IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 31, NO.

3, MAY 2016 2289

Stochastic Optimization of Renewable-Based

Microgrid Operation Incorporating Battery Operating
Cost
Tu A. Nguyen and M. L. Crow, Fellow, IEEE

Abstract—Integration of renewable energy resources in micro- of RE and ES. Specifically, due to the stochastic nature of the
grids has been increasing in recent decades. Due to the random- renewable resources such as solar and wind, the mismatch in
ness in renewable resources such as solar and wind, the power forecast and realized power may result in extra operating costs
generated can deviate from forecasted values. This variation may for committing costly reserve units or penalty cost for curtailing
cause increased operating costs for committing costly reserve units
demand. In addition, to better utilize the renewable energy in
or penalty costs for shedding load. In addition, it is often desired
to charge/discharge and coordinate the energy storage units in an the microgrid it is necessary to charge/discharge and coordi-
efficient and economical way. To address these problems, a novel nate multiple energy storage units in an efficient and economical
battery operation cost model is proposed which considers a battery way. To address these problems, the stochastic model of renew-
as an equivalent fuel-run generator to enable it to be incorporated able energy and load demands as well as the working character-
into a unit commitment problem. A probabilistic constrained ap- istics and operating cost of the ES devices should be incorpo-
proach is used to incorporate the uncertainties of the renewable rated into the UC and ED.
sources and load demands into the unit commitment (UC) and eco- Several approaches related to stochastic optimization of op-
nomic dispatch problems. eration for renewable-based microgrids have been conducted.
Index Terms—Energy storage, microgrids, renewable energy, In [4], the day-ahead scheduling of a microgrid is developed as
unit commitment. a two-stage stochastic problem in which the first stage identifies
the optimal dispatch for the distributed units while the second
stage considers the variability and uncertainty of photovoltaic
I. INTRODUCTION (PV) and wind energy generation. The probabilistic UC in [5]

D URING the past decades, the electric power industry has

undergone significant changes in response to the rising
concerns of global climate change and volatile fossil fuel prices.
is similarly formulated as a two-stage stochastic programming
problem in order to incorporate the uncertainty in load and PV
forecast. Forecast errors are modeled by normal distribution. A
For more efficient, reliable, and environmentally friendly en- two-stage stochastic programming is also used in [6] and [7].
ergy production, it is critical to increase the deployment of dis- Most of the existing studies are based on scenario-based
tributed generation, especially from renewable energy resources stochastic programming [4]–[9]. This approach is based on
(RE), as well as distributed energy storage (ES). This trend the replication of deterministic models across scenarios which
has evolved into the concept of a “microgrid” which can be are generated by Monte Carlo simulations. The computational
described as a cluster of distributed energy resources, energy burden in this approach increases exponentially with the
storage and local loads, managed by an intelligent energy man- number of investigated scenarios [2]. Scenario reduction using
agement system [1], [2]. different techniques might ease the problem of computational
Similar to bulk power grid operation, microgrid operation can overhead; however, this approach may overlook low-proba-
be determined by unit commitment (UC) and economic dispatch bility but high-impact scenarios. Therefore in this paper, we
(ED). The UC is performed from one day to one week ahead pro- propose a probabilistic constrained approach to incorporate
viding the start-up and shut-down schedule for each generation the uncertainties into the UC and ED for microgrids. In this
and storage unit, which can minimize the operating cost of the approach, the hard constraint of exact power balance is relaxed
microgrid. After the UC is determined, then ED is performed by introducing a probabilistic constraint which contains renew-
from few minutes to one hour in advance to economically allo- able powers and load demands as random variables. The power
cate the demand to the on-line units while considering all unit balance constraint is enforced with high probability while
and system constraints [3]. the penalty for the constraint violation is applied in the cost
Although operation optimization for bulk power systems has function. The advantage of this method over the scenario-based
been well studied in the literature, the traditional UC methods method is that all possibilities of load demands and renewable
cannot be applied directly to microgrids with high penetration generations are covered without the need to consider a large
number of scenarios. Furthermore a stochastic dynamic pro-
Manuscript received January 07, 2015; revised May 04, 2015 and May 20, gramming is used to solve the UC.
2015; accepted July 06, 2015. Date of publication July 29, 2015; date of current Most previous studies assume the energy storage system to
version April 15, 2016. This work has been supported in part by the Department be modeled with a constant efficiency and zero operating cost
of Energy SunShot program under DE-0006341 and in part by the National Sci- [4]–[16]. We propose a cost model for the battery that explic-
ence Foundation FREEDM ERC program. Paper no. TPWRS-00004-2015.
The authors are with the Department of Electrical and Computer Engineering, itly models the operating cost as a function of efficiency, life
Missouri University of Science and Technology, Rolla, MO 65401 USA. cycle, and state of health (SOH). This new cost model enables
Digital Object Identifier 10.1109/TPWRS.2015.2455491 the battery to be treated as an equivalent fossil fuel generator

0885-8950 © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 2290 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 31, NO. 3, MAY 2016

in the economic dispatch problem. This cost model also makes of storage capacity available:
economic dispatch for multiple batteries possible in the micro-
grid system without introducing additional objective functions (4)
to maximize their efficiencies and their life spans. Specifically,
the following contributions are made: where is the total lifetime cycling capacity of a battery.
• A novel battery operation cost model is proposed ac- By convention, an electrochemical battery, such as lead acid or
counting for charge/discharge efficiencies as well as lithium-ion, is often considered to be at the end of its life when
cycling life time of the batteries. The model enables the its has degraded to 80% of its rated energy capacity [17]. As-
battery to be treated as an equivalent fossil fuel generator suming that a battery will be discharged to its rated depth of
in the UC and ED. discharge every cycle, the average capacity degradation rate is
• A probabilistic constrained approach is proposed to incor- in which is the battery rated capacity and is
porate the uncertainties of the renewable sources and load the rated life time.
demands in microgrids into the UC and ED problems. The As opposed to a lead acid or a lithium-ion battery, a vanadium
UC is solved by stochastic dynamic programming. redox battery (VRB) has negligible capacity degradation from
repeated deep discharges and recharges. The cycle life of a VRB
II. BATTERY OPERATION COST MODEL mainly depends on the life expectancy of its proton exchange
For a small-scale fossil fuel generator in a microgrid, the op- membrane and its pumps. A VRB can last over 10 000 cycles
erating cost is typically the fuel cost. The cost can be character- until its membrane degrades or the pumps fail.
ized as a function of its output power [3], [10]: Therefore, the total lifetime usable capacity of a battery can
be estimated as follows:
• Lead acid and lithium-ion battery:

(1) (5)
where (in $/gal) is the fuel price, (in gal/h) is (6)
fuel consumption and is output power of generator .
As opposed to a generator, a battery consumes no fuel to op- • Vanadium redox battery:
erate. This makes it a challenge to evaluate the operating cost of
a battery. However, in terms of the energy conversion process, (7)
a battery and a generator are analogous. In a generator, energy
is stored in fuel form and generated into electricity via combus- where is the depth of discharge.
tion process. Similarly, in a battery, electricity is charged and The operating cost model of the battery is built based on the
discharged via an electrochemical process. In general, charging similarity with fuel cost model of a generator, therefore there
a battery is analogous to refilling fuel for a generator; thus, the is little added complexity over standard approaches. The kWhf
input electricity (kWh) can be considered as the “fuel” for the price for a battery does not change as frequently as fuel prices.
battery. The input electricity cost is denoted as to empha- The price includes the replacement cost, the rated ca-
size this analogy. Therefore, the operating cost of a battery can pacity, and life cycle which are determined at the time of pur-
be determined in the same form as (1) by deriving price chase and do not need to be updated.
and consumption of the battery. In this paper, lead-acid,
lithium-ion, and vanadium redox batteries are considered. B. Consumption for Batteries

A. Price for Battery The consumption of a battery during discharge is de-

fined as the energy usage for supplying a load during
For a generator, the price of fossil fuel is composed of a unit time:
two components:
(8)
(2)

in which represents the cost for fuel and represents in which is the battery output power and is the power
availability cost. The availability cost includes fuel transporta- loss during discharge.
tion cost and other service costs such as cost for on-site storage The consumption of a battery during charge is defined
facility. Depending on the location of the generator, can be as the energy loss for charging the battery during a unit
much larger than due to transportation and other service time:
costs.
(9)
Similarly, the price for a battery can be determined:
in which is the battery charge power, is
(3) the power loss during charge.
Depending on battery type, and can be character-
where is the price of energy used to charge the battery ized as functions of and , respectively. In this section,
and represents the availability cost of battery capacity. In and are characterized for lead-acid, lithium-ion, and
a microgrid, if renewable energy is used to charge the battery, vanadium redox batteries.
can be zero; therefore, is the main portion of the 1) Lead Acid and Lithium-Ion Battery: The power loss in
price. In this paper, is be defined as the cost to have 1 kWh lead-acid or li-ion batteries is mainly caused by the heat loss
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 NGUYEN AND CROW: STOCHASTIC OPTIMIZATION OF RENEWABLE-BASED MICROGRID OPERATION 2291

during charge or discharge. The heat is generated by ohmic re- TABLE I

sistances of the electrodes and electrolytes, and also by polariza- VRB LOSS MODEL COEFFICIENTS
tion effects [18]. The power loss is proportional to the voltage
drop (polarization) caused by the current:
(10)

For lead-acid and li-ion batteries, the voltage drop can be de-
termined based on the empirical method proposed in [19]:
• During discharge:
From (16), (17), the consumption during charge and
(11) discharge for the VRB can be determined:
• During discharge:
• During charge:

(12)
(18)
where is the internal ohmic resistance, is a constant which
can be calculated from manufacturer's data, and is the rated • During charge:
capacity of the battery.
From (11), (12), the consumption during charge and
discharge for lead acid and lithium-ion batteries can be deter-
mined: (19)
• During discharge:

III. STOCHASTIC UNIT COMMITMENT FOR MICROGRIDS

A. Problem Formulation
In this paper, the stochastic unit commitment is formulated to
(13) minimize the expected operation cost of a microgrid over a time
horizon . The objective function is therefore
• During charge:
(20)

(21)

in which
(14)

where is the state of charge and is the rated voltage of

battery. (22)
2) Vanadium Redox Battery: The power loss of a vanadium
redox battery during charge and discharge includes two compo-
nents: power for pumping the electrolytes and stack loss power
due to internal resistance and electrochemical process. Based on (23)
the empirical approach proposed in [13], the open circuit voltage
and stack voltage can be characterized as functions of and
charge/discharge power:
• Open circuit voltage: (24)
(25)
(15)
where
• During discharge: • is the time horizon, is the time step;
• is the total expected operation cost in period ; is
(16) the total transition cost which accounts for the start-up and
• During charge: shut-down cost of the generators in period .
• and are the number of generators and batteries, re-
(17) spectively;
• and denote generator and battery , respectively;
All model coefficients are given in Table I where and • are respectively the total operation cost
are the rated voltage and rated current of the VRB. during period of the generators, the discharging batteries
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 2292 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 31, NO. 3, MAY 2016

and the charging batteries; is the cost due to power R6 Once a generator is brought online, it should remain on-
mismatch; line for a minimum set time; when a generator is powered
• are respectively the binary status off, it should remain off a minimum time before it can be
during period of the generator and battery ; restarted.
due to the fact that a battery could In a small system such as a microgrid, grid power should not be
not charge and discharge at the same time. used to charge the energy storage because of the relatively low
• is the fuel price for generator ; ESS round-trip efficiency. Therefore, ESS should not charge
is the price for battery ; other ESS, nor should grid power be used for energy storage.
• is the fuel consumption of generator ; Only renewable energy should be used to charge the ESS; this
and are respectively the is reflected in constraint R2.
consumptions during discharging and charging of The above constraints are formulated as follows:
battery ;
• are respectively the dispatched
power during period to generator , discharging battery
and charging battery ; is the power mismatch during
period .
To better define the problem, the following conventions are
introduced:
• Charging power is considered as negative generation.
• Renewable sources (PV and Wind turbines generators) are
not dispatchable and considered as a negative load. The net if is discharging
load at period is defined as
if is charging

if is online
Since are random, is consid-
ered as a random variable. if is offline
• The batteries are charged only when .
• The power mismatch during period is the differ- where is the state of charge of battery in period and
ence between total generation and the net load defined as can be updated:
follows:
— If if discharging

if charging.

(26) By enforcing constraint R1, will happen

of the time when and of the time when
. From (26) and (27), R1 can be rewritten as
with is the total generation.
— If

(27) The parameter governs the level of probability of importing/

exporting electricity to the microgrid from the grid. If ,
then the probability that the microgrid produces sufficient power
with is the total charge. internally is zero and it must import the required power from the
• The microgrid is grid connected. The electricity price and grid. At the other extreme, if , then the probability that
buy back price are considered as deterministic. sufficient power is always produced within the microgrid is 1.
The constraints of the problem are defined based on the energy This is not a realistic condition, therefore is constrained to
management strategies and physical limits of the devices in the be strictly less than 1.0 (but may ideally be quite close to 1.0).
microgrid. The following constraints are considered: Similarly, if , then all net renewable energy will be used
R1 The power mismatch is greater than zero with a prede- for charging the energy storage. For example, if , there
fined probability. is 50% probability that the excess power from the renewable
R2 One battery should not be discharged to charge another energy will be exported to the grid. Choosing a larger will re-
battery; Generators should not be used to charge the bat- duce the probability that renewable energy is used to charge the
teries. energy storage, enabling more renewable energy to be exported
R3 Each storage device cannot be charged (or discharged) to the grid. Therefore, depending on energy management poli-
beyond the maximum (or minimum) SOC. cies, and can be flexibly selected and potentially vary by
R4 The charge (or discharge) rate for each storage device time period (or remain constant).
should not exceed the maximum (or minimum) rates. The amount of imported/exported power is governed by the
R5 Each generator should exceed its minimum output set- choice of the and parameters. Choosing a smaller will
point when online. increase the probability that the system will import power from
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 NGUYEN AND CROW: STOCHASTIC OPTIMIZATION OF RENEWABLE-BASED MICROGRID OPERATION 2293

the grid to supply its load, while choosing a larger will in- where is the CDF of Standard Normal distribution .
crease the probability that the system will export excess renew- Therefore constraint R1 can be realized as
able power to the grid. The selection of and provide a level
of freedom to decide whether or not and how much to import/ex-
port power from/to the grid, depending on desired energy man- if
agement policies.
if
B. Uncertainties in Forecasting Error of Load Demands and The system will generate (or charge) more or less power by
Renewable Sources choosing different values of and . The expected operation
To realize the cost function (21) and constraint R1, the cu- cost in (21) can be expressed:
mulative distribution function (CDF) and mean value of
need to be specified. In practice, the predicted values of load,
PV, and wind generation at time period can be obtained be- (33)
forehand based on forecast. Therefore, the realization of actual
load, PV, wind generation and net load can be expressed as [5], in which
[8], [9]

(28)
(29) (34)

(30)

(31) (35)
where is the price of the electricity which is ex-
where are the forecasting errors which
ported (imported) to the grid.
depend on the forecasting method and forecasting horizon. To
model the uncertainties of load and renewable source forecast, C. Stochastic Dynamic Programming
are considered to be random vari-
ables. Although the wind power forecast error can be more This UC problem can be categorized as a sequential deci-
precisely described with a Weibull, Cauchy [20] or mixed sion-making problem for which dynamic programming (DP)
Normal-Laplace distribution [21], it can be approximated with is well known. Dynamic programming is a method to find the
a zero mean Normal distribution [22]. Furthermore, the net shortest route to the destination by breaking it down to a se-
load error , which is the sum of all errors, can be quence of steps over time; at each step DP finds the possible op-
approximated with a zero mean Normal distribution [22] due to timum sequences (routes) based on the possible optimum sub-
the fact that the load demands and PV generation forecast errors sequences in the previous steps and finally find the optimum
are very close to a Normal distribution [5], [8], [9], [22]. The sequence at last step. The main advantage of DP is it can main-
standard deviation of can be calculated as follows: tain solution feasibility by its ability to find the optimum subse-
quence while searching for the optimum sequence. The primary
drawback to DP is it can be computationally burdensome. For
example, in an -unit system, there are combinations
(32) at each period and for periods the total number of combina-
tions is [3]. For a large scale system the computation
As a result, the following expected values and probabilities can required to traverse this space can be overwhelming. However,
be calculated: in microgrid applications, the small number of units and large
number of constraints significantly decrease the search space,
therefore DP can be an appropriate choice of algorithm [13],
[23].
Due to the uncertainties associated with the stochastic
problem, the cost at each stage is generally a random variable
[24]. Therefore, in the stochastic DP technique, the problem is
formulated to minimize the expected cost. For applying DP, the
states space at stage is defined as follows:
(36)

(37)
where is the set of feasible states in stage ; is the number
of states of ; is the binary status of unit which can be
a generator, a discharging battery or a charging battery. is a
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 2294 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 31, NO. 3, MAY 2016

Fig. 2. Typical microgrid.

Fig. 1. Forward DP algorithm.

TABLE II
DIESEL GENERATOR DATA
valid state if it satisfies constraints R2, R3, R6 and the following
conditions:

if
TABLE III
otherwise. BATTERIES DATA

In this paper, the forward DP algorithm is used (shown in

Fig. 1). The algorithm to compute the minimum cost to arrive
at state in stage is

(38)
where is the minimum cost to arrive to state ;
is the operating cost for state and is the
transition cost from state to state .
The operating cost can be found by performing an
economic dispatch (ED) to minimize the cost function (33) with
constraints R1, R4, and R5. In this paper, a steepest descent
algorithm is used to solve the ED.

IV. CASE STUDY AND RESULTS

A case study is developed to test the proposed approach.
Fig. 2 shows a typical microgrid which is connected to the low
voltage side of a distribution transformer to power residential
loads. The microgrid includes a 50 kW diesel generator, 2 20 Fig. 3. Load and renewable power forecast.
kW wind turbines, a 50 kW PV array, a 10 kW/40 kWh Vana-
dium Redox battery, and an 12 kW/30 kWh AGM lead-acid bat-
tery. The total load is 50 kW at peak. The cost for the AGM bat- The data of generator and batteries, extracted from manu-
tery is estimated at $8000. The replacement cost for the VRB is facturers' data [26]–[28], and their initial states are given in
estimated at $20 000 [25]. Table II and Table III. Day-ahead forecast values of total load
The use of diesel generators in this case study is application demands, PV and wind generations are given in Fig. 3 based on
specific. While most urban grids do not rely on auxiliary gen- per-unit data from [29] and [30].
eration, many rural, developing nation, and military microgrids The standard deviations for load, PV, and wind power fore-
are in areas that rely on auxiliary diesel units. This case study cast errors are respectively 3.12%, 12.5%, and 13.58% [31],
approach would be valid in these applications. [32]. The parameters and are chosen to be 0.9 and 0.1,
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 NGUYEN AND CROW: STOCHASTIC OPTIMIZATION OF RENEWABLE-BASED MICROGRID OPERATION 2295

TABLE IV
NET LOAD FORECAST ERROR STANDARD DEVIATION

Fig. 5. Stochastic UC solution.

• When the load is low and there is no renewable generation,

the stochastic commitment algorithm closely tracks that of
the deterministic case (hours 0 through 8).
Through the choice of and , the amount of allowable risk
in the system can be adjusted. In this example, the values of both
Fig. 4. Deterministic UC solution. and were held constant throughout the 24 hour period, but
in practice, these values may be varied to account for anticipated
variability in load or renewable generation.
respectively. The high value of indicates a high probability Furthermore, note that although not explicitly expressed, the
of meeting all loads internally. The low value of indicates a two energy storage units were committed in accordance with
low probability of exporting renewable energy to the grid (i.e. their individual operating profiles as detailed in Section II to
the preference is to use excess generation to charge the energy maximize their life spans.
storage units). This case study provides the framework for assessing the im-
The standard deviations of the net load forecast error at each pact of the proposed approach. The primary attribute of the pro-
hour are calculated from (32) and given in Table IV. The results posed method is that it better describes the actual performance
for the deterministic UC is shown in Fig. 4. By incorporating of the energy storage system performance and adjusts the allo-
the operation cost functions of the batteries, the ED tends to cation of resources in response. Specifically, most energy man-
dispatch power to the batteries which have a longer cycle life, agement and resource allocation approaches do not explicitly
lower replacement cost, and higher efficiency. In this case, the consider lifecycle degradation due to deep discharge nor do they
VRB has lower , however the AGM battery has higher consider efficiency as a function of output (or input) power. In
efficiency, therefore their dispatched powers, as shown in the re- this case study, the VRB and AGM lead-acid batteries fared sim-
sults, are close. Compared to the diesel generator, the batteries ilarly; even though the VRB was considerably more expensive
have lower operating cost due to lower “fuel” price and higher to install, its better lifecycle attributes and efficiency character-
efficiency. However, the batteries are limited by their maximum istics produced similar long-term economic profiles to the AGM
depth of discharge. For that reason, the batteries can only dis- lead-acid batteries. Furthermore, both batteries were more cost
charge for few hours at night, as observed in the results. effective than the diesel generator. Thus, it makes better eco-
The results for the stochastic UC is shown in Fig. 5 in com- nomic sense to increase the size of the energy storage system
parison with the deterministic case. The impact of the choice of with respect to the diesel generator. Furthermore, note that the
and can be seen by comparing the stochastic and deter- results from the stochastic analysis indicate a heavier reliance
ministic cases. on the diesel generator with respect to the deterministic case,
• When the load is high and renewable generation is low (ap- since with increased uncertainty the diesel generator becomes a
proximately hours 15 through 24), the stochastic algorithm more reliable resource. This can be counteracted by decreasing
over-commits the diesel generator compared to the deter- the size of .
ministic case as indicated by . Since is
chosen to be large (0.9), this indicates that the load must be
V. CONCLUSIONS
met internally with high probability. Since renewable en-
ergy is not available during these hours, this requires that In this paper, a novel battery operating cost model has been
the diesel generator must be available to accommodate any proposed. The model considers cycle life and charge/discharge
potential variability in the load. efficiencies for the batteries. The model makes economic
• When the load is high and the renewable generation is dispatch for multiple batteries possible in the microgrid system
high (approximately hours 9 through 14), the stochastic without introducing additional objective functions to maximize
algorithm over-commits charging the energy storage their efficiencies and their cycle lives. In addition, a proba-
units. Since is chosen to be low (0.1), this indicates a bilistic constrained approach has been proposed to consider
low probability of sending excess generation to the grid, the uncertainties in load and renewable power forecast errors.
thereby increasing the likelihood of charging the batteries. Stochastic dynamic programming is applied in this method
Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.
 2296 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 31, NO. 3, MAY 2016

to find the optimal day-ahead scheduling for a typical micro- [20] B. Hodge and M. Milligan, “Wind power forecasting error distributions
grid with a diesel generator, PV arrays, wind turbines, VRB, over multiple timescales,” in Proc. 2011 IEEE Power and Energy Soc.
General Meeting, 2011, pp. 1–8.
and AGM batteries. Results show the proposed approach can [21] J. Wu, B. Zhang, H. Li, Z. Li, Y. Chen, and X. Miao, “Statistical dis-
maintain the system optimal operation with a high probability tribution for wind power forecast error and its application to determine
without investigating a vast number of scenarios. Future work optimal size of energy storage system,” Int. J. Elect. Power Energy
in this area will include the consideration of electricity prices Syst., vol. 55, pp. 100–107, 2014.
[22] R. Doherty and M. O'Malley, “A new approach to quantify reserve de-
in selecting and to maximize the microgrid's profit. mand in systems with significant installed wind capacity,” IEEE Trans.
Power Syst., vol. 20, no. 2, pp. 587–595, May 2005.
REFERENCES [23] L. M. Costa and G. Kariniotakis, “A stochastic dynamic programming
[1] , N. Hatziargyriou, Ed., Microgrids: Architectures and Control. New model for optimal use of local energy resources in a market environ-
York, NY, USA: Wiley, 2014. ment,” in Proc. 2007 IEEE Lausanne Power Tech, 2007, pp. 449–454.
[2] H. Liang and W. Zhuang, “Stochastic modeling and optimization in a [24] D. P. Bertsekas, Dynamic Programming: Deterministic and Stochastic
microgrid: A survey,” Energies, vol. 7, no. 4, pp. 2027–2050, 2014. Models, D. P. Bertsekas, Ed. Englewood Cliffs, NJ, USA: Prentice-
[3] A. J. Wood and B. F. Wollenberg, Power Generation, Operation, Con- Hall, 1987.
trol. New York, NY, USA: Wiley, 1996, vol. 2. [25] G. Kear, A. A. Shah, and F. C. Walsh, “Development of the all-vana-
[4] W. Su, J. Wang, and J. Roh, “Stochastic energy scheduling in mi- dium redox flow battery for energy storage: A review of technological,
crogrids with intermittent renewable energy resources,” IEEE Trans. financial and policy aspects,” Int. J. Energy Res., vol. 36, no. 11, pp.
Smart Grid, vol. 5, no. 4, pp. 1876–1883, Jul. 2014. 1105–1120, 2012.
[5] X. Peng and P. Jirutitijaroen, “A probabilistic unit commitment [26] MQ Power [Online]. Available: http://www.powertechengines.com/
problem with photovoltaic generation system,” in Proc. TENCON MQP-DataSheets/MQP050IZ_Rev_0.pdf
2009—2009 IEEE Region 10 Conf., Jan. 2009, pp. 1–6. [27] Prudent Energy [Online]. Available: http://www.pdenergy.com/pdfs/
[6] M. Lubin, C. G. Petra, M. Anitescu, and V. Zavala, “Scalable sto- Prudent_Energy_Product_Brochure_2011.pdf
chastic optimization of complex energy systems,” in Proc. 2011 Int. [28] Sunxtender [Online]. Available: http://www.sunxtender.com/solarbat-
Conf. IEEE High Performance Computing, Networking, Storage and tery.php?id=11
Analysis (SC), 2011, pp. 1–10. [29] D. Hung, N. Mithulananthan, and K. Lee, “Determining PV penetration
[7] G. Martinez, N. Gatsis, and G. B. Giannakis, “Stochastic program- for distribution systems with time-varying load models,” IEEE Trans.
ming for energy planning in microgrids with renewables,” in Proc. Power Syst., to be published.
2013 IEEE 5th Int. Workshop Computational Advances in Multi-Sensor [30] A. A. Moghaddam, A. Seifi, and T. Niknam, “Multi-operation man-
Adaptive Processing (CAMSAP), 2013, pp. 472–475. agement of a typical micro-grids using particle swarm optimization: A
[8] S. Mohammadi, S. Soleymani, and B. Mozafari, “Scenario-based comparative study,” Renew. Sustain. Energy Rev., vol. 16, no. 2, pp.
stochastic operation management of microgrid including wind, pho- 1268–1281, 2012.
tovoltaic, micro-turbine, fuel cell and energy storage devices,” Int. J. [31] N. Lu, R. Diao, R. Hafen, N. Samaan, and Y. Makarov, “A comparison
Elect. Power Energy Syst., vol. 54, no. 0, pp. 525–535, 2014. of forecast error generators for modeling wind and load uncertainty,”
[9] T. Niknam, R. Azizipanah-Abarghooee, and M. R. Narimani, “An effi- in Proc. 2013 IEEE Power and Energy Soc. General Meeting (PES),
cient scenario-based stochastic programming framework for multi-ob- 2013, pp. 1–5.
jective optimal micro-grid operation,” Appl. Energy, vol. 99, no. 0, pp. [32] F. Kreikebaum, R. Moghe, A. Prasai, and D. Divan, “Evaluating the
455–470, 2012. application of energy storage and day-ahead solar forecasting to firm
[10] M. Mazidi, A. Zakariazadeh, S. Jadid, and P. Siano, “Integrated sched- the output of a photovoltaic plant,” in Proc. 2011 IEEE Energy Con-
uling of renewable generation and demand response programs in a mi- version Congr. Expo. (ECCE), 2011, pp. 3556–3561.
crogrid,” Energy Convers. Manage., vol. 86, no. 0, pp. 1118–1127, [33] A. Zakariazadeh and S. Jadid, “Energy and reserve scheduling of mi-
2014. crogrid using multi-objective optimization,” in Proc. 22nd Int. Conf.
[11] H. Bilil, G. Aniba, and M. Maaroufi, “Probabilistic economic emission Exhib. Electricity Distribution (CIRED 2013), 2013, pp. 1–4.
dispatch optimization of multi-sources power system,” Energy Pro-
cedia, vol. 50, no. 0, pp. 789–796, 2014.
[12] A. A. ElDesouky, “Security and stochastic economic dispatch of power Tu A. Nguyen received the B.S. degree in Power Sys-
system including wind and solar resources with environmental consid- tems from Hanoi University of Science and Tech-
eration,” Int. J. Renew. Energy Res. (IJRER), vol. 3, no. 4, 2013. nology, Hanoi, Vietnam, in 2007 and the Ph.D. de-
[13] T. Nguyen, X. Qiu, J. Guggenberger, II, M. Crow, and A. Elmore, gree from Missouri University of Science and Tech-
“Performance characterization for photovoltaic-vanadium redox bat- nology, Rolla, MO, USA, in 2015.
tery microgrid systems,” IEEE Trans. Sustain. Energy, vol. 5, no. 4, He worked as a Power Transformer Test Engineer
pp. 1379–1388, Oct. 2014. in ABB's High Voltage Test Department in Vietnam
[14] Y. Tan, Y. Cao, C. Li, Y. Li, L. Yu, Z. Zhang, and S. Tang, “Microgrid from 2008 to 2009. His research interests include
stochastic economic load dispatch based on two-point estimate method microgrid system modeling/analysis and power
and improved particle swarm optimization,” Int. Trans. Elect. Energy electronics applications in microgrid systems.
Syst., 2014.
[15] J. M. Lujano-Rojas, G. J. Osorio, and J. P. Catalao, “A probabilistic
approach to solve economic dispatch problem in systems with inter-
mittent power sources,” in Proc. 2014 IEEE PES T&D Conf. Expo.,
2014, pp. 1–5. M. L. Crow (S'83–M'90–SM'94–F'10) received the
[16] Y. Zhang, N. Gatsis, and G. B. Giannakis, “Risk-constrained energy B.S.E. degree from the University of Michigan, Ann
management with multiple wind farms,” in Proc. 2013 IEEE PES In- Arbor, MI, USA, and the Ph.D. degree from the Uni-
novative Smart Grid Technologies (ISGT), 2013, pp. 1–6. versity of Illinois, Urbana/Champaign, IL, USA.
[17] T. Markel, K. Smith, and A. Pesaran, “Improving petroleum displace- She is the F. Finley Distinguished Professor of
ment potential of phevs using enhanced charging scenarios,” in EVS-24 Electrical Engineering at the Missouri University
NREL/CP-540-45730, 2009. of Science & Technology, Rolla, MO, USA. Her
[18] H. A. Kiehne, Battery Technology Handbook. Boca Raton, FL, USA: research interests include computational methods for
CRC Press, 2003, vol. 60. dynamic security assessment and the application of
[19] O. Tremblay and L.-A. Dessaint, “Experimental validation of a battery power electronics in bulk power systems.
dynamic model for ev applications,” World Elect. Vehicle J., vol. 3, no. Prof. Crow is a Registered Professional Engineer
1, pp. 1–10, 2009. in the State of Missouri.

Authorized licensed use limited to: SHIBAURA INSTITUTE OF TECHNOLOGY. Downloaded on February 28,2024 at 02:07:23 UTC from IEEE Xplore. Restrictions apply.