2. A coalitional operation model of multi-microgrid system

(4) Power balance constraint:
NMT NPV NWT
2.2.1. Objective function
A typical network topology of distribution system with 10 micro-
Pgrid (t ) + ∑ PMT ,m (t ) + ∑ cm PPV ,m (t ) + ∑ cm PWT ,m (t )
MT = 1 MT = 1 MT = 1
grids integrated is shown in Fig. 1. In this figure, each microgrid con- M n
tains different types of DGs on the generation side, including micro − ∑ cm PmD (t )− ∑ PiD (t )−PLoss (t ) = 0
turbines (MT), boilers, thermal energy storage (TES), wind turbines and m=1 i=1 (9)
photovoltaic solar panels, and electrical load and thermal load on the n
demand side. All microgrids form one grand coalition based on binding Qgrid (t )− ∑ QiD (t )−QLoss (t ) = 0
contract, and their objective is to minimize the total operation cost: i=1 (10)
NT Eqs. (9) and (10) are the active and reactive power balance con-
⎛
minv (c ) = ∑ ⎜πP (t ) Pgrid (t ) + πQ (t ) Qgrid (t ) straint of the distribution system, where PLoss(t) and QLoss(t) are the
t=1 ⎝ network losses. Since both terms are nonlinear and nonconvex func-
NMT Nb tions of the other control variables, i.e. generator output, which adds
⎛ PMT ,m (t ) Pb,m (t ) ⎞ ⎞
+ ∑ ⎜ ∑ Cfuel + ∑ Cfuel great model complexity. We further apply a linearized optimal power
ηMT ηb ⎟ ⎟ (1)
m∈c ⎝ MT = 1 b=1 ⎠⎠ flow for distribution (LOPF-D) model to overcome this computation
Eq. (1) calculates the total operation cost of microgrid coalition c, in difficulty. More details of the LOPF-D model will be revealed in Section
which the first two terms are the cost for distribution system to ex- 3. In Eq. (10), for simplicity, we assume that the reactive load is sup-
change active and reactive of power exchange with the transmission ported by power exchange from the transmission system, and local
system. The power exchange is evaluated as the power flow at the point microgrids only generate active power.
of common coupling (PCC) shown in Fig. 1. NT is the length of opera-
tion cycle, in this study it is set to 24 h. The following two terms are the (5) Bus voltage constraint:
generation cost of micro turbines and boilers that reside in microgrid m. Lk Lk
Vj (t ) = Vi (t )−((PLk (t ) + PLoss (t )) rk + (Q Lk (t ) + QLoss (t )) xk )/ V1 (t ) for i
= 2,3,...,n (11)
2.2.2. Constraints
Operation of microgrid coalition c should obey the following con- V1 (t ) = 1.05 p.u. (12)
straints to ensure system-wide economy and stability:
Vimin ⩽ Vi (t ) ⩽ Vimax (13)
(1) The scale of microgrid coalition:
Eqs. (11)–(13) is based on the line model illustrated in Fig. 2 [33].
0 < 1Tc ⩽ M (2) The line that has a tail bus k + 1 is numbered as Lk. For a line that
In Eq. (2), M is the total number of microgrids, c is an M × 1 vector connects a head bus i to a tail bus j (k + 1), the relationship between
Lk
composed of 0–1 binary indices [c1,c2,…,cm,…,cM]T, where cm indicates bus voltage Vi(t) and Vj(t) are shown in Eq. (11), where PLk (t ) + PLoss (t )
Lk
whether the mth microgrid belongs to the coalition c or not. It is obvious and Q Lk (t ) + QLoss (t ) stand for the total line power flow from bus i to
that in the grand coalition case, we have 1Tc = M.
Vi(t) rk+jxk Vj(t)(Vk+1(t))
(2) Capacity of DGs:
Lk Lk
min
cm PMT max
⩽ PMT ,m (t ) ⩽ cm PMT (3) PLoss (t ), QLoss (t )
PLk (t ) + jQLk (t )
cm Pbmin ⩽ Pb,m (t ) ⩽ cm Pbmax (4)
( Pi D (t ) − Pi G (t )) + j (QiD (t ) − QiG (t ))
Eqs. (3) and (4) are the capacity constraints of DGs, where P{min
MT ,b}
and P{max Fig. 2. Line model in distribution system.
MT ,b} are the lower bound and upper bound, respectively, of the

386
Y. Du et al. Applied Energy 222 (2018) 383–395

bus j. In addition, the vertical voltage drop is neglected, and the head Lk PL2k (t ) + Q L2k (t )
PLoss (t ) = rk
bus voltage is assumed to be close to the rated value. Vk2+ 1 (t ) (19)

(6) Heat balance constraints: Lk PL2k (t ) + Q L2k (t )

QLoss (t ) = xk
Vk2+ 1 (t ) (20)
∑ (νh PMT ,m (t ) + Pb,m (t ) + Pth,m (t ) + cm Psolar ,m (t ))− ∑ h
cm pLoad ,m (t )
m∈c m∈c The LF-D is calculated as the partial derivative of line losses to the
=0 (14) bus generation:
In Eq. (14), νh is the heat to electricity ratio of micro turbines, Lj−1
∂PLoss (t ) ∂PL (t ) ∂Q Lk (t ) ⎞ rk
h
Psolar,m(t) is the heat generated by solar energy, and PLoad = ⎜⎛2PLk (t ) Gk + 2Q Lk (t )
,m (t ) is the
⎟
∂PiG (t ) ⎝ ∂Pi (t ) ∂PiG (t ) ⎠ Vk2+ 1 (t ) (21)
thermal load of the mth microgrid system.
In the above coalitional operation model, when all microgrids form L
j−1
∂PLoss (t ) ∂PLk (t ) ∂Q Lk (t ) ⎞ rk
as one grand coalition, DGs, energy storages and loads owned by dif- = ⎜⎛2PLk (t ) + 2Q Lk (t ) ⎟
∂QiG (t ) ⎝ ∂QiG (t ) ∂QiG (t ) ⎠ Vk2+ 1 (t ) (22)
ferent microgrid operators are dispatched indiscriminately in a cen-
tralized manner. In this way, neighboring microgrids can provide en- L
j−1
∂QLoss (t ) ∂PL (t ) ∂Q Lk (t ) ⎞ xk
ergy support to each other via power exchange, and share excessive = ⎛⎜2PLk (t ) Gk + 2Q Lk (t ) ⎟
G
∂Pi (t ) ⎝ ∂Pi (t ) ∂PiG (t ) ⎠ Vk2+ 1 (t ) (23)
resources like low-cost renewable energy, which consequently lead to
reduction of total operation cost. On the other hand, the individual L
j−1
∂QLoss (t ) ∂PLk (t ) ∂Q Lk (t ) ⎞ xk
benefit of each microgrid operator is downplayed under the global = ⎛⎜2PLk (t ) + 2Q Lk (t ) ⎟
∂QiG (t ) ⎝ ∂Qi
G
(t ) ∂QiG (t ) ⎠ Vk2+ 1 (t ) (24)
optimization goal of minimizing the total cost. In order to guarantee the
interest of local microgrid operators, a fair cost allocation method is In Eqs. (21)–(24), index j is equal to k + 1, as is shown in Fig. 2. LF-
proposed in this paper, in which case any microgrid operator, and any D is related to the generation shift factors (GSF),
subset of microgrid operators can receive some cost saving by partici- ∂PLk (t )/ ∂PiG (t ),∂PLk (t )/ ∂QiG (t ),∂QLk (t )/ ∂PiG (t ) , and ∂QLk (t )/ ∂QiG (t ) ,
pating in the grand coalition. The detailed modeling of the cost allo- which is the sensitivity of line power flow to bus power injection, and
cation will be demonstrated in Section 4. can be calculated as follows:
n L
j−1
∂PLk (t ) ∂PLoss (t )
3. Linearized optimal power flow for distribution with multi- = −Sub (k + 1,i) + ∑ Sub (k + 1,j )
microgrid coalition ∂PiG (t ) j=k+2
∂PiG (t ) (25)
n j−1 L
As has been observed in Section 2, Eqs. (9) and (10), the active and ∂PLk (t ) ∂PLoss (t )
= ∑ Sub (k + 1,j )
reactive network losses PLoss(t) and QLoss(t) are nonlinear and non- ∂QiG (t ) j=k+2
∂QiG (t ) (26)
convex terms, which adds to model complexity. To improve computa-
n j−1L
tion efficiency, we apply the linearization technique derived from [33] ∂Q Lk (t ) ∂QLoss (t )
in the model, which are presented as follows:
= ∑ Sub (k + 1,j )
∂PiG (t ) j=k+2
G
∂Pi (t ) (27)
For the kth line in the distribution system, the active and reactive
line flow can be calculated as follows: ∂Q Lk (t ) n j−1
∂QLoss
L
(t )
= −Sub (k + 1,i) + ∑ Sub (k + 1,j )
n n
L
∂QiG (t ) j=k+2
∂QiG (t ) (28)
PLk (t ) = ∑ Sub (k + 1,j )·(P jD (t )−P Gj (t )) + ∑ j−1
Sub (k + 1,j )·PLoss (t )
j=k+1 j=k+2 It can be observed from Eqs. (21)–(28) that the calculation of gen-
(15) eration shift factors and loss factors are nested within each other. Hence
n n
a recursive method is applied to obtain their values: to begin with, both
Q Lk (t ) = ∑ Sub (k + 1,j )·(QjD (t )−QGj (t )) + ∑ j−1
Sub (k + 1,j )·QLoss
L
(t ) GSF and LF-D are set to 0. GSF are first calculated based on Eqs.
j=k+1 j=k+2 (25)–(28). Then LF-D are calculated based on both GSF and the line
(16) power flow results from a linearized power flow model, which can be
expressed as follows:
In Eqs. (15) and (16), both j and k are bus indices of the distribution
inj
system, n is the set of buses. Sub is a n × n matrix, where each element ⎡ P ⎤ = ⎡ B2 B1 ⎤ ⎡ δ ⎤
inj ⎥
(k + 1, j) denotes if bus j belongs to the sub-tree of bus (k + 1). ⎢
⎣ Q ⎦ ⎣− B1 B2 ⎦ ⎣V ⎦ (29)
The active and reactive line loss in Eqs. (15) and (16) can be ex- inj inj
In Eq. (29), P and Q are (n−1) × 1 vectors accounting for active
pressed as the following linearized function of bus generation:
and reactive power injection for all buses except for the slack bus. δ and
n Lj−1 n L
j−1 V are the voltage angle vector and voltage magnitude vector, both with
L L ∗ ∂PLoss ∂PLoss
j−1
PLoss j−1
(t ) ≈ PLoss (t ) + ∑ (PiG (t )−PiG ∗ (t )) + ∑ (QiG (t ) dimension of n × 1. The voltage angle and voltage magnitude at slack
i=1 ∂PiG i=1 ∂QiG
bus are set to 0 and 1 p.u., respectively. The B matrix contains the re-
−QiG ∗ (t )) (17) sistance and reactance information of the system, and is expressed as
follows:
n L
j−1 n L
j−1
L L ∗ ∂QLoss ∂QLoss rij x ij
j−1
QLoss j−1
(t ) ≈ QLoss (t ) + ∑ (PiG (t )−PiG ∗ (t )) + ∑ (QiG (t ) B1 (i,j ) = ,B2 (i,j ) = ,i≠j
i=1 ∂PiG i=1 ∂QiG rij2 + x ij2 rij2 + x ij2 (30)
−QiG ∗ (t )) (18) n n
rij x ij
In Eqs. (17) and (18), the partial derivative parts, i.e. B1 (i,i) = ∑ ,B2 (i,i) = ∑
Lj − 1 Lj − 1 Lj − 1 Lj − 1 j = 1,j ≠ i
rij2 + x ij2 j = 1,j ≠ i
rij2 + x ij2 (31)
∂PLoss (t )/ ∂PiG (t ),∂PLoss (t )/ ∂QiG (t ),∂QLoss (t )/ ∂PiG (t ) , and ∂QLoss
G
(t )/ ∂Qi (t ) , are called loss factors for distribution (LF-D), which de- Since the power injection equations do not include the slack bus, the
scribe the sensitivity of the (j−1)th line loss to the ith bus generation. B1 and B2 matrices are both (n−1) × n matrices with i traversing from
The LF-D can be obtained as follows: 2 to n and j traversing from 1 to n. Eq. (29) holds because originally we
The losses of the kth line are given by have:

387
Y. Du et al. Applied Energy 222 (2018) 383–395

∗
⎧ ⎛ Vi −Vj ⎞ ⎫ rij x ij Vi (Vi −Vj cosδij ) x ij2 Vi Vj sinδij
Pij = Re Vi ⎜ = · + ·
⎨ rij + jx ij ⎟ ⎬ rij2 + x ij2 x ij rij2 + x ij2 x ij
⎩ ⎝ ⎠⎭
(32)
∗
⎧ ⎛ Vi −Vj ⎞ ⎫ rij x ij Vi Vj sinδij
Qij = Im Vi ⎜ =− 2 ·
⎨ rij + jx ij ⎟ ⎬ rij + x ij2 x ij
⎩ ⎝ ⎠⎭
2
x ij Vi (Vi −Vj cosδij )
+ 2 ·
rij + x ij2 x ij (33)
Then, based on the assumptions that in the distribution system, the
bus voltage angle difference along one branch is close to zero, and the
bus voltage magnitude is around 1p.u., Eqs. (32) and (33) can be
simplified as follows:

rij x ij (Vi −Vj ) x ij2 (δi−δj )

Pij ≈ · +
rij2 + x ij2 x ij rij2 + x ij2
rij x ij (δi−δj ) x ij2
Qij ≈ − · +
rij2 + x ij2 x ij rij2 + x ij2
Which leads to the matrix formation of the relationship between bus
power injection and bus voltage, as is shown in Eq. (29). δ and V can
thus be calculated by multiplying the inverse of B and the LHS of Eq.
(29). Since this paper mainly focuses on multi-microgrid cooperation,
and the linearized power flow model is only used to calculate the net-
work losses during the cooperation, more justification of the model is
skipped to make the paper concentrated. Readers who are interested in
the power flow linearization method can refer to [33] for more details.
By obtaining the GSF and LF-D, the network losses can be calculated
using Eqs. (17) and (18). Both equations are in the form of first-order
Lj − 1 ∗ Lj − 1 ∗
Taylor expansion series, where PLoss (t ) and QLoss (t ) are the line losses
in the case where there is no generation output, i.e.
PiG∗ (t ) = QiG∗ (t ) = 0 , and can be computed using the linearized power
flow model.
Compared with [33], a considerable improvement of the model in
this paper is the development of voltage constraints (11)–(13), which
further guarantees that the multi-microgrid operation does not violate
physical constraints of a distribution system. In addition, according to
[33], in the LOPF-D model, the loss factors for distribution are obtained
from a “cold-start” algorithm, which means that they are estimated
values with no knowledge of the present operating point of the system.
This method can provide results accurate enough when there is no
voltage constraints in the model. However, after adding Eqs. (11)–(13),
it is discovered that the “cold-start” method becomes inapplicable. The
loss factors have to be repeatedly updated based on the current oper-
ating point in order to reach a solution close enough to the standard
ACOPF result. Hence, it leads to the design of an iterative algorithm.
The implementation of bus voltage constraints and the development of
the associated algorithm is one of the main contributions of this paper.
More details of the proposed iterative algorithm will be demonstrated
in Section 4.
To conclude, the original nonconvex coalitional operation model of
multi-microgrid system within the distribution system can now be re-
presented as the following LOPF-D model: minimizing (1), subject to
(2)-(18). Comparisons with the conventional ACOPF model in Section 5
verifies the accuracy and high computation efficiency of the proposed
LOPF-D model.
4. A Nucleolus-based fair cost allocation method
As has been stated in Section 2, the coalitional operation model of
multi-microgrid system realizes a global optimum without considering
the profits of local microgrids, which may violate the interest and au-
tonomy of individual microgrid operators. In this section we introduce
a Nucleolus-based cost allocation method derived from cooperative
·
x ij (34) Fig. 3. The core solution in a cooperative game (the red spot within the poly-
hedron). (For interpretation of the references to colour in this figure legend, the
(Vi −Vj ) reader is referred to the web version of this article.)
·
x ij (35)
game theory to realize a fair cost share among coalition members, as
well as to improve local economic benefit.
4.1. A Nucleolus-based cost allocation method
According to definitions (1)–(3) in Section 2.1, a core solution to a
cooperative game must meet the requirements of individual rationality,
efficiency and coalitional rationality. From a mathematical perspective,
since all the three constraints are affine functions, finding the core of
the game can be formulated as a convex optimization problem, where
the feasible region is a polyhedron in high dimensional space, as is
shown in Fig. 3. Under certain circumstances, multiple cores can exist,
and additional measures are required to evaluate their optimality.
While in other cases, the core set of the game may turn out to be empty.
In this section we will present a Nucleolus-based core solution to the
multi-microgrid cooperative operation case, as well as the associated
calculation method.
In this paper, we use Nucleolus as the potentially feasible core so-
lution to fair cost distribution among microgrids, as has been previously
applied in [21,34]. The Nucleolus allocation possesses several favorable
properties: its existence is unique, and if the core set of the game is
nonempty, it always belongs to the core. The main idea behind the
Nucleolus allocation is to first find out the coalition that is most dis-
satisfied with the current allocation of the total cost, which implies that
it may defect from the grand coalition. Then the grand coalition adjusts
the allocations to minimize the dissatisfaction of this coalition. Nu-
cleolus refers to this adjusted allocation. By conducting a Nucleolus al-
location, the dissatisfaction of any coalition is minimized, and the sta-
bility of the grand coalition is ensured.
The dissatisfaction of a coalition under the current cost allocation is
measured quantitatively as follows:
g (c,x ) = c Tx −v (c ) x = [x1,x2,...,x m,...,xM]T , 1Tx = v (1) (36)
As we can see, in Eq. (36), the dissatisfaction is the difference be-
tween the total cost allocated to a coalition c and the cost generated
when c operates separately from the grand coalition, where x is the
allocation vector indicating the cost distributed to each microgrid, and
v(1) is the total cost of the grand coalition, when every cm is equal to 1.
The expression is meaningful because a positive value of g(c,x) in-
dicates that staying in the grand coalition will cost the coalition c more
than the case in which it operates independently, therefore it may de-
fect from the grand coalition. The adjustment model of the cost allo-
cation based on dissatisfaction measurement can then be expressed as
follows:

388
Y. Du et al.
f = min δ
c,x
s. t . δ ⩾ g (c,x ) = c Tx −v (c )
0 < 1Tc ⩽ M−1c = [c1,c2,...,cm,...cM ], cm ∈ {0,1}
In Eq. (37), the decision variables are c and x. An auxiliary variable
δ is introduced to ensure that the maximum level of dissatisfaction
among all coalitions is minimized, as is shown in Eq. (38). In this way,
the final allocation x satisfies the coalitional rationality. Eq. (39) en-
compasses all the coalitions except for the empty set and the grand
coalition.
4.2. Finding Nucleolus solution via benders decomposition algorithm
The difficult part of Eqs. (37)–(39) is that the number of coalitions
grows exponentially with an increasing number of microgrid players,
which makes it impossible to obtain all the values of v(c) in (36). To
overcome this combinatorial explosion obstacle, we adopt a Benders
Decomposition (BD) method, which has been previously applied in
[35]. To begin with, Eq. (37) can be rewritten as follows:
min{max c Tx −v (c )}
x c
Which is a min–max problem, and can be further decomposed into
the following sub-problem and master problem:
Sub-problem:
g ∗ (x ) = max c Tx −v (c )⋯⋯UB (k )
c ∈ C ⧹ {1}
s. t . 0 < 1Tc ⩽ M−1
(2)−(18)
Master-problem:
f (k ) = min δ⋯⋯LB (k )
x (43)
s. t . δ ⩾ g ∗ (x (s) ) + ∂g ∗ (x (s) )T (x −x (s) ) , ∀ s = 1,2,...,k−1
x m ⩽ v (cm),cm = 
[0,0,...,1,...,0]
 
c1,c 2,...,cm ,...,cM
1Tx = v (1)
In the sub-problem (41), cost allocation x is fixed, and the variable
is the binary coalition index cm. The objective is the difference between
the allocated cost and the operating-alone cost, which stands for the
dissatisfaction of coalition c. Eq. (41) aims to find the coalition c that is
most dissatisfied under the current cost allocation x, and determines an
upper bound of the optimum.
In the master-problem (43), the dissatisfaction g(x) from sub-pro-
blem (41) is used as input, and the variable is the cost allocation x. The
objective is to minimize the maximum dissatisfaction among all the
previously found coalitions c, and it determines a lower bound of the
optimum. The algorithm iterates until the upper bound and the lower
bound converge to the same value. Since the master problem is linear,
the convergence of the algorithm is guaranteed. Eq. (44) is the Benders
cut generated from the sub-problem in the form of first-order approx-
imation of a Taylor series expansion, where ∂g∗(x(s))T is the partial
derivative of g∗ to x(s) in the sth iteration, and is equal to c(Ts) . Eqs. (45)
and (46) are the representations of individual rationality and efficiency,
respectively.
It can be observed that via Benders Decomposition method, there is
no need to calculate v(c) for all coalitions c, but only have to focus on
the coalition that is most likely to refuse to cooperate, which con-
siderably reduces computational efforts. Although the algorithm will
take several iterations to converge, the number of iterations is still
much less than the number of coalitions if applying an enumeration
method. The computation efficiency of Benders Decomposition method
Applied Energy 222 (2018) 383–395
will be further proved in the case study in Section 5.
(37)
It should be further pointed out that although the above Benders
(38) Decomposition method has already been applied in [35], the major
difference between the current work and Ref. [35] is that an economic
(39) dispatch model is involved in this work, since each microgrid individual
contains RESs, dispatchable distribute generators, and energy storage
devices, among which the dispatch of the latter two are variables,
which adds to higher complexity of the model, and leads to a multi-
variable, high dimensional, and nonconvex problem. While Ref. [35]
mainly focuses on renewable energy generation. In summary, in this
paper we include a more complicated economic dispatch model within
the fair allocation model to suit to multi-microgrid cooperation case,
and efficiently combine optimal power flow linearization technique
with Benders Decomposition to tackle the original high dimensional,
nonconvex problem.
Another explanation that has to be made is that although there
exists other core solutions in the cooperative game, such as Shapley
value [36] and Nash-Harsanyi solution [37], these solutions have some
undesirable properties like nonlinearity and nonconvexity, and cannot
be efficiently decomposed as in the case of Nucleolus. Hence in this
paper we mainly consider the Nucleolus allocation as the way of a fair
cost distribution.
(40)
4.3. A panorama of solution process for fair cost allocation
It can be noticed that the above sub-problem (41) also includes the
constraints from LOPF-D model, i.e. constraints (2)–(18), since the sub-
(41) problem involves economic dispatch of coalition c to find v(c), which
requires an iterative algorithm to solve, as is stated in Section 3. The
major difference between the model described by Eq. (41) and the
(42) multi-microgrid coalitional operation model (1) is that in the former
only microgrid coalition c is connected with the distribution system,
while in the latter case all microgrids, namely the grand coalition is
connected with the distribution system. This leads to the result that in
Eq. (41), c is 0–1 binary variable, while in model (1) all cm indices are 1,
(44) since all microgrids are included in the grand coalition. Another trivial
difference is that Eq. (41) computes the difference between the cost
allocated to the coalition c when it stays in the grand coalition, and the
(45) total operation cost of c when it operates alone, while Eq. (1) calculates
the total operation cost of grand coalition.
(46)
The combination of LOPF-D constraints with Benders
Decomposition method leads to a double-loop iteration algorithm to
solve the fair cost allocation problem, as is shown in Fig. 4.
As can be seen from the figure, the first iteration takes place in the
sub-problem of fair cost allocation model, with a known cost allocation
x. The loss factors are constantly updated based the current operating
points, until the objective function g(x) = cTx – v(c) reaches a fixed
value. After the first iteration converges, we get the coalition c that has
the highest dissatisfaction with the current cost allocation x. We input g
(x) into the master problem and readjust the cost allocation x to
minimize this dissatisfaction. If the output from the master problem,
which is the minimized dissatisfaction, does not equal the result from
the sub-problem, we input the updated cost allocation x calculated from
the master problem into the sub-problem, and repeat the above steps.
The iterative calculation of the sub-problem and the master problem
constitutes the second loop. If solving the multi-microgrid coalitional
operation problem (1)–(18), only the first loop is required, since the
problem only calculates the total operation cost and there is no de-
composition of problem.
5. Case study
In this paper, the above established cooperative operation model of
the multi-microgrid system and the associated cost allocation method
are tested on both IEEE 33-bus distribution system and IEEE 123-bus
distribution system. The first case is used to verify the economic
389
Y. Du et al. Applied Energy 222 (2018) 383–395

Initialization effectiveness of multi-microgrid cooperation as well as the stability of

k 1, x(1)=v(1)·M-1·1
UB (k)
10e4, LB(k) 0 the formulated grand coalition. The second test case is intended to
demonstrate the computation efficiency of Benders Decomposition
Set the initial values for method for cost allocation.
loss factors for distribution
and generation shift factors

5.1. Simulation results of IEEE 33-bus distribution system

h=1

Solve the sub-problem (41), Update the loss factors In this test case, ten microgrids are connected to the distribution
subject to (2)-(18), (42) to based on the current system at different buses, and are formed as one grand coalition under
find the sub-coalition c with operating point The 2nd loop:
highest dissatisfaction Iteratively calculate ex-ante binding contracts, where there exist both electrical energy ex-
h=h+1
the sub-problem and
master problem change and thermal energy change among coalition members. The
No
(h)
|g (x)-g (h-1)
(x)| 1?
from BD to decide multi-microgrid system topology has already been shown in Fig. 1. The
the optimal cost
allocation power base of the system is 1 MVA. The voltage base is 12.66 kV. The
Yes
Set UB(k) to g(h)(x), record
The 1st loop: acquire parameters for DGs and energy storages are given in Table 1. The re-
the value of x(h), g(h)(x) and the operation cost for
sub-coalition c and its newable energy generators, i.e. wind turbines and PVs, are assumed to
g(h)(x)
dissatisfaction with work at MPPT mode with zero cost. Wind speed data and solar irra-
the current cost
Solve the master problem allocation x diation data are acquired from [38]. Load data is acquired from [39].
(43)-(46), Set LB(k) to f (k), x = x(k)
k=k+1
update x
5.1.1. Comparison between LOPF-D and ACOPF
No As discussed in Section 3, the coalitional operation model of a mult-
(k) (k)
|UB -LB | 2?
microgrid system is established as a LOPF-D problem. To ensure the
Yes
accuracy of the proposed LOPF-D model, we compare the results of the
Algorithm terminates
Output the optimal allocation x coalitional operation model (1)–(18) with a conventional ACOPF
model. We simulate the coalitional operation of the multi-microgrid
Fig. 4. Flow chart of fair cost allocation process. system for 7 consecutive days, with the time interval set as 1 h, which is
168 h in total. To fully validate that the benefits from coalitional op-
eration for each microgrid is not occasional, and can be maintained in
Table 1
the long term, we choose the time horizon as 168 h instead of the 24 h
Parameters of DG and energy storage.
used in the daily schedule. The comparison of the results from LOPF-D
DG P{min P{max
MT ,b} (kW) η (%) νh Cfuel ($/kWh) and ACOPF are shown in Table 2.
MT ,b}
(kW) In Table 2, the total cost refers to the operation cost for the simu-
lated 168 h. The relative error of bus voltage, active and reactive net-
MT 0 60 33 2.69 0.042
work losses are the maximum relative error among all the buses and
Boiler 0 100 85 –
over all the time intervals. As seen from the table, the results from
Energy Capacity cmax / P dmax
Pth η (%) Initial Δt(h)
th LOPF-D is very close to the one from ACOPF, which verifies the accu-
Storage (kWh) (kW) state
(kWh) racy of the former. Furthermore, it should be noted that the calculation
time of LOPF-D model is 4 times faster than the ACOPF model, which
TES 100 50 98 50 1 substantiates its high computation efficiency.

5.1.2. Optimal cost allocation among microgrids

Table 2
Comparison of results from LOPF-D and ACOPF (33-bus System).
We run the above multi-microgrid coalitional operation model and
fair cost allocation model in two different scenarios: a winter scenario
Relative errors (%) and a spring scenario. The two are different from each other in para-
Total cost: 0.0083 Bus voltage: 0.4339 meters including solar radiation, wind speed, electrical load and
Active network losses: 0.4761 Reactive network losses error: 0.4241 thermal load. The simulation length for both scenarios are 7 days, or
Calculation time (s)
168 h. The aim is to verify the economy of multi-microgrid coalition
over a long-time scale in diversified real-world scenarios. The final cost
LOPF-D: 11.9291 ACOPF: 57.7869 savings for each microgrid member in the two scenarios are presented
in Table 3:

Table 3
Final cost savings based on Nucleolus method (33-bus System).
No. Nucleolus allocation ($) Independent operation ($) Cost saving (%) No. Nucleolus allocation ($) Independent operation ($) Cost saving (%)

Winter
1 340.6064 390.4244 12.7599 6 357.8588 408.4880 12.3943
2 −100.7641 −81.1947 24.1017 7 554.0574 565.9426 2.1001
3 345.8551 350.5885 1.3501 8 388.7359 407.0190 4.4920
4 211.5046 241.3738 12.3747 9 434.3381 466.2974 6.8539
5 182.0589 197.0653 7.6149 10 299.9779 315.1441 4.8125

Spring
1 174.4132 204.9736 14.91 6 108.6317 126.5169 14.14
2 −47.6777 -28.422 67.75 7 177.4422 178.7913 0.75
3 313.0457 314.3863 0.43 8 241.6237 244.3965 1.13
4 278.315 278.6241 0.11 9 141.3924 141.3924 0.00
5 203.6429 229.6581 11.33 10 225.1895 225.1895 0.00

390
Y. Du et al. Applied Energy 222 (2018) 383–395

300
600
300
200
(kWh)

400

(kWh)
(kWh)
200

100

solar
MT

200

MT
100
P

P
P
0 0
0
0 0 0
2 2 2
4 4 4
6 6 6
8 8 8
10 10 10
12 12 12
14 14 14
16 16 16
18 18 18
20 67 20 7 20 7
22 45 22 56 22 56
Day 24 23 Day 24 23
4 Day 24 34
1 1 12
Time(h) Time(h) Time(h)

Fig. 5. Comparison of MT generation in coalitional operation and independent operation.

1
P grid
D
Pm
0.8
P
WT

0.6
P grid (p.u.)

0.4

0.2

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Time(h)

Fig. 6. DSO power exchange with transmission system.

Fig. 7. Daily dynamic changes of TES state of charge in coalitional operation case.

As is shown in the table, under the coalitional operation, each mi- sell power back to the distribution system to support power consump-
crogrid receives some cost saving in different scenarios, which sub- tion in its neighboring microgrids and make profits.
stantiates the economic efficiency of multi-microgrid cooperation.
Notice that in both winter and spring scenario cases, microgrid 2 has a 5.1.3. Economy analysis of multi-microgrid cooperation
negative cost under both Nucleolus allocation and independent opera- We summarize two reasons behind the operation economy of the
tion. This is because when microgrid 2 has extra power supply, it can multi-microgrid coalitional operation. First of all, mutual power

391
Y. Du et al.
Fig. 8. Daily dynamic changes of TES state of charge in independent operation case.
Fig. 9. Bus voltage increase due to multi-microgrid penetration (winter scenario).
exchange among local microgrids increases the utilization efficiency of
the zero-cost renewable energy and consequently reduces generation
cost.
We first take a look at the micro turbine generation of microgrids in
different situations. Fig. 5 demonstrates the micro turbine generation of
microgrids in the winter scenario in both coalitional operation case and
independent operation case, as well as the solar thermal generation. In
the independent operation case, each microgrid has to supply the local
thermal load on their own, with no power exchange with other mi-
crogrids. As can be observed from the three figures, during the time
period with high solar thermal energy (i.e., 9 a.m.–4p.m.), the coali-
tional operation case shows much less micro turbine generation than
the independent case (the former is 0). This is because when forming as
one grand coalition, thermal energy is transferred from microgrids with
surplus solar panel generation to the ones with higher thermal load in
the grand coalition, while in the independent case, since there is no
local energy exchange, the more expensive micro turbine has to be
applied to provide thermal energy.
The utilization of wind power in the multi-microgrid coalitional
operation case is further demonstrated in Fig. 6. As can be observed
from the figure, the power exchange between DSO and transmission
system, Pgrid, follows the tendency of the microgrid demand PmD . In
addition, when the wind power generation is high, i.e. in the 17th,
84th, 85th and 98th hour, DSO purchased less power from the
392
Applied Energy 222 (2018) 383–395
transmission system; while during other periods with lower wind power
generation, more power is sent to the DSO to support the microgrid
power demand. The correlation coefficients between DSO power ex-
change and microgrid demand is 0.6124, and the correlation coefficient
between DSO power exchange and wind power is −0.7253, which
validates the above observations. Hence it can be concluded that the
interconnections among microgrids in the coalitional operation case
makes it possible for wind power share, which in return reduces power
purchase cost.
The second reason for the cost saving effects is related to the energy
storage devices in the microgrids, and is shown in Figs. 7 and 8:
The above two figures demonstrate the daily dynamic changes of
four thermal energy storages owned by MG1-MG4 in the winter sce-
nario and in both coalitional operation case and independent operation
case. As is shown in the figure, the energy storages reaches their
maximum capacity more often in the coalitional operation case than in
the independent operation case, which is especially evidently shown in
TES4. This is because when operating cooperatively, the surplus
thermal solar power of the microgrid without energy storage can be
fully stored by the energy storage of another microgrid via local power
exchange, therefore makes full use of zero-cost renewable energy and
saves generation cost. While in the independent operation case, since
each microgrid has a large amount of thermal load to supply, there is
less extra thermal power to be stored.
Y. Du et al.
Fig. 10. IEEE 123-bus system with 30 microgrids.
Table 4
Comparison of results from LOPF-D and ACOPF (123-bus system).
Relative errors (%)
Total cost: 0.0096 Bus voltage: 0.6196
Active network losses: 0.7584 Reactive network losses error: 0.7467
Calculation time (s)
LOPF-D: 15.267 ACOPF: 355.7584
Further comments on the above analysis is that although in Fig. 1
the 10 microgrids are connected with the distribution system at specific
locations, if randomly located, the economy of cooperative operation
will still take place because the above two reasons behind the operation
economy improvement does not relate to microgrid locations. This has
been proved by simulations with microgrids arbitrarily located in dif-
ferent buses. For the sake of conciseness, the simulation results, which
are much similar to Table 3 are skipped in the paper.
5.1.4. Impacts of multi-microgrid penetration on distribution system
We first investigate the impact of multi-microgrid penetration on
the system losses. Network losses rate is implemented here, which is
defined as the ratio of network work losses to the total power genera-
tion. In the winter scenario case, when there is no microgrids in the
distribution system, the network losses rate is 4.65%; with 10 micro-
grids, the maximum hourly network losses rate over the entire simu-
lation time span (168 h in total) is 4.39%. Hence it can be concluded
that multi-microgrid penetration can help reduce network losses. This is
because local energy consumption can be supplied by DGs in micro-
grids, therefore avoids long-distance power transmission and decreases
network losses.
We further consider the impact of multi-microgrid penetration on
system voltage level. Fig. 9(b) demonstrates the percentage of bus
voltage increase in the multi-microgrid coalitional operation case
compared with the original voltage level of the distribution system in
the winter scenario. The grey plane in the figure is the zero-level. As
seen from the figure, the majority of bus voltages experienced a voltage
increase during the entire simulation time span, with the maximum
393
Applied Energy 222 (2018) 383–395
voltage increase percentage reaching 1.11%. Notice that at some buses,
the voltage increase is below zero, which indicates a voltage decrease.
This is because of the penetration of the microgrid demand. Still,
Fig. 9(a) shows that all the bus voltage levels are within the feasible
region [0.95p.u., 1.05p.u], which indicates that multi-microgrid pene-
tration can provide reliable voltage support to the distribution system.
5.2. Simulation results of IEEE 123-bus distribution system
One of the key contributions of our paper lies in the application of
Benders Decomposition method to overcome the combinatorial explo-
sion in fair cost allocation among microgrid members. To fully verify
the efficiency of the applied algorithm, we further test the coalitional
operation model and the fair cost allocation model on a larger-scale
123-bus distribution system, with 30 microgrids involved in the co-
operation, which leads to a tremendous number of coalitions, i.e. 230 –
1 in total. The topology of the 123-bus system is shown in Fig. 10:
In Fig. 10, 30 microgrids are connected to IEEE-123 bus distribution
system and operate as one grand coalition. Similarly to the 33-bus test
case, in this test case the multi-microgrid cooperation is simulated in
two scenarios with a time span of 168 h. The comparison with ACOPF
and final cost allocations are shown in Tables 4 and 5:
As can be seen from Table 4, the LOPF-D has adequate accuracy
even on a larger system with far more variables, and the computation
speed is more than 20 times faster than the ACOPF. In Table 5, in the
spring scenario, MG9-10, MG19-20, MG29-30 receive zero cost saving.
Although those microgrids cannot receive any cost savings in the spring
scenario, they indeed benefit from coalitional operation in winter sce-
nario (the cost saving percentage is greater than 0). If viewed from a
long-term perspective, they would still be willing to stay in the grand
coalition. For all the other microgrids, they receive some cost savings in
both winter scenario and spring scenario by cooperation.
The computation efficiency of the Benders Decomposition method
in finding Nucleolus cost allocation is demonstrated as follows: in the
sub-problem (41), the number of constraints is 115,839 and the total
number of variables is 113,431, including 30 binary variables; in the
master-problem (43), the total number of variables is 30, the number of
constraints grows with each iteration. The simulation is carried out on a
hybrid platform, MATLAB 2016a plus GAMS 24.7, where MATLAB is
Y. Du et al. Applied Energy 222 (2018) 383–395

Table 5
Final cost savings based on Nucleolus method (123-bus system).
No. Nucleolus allocation ($) Independent operation ($) Cost saving (%) No. Nucleolus allocation ($) Independent operation ($) Cost saving (%)

Winter
1 364.5519 401.2525 9.1465 16 358.9493 401.7175 10.6463
2 −78.8023 −72.5693 8.5891 17 555.9820 559.5865 0.6441
3 367.8911 375.0287 1.9032 18 388.8479 404.5199 3.8742
4 234.4722 265.9025 11.8202 19 434.9585 468.4101 7.1415
5 186.4434 200.3160 6.9254 20 299.8734 314.5705 4.6721
6 359.0958 403.6232 11.0319 21 362.1107 393.5861 7.9971
7 556.2037 560.5060 0.7676 22 −78.1103 −73.9583 5.6139
8 388.2792 405.1344 4.1604 23 368.6778 372.2232 0.9525
9 436.2354 468.2867 6.8444 24 235.1628 263.7281 10.8313
10 301.5989 317.1347 4.8988 25 186.5898 197.8237 5.6787
11 363.6436 396.8117 8.3586 26 359.7733 398.0551 9.6172
12 −78.2742 −73.3835 6.6646 27 555.0086 556.3440 0.2400
13 367.4488 373.4819 1.6154 28 387.3975 402.9131 3.8509
14 235.3319 265.7271 11.4385 29 434.9293 465.7893 6.6253
15 184.9453 197.1955 6.2122 30 302.4844 315.0619 3.9921

Spring
1 195.8652 212.6706 7.9021 16 109.0740 124.9667 12.7176
2 −26.9551 −22.5069 19.7638 17 176.8997 178.4936 0.8930
3 334.1948 336.9596 0.8205 18 240.5403 243.7097 1.3005
4 298.5242 300.4783 0.6503 19 141.3924 141.3924 0
5 206.3893 231.4749 10.8373 20 225.1895 225.1895 0
6 109.0474 125.4055 13.0442 21 195.8315 209.8261 6.6696
7 176.8823 178.5370 0.9268 22 −26.9405 −22.9849 17.2092
8 240.4132 243.8768 1.4202 23 334.2929 335.2857 0.2961
9 141.3924 141.3924 0 24 298.3843 299.0838 0.2339
10 225.1895 225.1895 0 25 206.3761 230.0033 10.2725
11 196.3429 211.0214 6.9559 26 108.8652 124.1300 12.2975
12 −26.8926 −22.7858 18.0239 27 176.8709 178.3374 0.8223
13 334.1976 336.0424 0.5490 28 240.2067 243.2668 1.2579
14 298.3681 300.3638 0.6644 29 141.3924 141.3924 0
15 206.0727 229.6383 10.2621 30 225.1895 225.1895 0

Table 6 contribute to the generation cost reduction;

Computation efficiency of benders decomposition method (123-bus system).
Simulation scenario Number of iterations in benders Computation time (s)
decomposition
Winter 278 12,316
Spring 178 7482
used for creating input data profiles and recording computation results.
The above min–max problem is solved by CPLEX on GAMS. The hard-
ware environment is a laptop with Intel®Core™ i5-6300U 2.4 GHz CPU,
and 4.00 GB RAM. Computation time and iterations for calculating the
Nucleolus cost allocation in each scenario is provided in Table 6:
In the 123-bus test case, the total number of coalitions is
2^30–1 = 1.0737 × 109. From Table 6, we can see that only a small
number of coalitions needs to be checked before reaching the core so-
lution, i.e. 278 in the winter scenario and 178 in the spring scenario.
Therefore, we may safely conclude that the Benders Decomposition
method holds considerable computation efficiency and can fairly be
applied to large-scale system simulation.
6. Conclusion
This paper highlights the potential advantages of cooperation
among multiple microgrids with different types of distributed energy
resources at the distribution system level. The main contributions of
this paper are summarized as follows:
(2). To secure the stability of the multi-microgrid coalition, a fair al-
location of the total operation cost among microgrids is further
proposed, which is a core solution to the cooperative game.
Benders decomposition algorithm is applied to obtain the core
solution with high computational efficiency. In this way, the local
optimum is achieved to guarantee the benefit of each microgrid
player;
(3). We apply a linearized optimal power flow for distribution (LOPF-
D) model in both multi-microgrid coalitional operation model and
fair cost allocation model to include distribution network losses,
since at distribution level the losses accounts for a significant part.
Comparison with the ACOPF model verifies the accuracy of the
proposed LOPF-D model as well as its computation efficiency.
The proposed work and conclusions hold considerable implications
for real-world applications: (1) with the increasing penetration of the
uncertain and intermittent renewable energy into the distribution
system, a coalitional operation mode of multi-microgrid provides a vi-
able solution to efficiently consume the surplus green power in case of
over generation, which can reduce the fuel cost of dispatchable gen-
erators and also the dependence on the main grid for power supply; (2)
the propose Nucleolus cost allocation method can realize local optimum
and guard the interest of individual microgrid, which constitutes a
strong prerequisite to fully realize a fair and profitable multi-microgrid
cooperation. Hence, the proposed model and solution can be well ex-
plored for a practical application.

(1). Inspired by cooperative game theory, a coalitional operation model Acknowledgement

of grid-connected microgrids is constructed to minimize the total
operation cost. Simulation results verify that via local power ex- This work is partly supported in part by SGCC Science and
change among microgrids, the utilization efficiency of renewable Technology Program and in part by CURENT, a US NSF/DOE
energy and also energy storage devices can be increased, which Engineering Research Center under the NSF award EEC-1041877.

394
Y. Du et al.
References
[1] IEEE Standards Association. IEEE guide for design, operation, and integration of
distributed resource island systems with electric power systems. IEEE Std, 1547;
2011.
[2] Gil NJ, Lopes JP. Hierarchical frequency control scheme for islanded multi-micro-
grids operation. In: 2007 IEEE Power Tech; 2007. p. 473–78.
[3] Ren L, Qin Y, Li Y, Zhang P, Wang B, Luh PB, et al. Enabling resilient distributed
power sharing in networked microgrids through software defined networking. Appl
Energy 2018;210:1251–65.
[4] Bullich-Massagué E, Díaz-González F, Aragüés-Peñalba M, Girbau-Llistuella F,
Olivella-Rosell P, Sumper A. Microgrid clustering architectures. Appl Energy
2018;212:340–61.
[5] Levron Y, Guerrero JM, Beck Y. Optimal power flow in microgrids with energy
storage. IEEE Trans Power Syst 2013;28:3226–34.
[6] Sanjari MJ, Gharehpetian GB. Game-theoretic approach to cooperative control of
distributed energy resources in islanded microgrid considering voltage and fre-
quency stability. Neural Comput Appl 2013;25:343–51.
[7] Moutis P, Skarvelis-Kazakos S, Brucoli M. Decision tree aided planning and energy
balancing of planned community microgrids. Appl Energy 2016;161:197–205.
[8] Qiu J, Zhao J, Yang H, Wang D, Dong ZY. Planning of solar photovoltaics, battery
energy storage system and gas micro turbine for coupled micro energy grids. Appl
Energy 2017.
[9] Mashayekh S, Stadler M, Cardoso G, Heleno M. A mixed integer linear programming
approach for optimal DER portfolio, sizing, and placement in multi-energy micro-
grids. Appl Energy 2017;187:154–68.
[10] Ouammi A, Dagdougui H, Dessaint L, Sacile R. Coordinated model predictive-based
power flows control in a cooperative network of smart microgrids. IEEE Trans
Smart Grid 2015;6:2233–44.
[11] Wang Z, Chen B, Wang J, Begovic MM, Chen C. Coordinated energy management of
networked microgrids in distribution systems. IEEE Trans Smart Grid
2015;6:45–53.
[12] Wu J, Guan X. Coordinated multi-microgrids optimal control algorithm for smart
distribution management system. IEEE Trans Smart Grid 2013;4:2174–81.
[13] Fathi M, Bevrani H. Statistical cooperative power dispatching in interconnected
microgrids. IEEE Trans Sustain Energy 2013;4:586–93.
[14] Kou P, Liang D, Gao L. Distributed EMPC of multiple microgrids for coordinated
stochastic energy management. Appl Energy 2017;185:939–52.
[15] Holjevac N, Capuder T, Zhang N, Kuzle I, Kang C. Corrective receding horizon
scheduling of flexible distributed multi-energy microgrids. Appl Energy
2017;207:176–94.
[16] Pei W, Du Y, Deng W, Sheng K, Xiao H, Qu H. Optimal bidding strategy and in-
tramarket mechanism of microgrid aggregator in real-time balancing market. IEEE
Trans Ind Inform 2016;12:587–96.
[17] Wang J, Zhong H, Tang W, Rajagopal R, Xia Q, Kang C, et al. Optimal bidding
strategy for microgrids in joint energy and ancillary service markets considering
flexible ramping products. Appl Energy 2017;205:294–303.
[18] Nikmehr N, Najafi-Ravadanegh S, Khodaei A. Probabilistic optimal scheduling of
networked microgrids considering time-based demand response programs under
uncertainty. Appl Energy 2017;198:267–79.
[19] Tsukamoto Y, Iyoda I. Allocation of fixed transmission cost to wheeling transactions
by cooperative game theory. IEEE Trans Power Syst 1996;11:620–9.
395
Applied Energy 222 (2018) 383–395
[20] Zolezzi Juan M, Rudnick H. Transmission cost allocation by cooperative games and
coalition formation. IEEE Trans Power Syst 2002;17:1008–15.
[21] Rao MSS, Soman SA. Marginal pricing of transmission services using min-max
fairness policy. IEEE Trans Power Syst 2015;30:573–84.
[22] Street A, Lima DA, Freire L, Contreras J. Sharing quotas of a renewable energy
hedge pool: A cooperative game theory approach. In: 2011 IEEE PowerTech; 2011.
p. 1–6.
[23] Lee W, Xiang L, Schober R, Wong VWS. Direct electricity trading in smart grid: a
coalitional game analysis. IEEE J Selected Areas Commun 2014;32:1398–411.
[24] Kristiansen M, Korpås M, Svendsen HG. A generic framework for power system
flexibility analysis using cooperative game theory. Appl Energy 2018;212:223–32.
[25] Saad W, Zhu H, Poor HV. Coalitional game theory for cooperative micro-grid dis-
tribution networks. In: 2011 IEEE international conference on communications
workshops; 2011. p. 1–5.
[26] Ni J, Ai Q. Economic power transaction using coalitional game strategy in micro-
grids. IET Generation, Transm Distribution 2016;10:10–8.
[27] Derakhshandeh SY, Golshan MEH, Masoum MA. Profit-based unit commitment with
security constraints and fair allocation of cost saving in industrial microgrids. IET
Sci, Meas Technol 2013;7:315–25.
[28] Wang H, Huang J. Cooperative planning of renewable generations for inter-
connected microgrids. IEEE Trans Smart Grid 2016;7:2486–96.
[29] Zenginis I, Vardakas JS, Echave C, Morató M, Abadal J, Verikoukis CV. Cooperation
in microgrids through power exchange: An optimal sizing and operation approach.
Appl Energy 2017;203:972–81.
[30] Lv T, Ai Q. Interactive energy management of networked microgrids-based active
distribution system considering large-scale integration of renewable energy re-
sources. Appl Energy 2016;163:408–22.
[31] Du Y, Li F, Kou X, Pei W. Coordinating multi-microgrid operation within distribu-
tion system: a cooperative game approach. In: Power and energy society general
meeting, 2017 IEEE; 2017. p. 1–5.
[32] Brandenburger A. Cooperative game theory: characteristic functions, allocations,
marginal contribution. http://www.uib.cat/depart/deeweb/pdi/hdeelbm0/arxius_
decisions_and_games/cooperative_game_theory-brandenburger.pdf.
[33] Yuan H, Li F, Wei Y, Zhu J. Novel linearized power flow and linearized OPF models
for active distribution networks with application in distribution LMP. IEEE Trans
Smart Grid 2018;9:438–48.
[34] Dabbagh SR, Sheikh-El-Eslami MK. Risk-based profit allocation to DERs integrated
with a virtual power plant using cooperative Game theory. Electric Power Syst Res
2015;121:368–78.
[35] Freire L, Street A, Lima DA, Barroso LA. A hybrid MILP and benders decomposition
approach to find the nucleolus quota allocation for a renewable energy portfolio.
IEEE Trans Power Syst 2015;30:3265–75.
[36] O'Brien G, El Gamal A, Rajagopal R. Shapley value estimation for compensation of
participants in demand response programs. IEEE Trans Smart Grid 2015;6:2837–44.
[37] Madani K, Hooshyar M. A game theory–reinforcement learning (GT–RL) method to
develop optimal operation policies for multi-operator reservoir systems. J Hydrol
2014;519:732–42.
[38] Jager D, Andreas A. NREL national wind technology center (NWTC): M2 Tower;
Boulder, Colorado (Data); NREL Report No. DA-5500-56489; 1996. http://dx.doi.
org/10.5439/1052222.
[39] Ercot Backcasted (Actual) Load Profiles-Historical. Available: http://www.ercot.
com/mktinfo/loadprofile/alp.