This paper presents the basis for the development of an intelligent and autonomous energy management strategy for hybrid powersystem (HPS). Two hierarchical levels are proposed to control and manage the HPS. The low level is performed by a local controlunit (DC-DC converters controller) of the different power sources. Dynamic equations describing the coupling of converters arederived, and a robust sliding mode dynamic controller is designed. The high level is performed by the online supervisor unit. Thisunit is designed by applying on-line Takagi-Sugeno fuzzy logic principles. As a result the robust control system gets rid of the limitsof the HPS, which has the imprecision, uncertainty, strong coupling, and nonlinearity, to achieve its tractability, robustness, andlow solution cost. Under the operation constraints related to each type of sources, the simulation results show that the optimaloperation objective of HPS has been achieved.
1. Introduction
The limited reserves of fuel oils, their pollution impact,and their unstable prices have significantly increased theinterest in renewable energy sources (RES: photovoltaicmodules, wind turbine, etc.) to produce electricity that is anessential factor for the development of the human societies.In this context, HPSs, which combine renewable energy andconventional energy sources with the storage systems, arevery interesting in terms of environmental protection andreduction of the effects of greenhouse gas emissions. In otherwords, the HPSs can provide an economic, environmentfriendly, and reliable supply of electricity.
To perform the hybrid systems a control strategy hasto be designed and implemented on the system. Numeroussolution methodologies have been proposed in this fieldin the last decade [1–4]. Many of them are based ona static approach. However, the development of robustcontroller is necessary to ensure stability and robustness ofthe multisources of the renewable energy systems.
Subsequently, others advanced control strategies, such aspredictive control, fuzzy logic, and neural network, have beendeveloped and successfully applied [5–7]. However, theseapproaches require considerable computing resources and
as a result their applicability for real-time applications isreduced.
In this paper, emphases are put on the energy man-agement and control from the viewpoint of control theory.The proposed management strategy and controller, designedby applying fuzzy logic and sliding mode principles, notonly is simple, stable, and robust, but also reduces thecomputational resources. Two hierarchical levels are requiredto control and manage the HPS. The low level is performedby a local control unit (DC-DC converters controller) of thedifferent power sources. This level manipulates the duty cycleconverter according to the variation of the sources and theload. Without the use of cascade structure commonly used inliteratures, I proposed a robust control law based on slidingmode control approach to operate the DC-DC converters inthe most efficient way.
The high level is performed by the online supervisor unit.This unit uses the data about the load, the meteorologicalconditions and the charge state of energy storage system(ESS) and combustion engine (CE), to correctly and effi-ciently share the load demand according to the availability ofconventional and renewable energy, in other words, to decidewhether to charge or discharge the ESS, to turn on or offthe CE, to reduce the renewable sources power production
2 ISRN Renewable Energy
Study zone
DC/DC converter
Inverter
Wind turbine
Energy storage
DCbus
Rectifier
Reversible DC/DC
converter
DC/DC converter
Rectifier
Variable load
DC/DC converter
Combustion engine
∼=
=
= =
=
=
−
−
−
PV panel
systems (B + U)
ϕi: duty cycle
ϕ1 ϕ2
ϕ3ϕ4
Vbus
Figure 1: HPS Scheme.
or not, and so forth. Such control level could be achieved byadaptive techniques control such as those employing fuzzylogic.
The paper is organized in five sections. In Section 2, thehybrid power system and their mathematical models aredescribed. Then, in Section 3, I will present the design oflow level controller witch performed by a local sliding modecontrol unit. By applying online dynamic Takagi-Sugenofuzzy logic principles, the high-level controller design ispresented in Section 4. An application to the multisourcesystem is presented in Section 5 in order to illustrate theefficiency of the developed approach. The conclusion sumsup the main characteristic of the proposed hierarchicalcontroller and introduces one of the most challenge foradaptive control in the future research.
2. System Configuration
The HPS under consideration is shown in Figure 1. It con-sists of a conventional generation (CE), a renewable energysources (RES: photovoltaic module and wind turbine), anESS (batteries and ultracapacitors), and a variable load. Allthese elements are connected onto a DC bus through DC/DCpower electronic converters.
In order to allow ESS charging as well as discharging, theDC/DC converter which connects the ESS to the DC bus isbidirectional unlike what other converters are unidirectional.The DC bus accumulates the generated energy and sends itto the variable load and, if necessary, to the energy storagesystem. In this configuration, renewable sources take over asmain energy source.
2.1. Models of Energy Sources
2.1.1. Renewable Sources. The most accessible energy sourcesare solar and wind energy which can be produced through
photovoltaic conversion and wind turbine, respectively.The available photovoltaic (PV) and wind turbine energyproduction are calculated based on the meteorologicalinformation (irradiation, temperature, and wind speed) andtheir characteristics. In this paper, the photovoltaic and windturbine power models were obtained, respectively, from [8,9] as follows.
For PV area, the electric model closest to the PVgenerator is a model with one diode.
The current output of a PV array ((Np − Ns) PV cellsinterconnected in a parallel series) Ip can be calculated by
Ip = Np
[Iph − Is ∗
(exp
(NpVp + NsIpRs
NpNsVT
)− 1
)]
− NpVp + NsIpRs
NsRsh,
(1)
where VT = (nKBT)/q is the thermodynamic voltage, Vp isthe voltage level on the PV panel array terminals, Iph is thephotocurrent, and Is the reverse saturation current, KB is theBoltzman constant, and n is the cell deviation from the idealp-n junction characteristic.
Based on the observation of voltage-power relationship,I can notice that the PV array has a unique voltage pointwhich provides the maximum power [10]. This maximumpower point (MPP) can be tracked in real time with acontrol algorithm called MPPT in order to increase theirperformances. Several algorithms have been proposed inexisting literature [11] such as: perturb and observe, incre-mental conductance, fractional open circuit voltage, fuzzycontrol, and neural network. As the objective of this paperis to develop the higher-level control algorithms (energymanagement strategy), I adopt in the simulation section oneof these algorithm (e.g., incremental conductance).
For the wind turbine, the output power of wind genera-tor at wind speed v is expressed as
PWT(t)
=
⎧⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎩
(v3 − v3
ci
)v3r − v3
ciPR vci < v < vr ,
PR vr < v < vco,
0 v ≤ vci or v ≥ vco,
v(t) = vr(t) ·(h
hr
)γ,
(2)
where PR is the nominal power; v, vr , vci, and vco, respectively,are the wind speed at desired height h, wind speed at thereference height hr , and cut-in and cutoff wind speed of thewind generator.
2.1.2. Combustion Engine. In the case of a renewable energysystem, the production of electrical energy is conductedwithin the resources (wind, sun, . . .) and not in demand.For standalone installations, in order to assist the renewablesources in situations of deficient environmental circum-stances, it is necessary to resort to storage and/or to add aCE.
ISRN Renewable Energy 3
The CE is a conventional power source which isindependent from environmental conditions (irradiation,wind velocity, etc.). This source is the site of a complexphenomena and strongly nonlinear. There are several levels,depending on the desired use and degree of complexity ofmodels. In this work, for simplification reasons of systemstudy, I assumed that the dynamic behavior of CE can beapproximated by a first-order system with a time constantτ [12].
2.1.3. Energy Storage Systems. Energy storage is a key factorin a HPS. It can provide energy when needed and store thesurplus of production otherwise. Since there is a wide rangeof storage options, for a given application, both the powerand energy rating of the storage device must be considered inthe device selection. The most common storage technologiescan be summarized in terms of transient and steady-statestorage devices. In our case, the energy storage systemcomprises the batteries and the supercapacitors banks. Thecombination of these sources makes the best use of theadvantages of each individual device and may meet therequirements of the load regarding both power and energydensities.
(1) Batteries. In general, the batteries allow the good storagecharacteristics for wide ranges of applications such as electricvehicles, hybrid electric vehicles, and electrical networks[13–15]. The mathematical dynamic model of the batterybank used in the simulation program is deducted fromthe CIEMAT model [16, 17]. Its input is the currentbattery profile which is positive when power is drawn fromthe batteries and negative when power is supplied to thebatteries. Its output is the state of charge (SOC), whichdefines the relationship between the battery’s total energyand its usable energy (3). We have
SOCmin ≤ SOC(t) = SOC(t − 1) +Ibat
Cbat≤ SOCmax, (3)
where Ibat is the charge (or discharge) instantaneous currentof the battery bank; this current is positive when the batteriesare charging and negative otherwise. Cbat is the capacityof battery bank calculated from the average battery currentIb m as follow:
Cbat = 1.67∗ C10
1 + 0.67∗ (Ib m/I10)0.9 ∗ (1 + 0.005∗ ΔT), (4)
where ΔT is a battery bank heating relative to the ambienttemperature, C10 is a nominal capacity of battery bank, andI10 is a nominal current of battery bank.
According to the specifications from the manufacturers,in this paper, the battery’s lifetime can be prolonged to themaximum if SOC is between 20 and 90% [17].
(2) Supercapacitors. Supercapacitors are electrical storagedevices with a high power density, used for peak powerleveling to assist the battery during transients.
ESS
Battery regime
ESS help mode
ESS recharge mode
Combustion engine
Max mode Limit mode
Production mode
Off mode
Load DC bus Ultracapacitor regime
Slow down mode
Renewable sources
Figure 2: Multisource operating modes.
In contrast to a battery, the voltage of a supercapacitorchanges linearly with its state of charge as follows [18]:
SOCUC = 43
(CV 2
2Emax− 1
4
), (5)
where C is the equivalent total capacity of supercapacitorsbank, V is the instantaneous voltage measured at itsterminals, and Emax is the maximum stored energy obtainedfor Vmax.
In this paper, to ensure good performances and a suffi-cient lifetime of supercapacitors and to avoid overchargingthe maximum voltage was limited to Vmax, and the dischargerate is fixed at 0.25.
When all these constraints due to limitations in voltage,current and power are combined, the momentary poweroutput of the supercapacitors can be calculated for every timestep.
2.2. Operating Modes of Energy Sources. Because of the vari-able structure nature of multisource system, the operatingmodes of each source are summarized in Figure 2.
For renewable sources, there are two possible operatingmodes: Maxi Mode and Limit Mode. The Maxi Mode occurswhen the maximum energy captured from the renewablesources is not enough to satisfy the total power demand (theload plus the ESS recharge requirements). In this case, theMPPT algorithm must be used (i.e., capturing the maximumrenewable power (Pren max) available). However, when theprovided power is higher than demanded one (Limit Mode),only the required amount of energy is produced.
For CE, there are three possible operating modes:Production Mode (PM), Off Mode (OM), and Slow DownMode (SDM).
In PM, the CE assists the renewable sources in order toprovide the load power demand and the ESS power recharge.
In OM, the load power demand is provided withrenewable sources assisted by ESS, and the CE is off.
In SDM, the load power demand is only provided withrenewable sources, and the CE is relaxed to avoid transientphase and overheating. This is important to avoid runningthe CE for very short period of time.
4 ISRN Renewable Energy
En
ergy
(Batteries)in transient state(Ultracapacitors)
Energy production Energy consumption
(Load )
Energy storagesystems
En
ergy
stor
age
hel
p
(CE + RS )
Charge/discharge Charge/dischargein steady state
−−
Figure 3: Energy storage management strategy.
VL
Ce
Re Rch
CDC
s
M
VgVp
iL
Vs
VB
VA
n
Figure 4: Structural diagram of the ZVS full bridge isolated DC/DCconverter.
For ESS, depending on the flow of power, there are twopossible operating modes, Help Mode (HM) and RechargeMode (RM).
In the HM, the power demand is provided by therenewable sources assisted by ESS while the CE is switched offand renewable sources power is lower then demanded one.
In the RM, when the load energy demand is smaller thana produced power, the redundant energy is used to charge theESS.
According to profile of the redundant energy (Figure 3),I can distinguish two charge/discharge cycles: steady-statecycle ensured by batteries and transient-state cycle ensuredby ultracapacitors, which are used for peak power leveling toassist the batteries during hard power demand.
In order to improve system operation, the power andenergy management strategy has to decide which operatingmode must be used.
3. Low-Level Controller
The DC/DC converters are basic constituents of the mul-tisource system. The aim of these converters is to regulatethe DC component of the output voltage to its reference bycontrolling the current provided by each source in spite ofthe voltage variations on their inputs.
3.1. Multimodel of Several DC/DC Power Converters Coupledon a DC Bus. An accurate model for the coupling of DC/DCpower converters on a DC bus has been developed previouslyin [19].
The ZVS full-bridge isolated DC/DC converter [20, 21],studied in this paper, is represented on Figure 4.
CDCRch
L
iL
d3 d5 d6oror
d8 d7ord4 ic
Ce
Re
s
rp LMiM
i
vdvd
vd
rmos(Q1)
rmos(Q2)
or
Figure 5: Phases 1 and 3.
ig i
Transformer
Vg
rp
LM e2e1
rsiM
L
rL
Coil
d8 d7or
vd
vd
d5 d6or
iL
ic ir
Ce
s
CDCRchRe
L
i
(to load)
DC bus
rmos(Q1) or rmos(Q3)
rmos(Q2) or rmos(Q4)
Figure 6: Phases 2 and 4.
It supposed that it runs in a continuous conductionmode.
Four basic structures of operation may be distinguished.
Phase 1: (Q1Q4) On and (Q2Q3) Off.
Phase 2: (Q1d3) On.
Phase 3: (Q2Q3) On and (Q1Q4) Off.
Phase 4: (Q2d4) On.
The structural diagram corresponding to the operationphases 1 and 3 is depicted in Figure 5 and the one corre-sponding to the operation phases 2 and 4 is representedon Figure 6. rp and rs are, respectively, the transformerprimary and secondary resistances. LM is the transformermagnetizing inductance. CDC is the DC bus capacity, and theload Rch is supposed resistive.
The voltage source (Vg) is disconnected during phases 2and 4 which correspond to the transformer demagnetization(Figure 4), so ig = 0.
Each phase leads to the following state space model:
x = Aix + Biu, (6)
where x = [iM iL Vbus]T is the state vector, and u =
[Vg Vd]T is the control vector.iM , iL, Vbus, vg , and vd, respectively, represent the
transformer magnetizing current, the inductance current,the DC/DC power converters output voltage, the source, andthe diode voltage. It is assumed that diodes are not ideal. Wehave
ISRN Renewable Energy 5
Converter (1) Converter (2)Bus
Converter (m− 1)
Vgm−1
rL
Re
ReRe
Re
Ce
CeCe
Ce
L1 rLrL L2
rL
iLm
Converter (m)
Vgm
iL1 iL2
iLm−1
Lm−1
Vg 2Vg 1
Cdc
Rch1
LM
−1
Figure 7: Structural diagram of the coupling of m DC/DC powerconverters.
A1 =⎡⎢⎢⎢⎢⎢⎢⎢⎢⎣
−(
2Rmos + rp)
LM
−n(
2Rmos + rp)
LM0
−n(
2Rmos + rp)
L
−n2(
2Rmos + rp)− (rs + rL)
L
−1L
01Ceq
− 1ReqCeq
⎤⎥⎥⎥⎥⎥⎥⎥⎥⎦
,
A3 =⎡⎢⎢⎢⎢⎢⎢⎢⎢⎣
−(
2Rmos + rp)
LM
+n(
2Rmos + rp)
LM0
−n(
2Rmos + rp)
L
−n2(
2Rmos + rp)− (rs + rL)
L
−1L
01Ceq
− 1ReqCeq
⎤⎥⎥⎥⎥⎥⎥⎥⎥⎦
,
A2/4 =
⎡⎢⎢⎢⎢⎢⎢⎢⎣
−(Rmos + rp
)LM
0 0
0−rLL
−1L
01Ceq
− 1ReqCeq
⎤⎥⎥⎥⎥⎥⎥⎥⎦
,
B1 =⎡⎢⎣
1LM0
n
L−2L
00
⎤⎥⎦T
,
B3 =⎡⎢⎣ −
1LM0
n
L−2L
00
⎤⎥⎦T
,
B2/4 =[
00
00
00
]T
.
(7)
Combining the above phase states model, the averagemodel of the DC/DC converter shown in Figure 4 can bederived as follows [22, 23]:
x(t) = f (x,u) + g(x,u)ϕ(t), (8)
where
f (x,u) = ax(t) + bu(t), (9)
g(x,u) = Ax(t) + Bu(t), (10)
A = 12
(A1 − A2 + A3 − A4), (11)
a = 12
(A2 + A4), (12)
B = 12
(B1 − B2 + B3 − B4), (13)
b = 12
(B2 + B4). (14)
The average model of the DC/DC converter developedabove can be extended to the coupling of many renewablesources [24]. I suppose that the DC/DC power converters areidentical.
The partial structural diagram of the coupling of mDC/DC power converters is represented in Figure 7.
The multimodel for m DC/DC converters coupled on aDC bus is express as follows:
X(t) = F(X ,u) + G(X ,u)ϕ(t), (15)
where
F(X ,U) = [ f1(x1,u1), . . . , fm(xm,um)]T ,
G(X ,U) = diag[g1(x1,u1), . . . , gm(xm,um)
],
X = [x1, . . . , xm]T ,
U = [u1, . . . ,um]T ,
ϕ = [ϕ1, . . . ,ϕm]T.
(16)
The multimodel parameters (F and G) change accordingto the DC/DC power converters coupled on the DC bus.
For simplicity of presentation, in this subsection, onlytwo coupling sources are considered (m = 2).
3.2. Dynamic Controller Design for DC/DC Converter.The variable structure systems (VSS) theory [25] can beextremely helpful in the study of the control of powersDC/DC converters. The switched mode DC/DC convertersare nonlinear, and it is not suitable to the application of linearcontrol theory.
A different approach, which complies with the nonlinearnature of switch-mode power supplies, is proposed. In thiscontext, the sliding mode control, which is derived fromthe VSS theory, appears as a powerful control techniquethat offers several advantages: stability even for large supply
6 ISRN Renewable Energy
Converter 1DC bus
Converter 2
Sliding mode controller
CeCe ReRe
Rch
Vbus
ϕ1 ϕ2
rLiL1
Cdc
L1 rLiL2
Figure 8: MIMO sliding mode robust control scheme.
and load variations, robustness, good dynamic response andsimple implementation. Their capabilities emerge especiallyin application to high-order converters, yielding improvedperformances as compared to usual control techniques.
In this subsection, I design a sliding mode dynamiccontroller “SMDC” which regulates the voltage level on theDC bus by controlling the current provided by each source.
In various works in the literature, a cascade controlstructure with two control loops for DC/DC converterscontrol is usually adopted [26, 27]. In this structure, aninner current loop regulates the inductor current, whereasan external control loop keeps a constant output voltage.
However, this structure has some drawbacks such as thevalue of reference current that may be poorly estimatedbecause of quick variations of the load and the sources.The current estimation error introduces a bus voltage errorcompared to the reference one. This affects the controllerperformances in terms of robustness.
To overcome this issue, I adopt, in this paper, a MIMOsliding mode robust control scheme (Figure 8).
The development of the sliding mode control schemeconsists of two phases. The first is to design a sliding surfacewhere the DC/DC converter exhibits desired properties. Thesecond is to design a control law to drive and maintain thesystem on the sliding surface [26, 27].
In order to act simultaneously on all state variables, let usdefine the following PI-type sliding surface:
ILi ref,Vbus ref are the references of inductor current andvoltage DC bus. k1,2,3 are the tracking weight factor error, andα > 0 is the sliding-surface integral parameter.
In the following, a bounded control input is designedto force δi(t) to converge to zero or make its absolute valuesmaller.
We obtain the equivalent control ϕeq1,2 by applying theinvariance condition:
δi(t) = δi(t) = 0, with ϕi(t) = ϕeqi(t), i = 1, 2. (19)
From (17), the time derivative of δi(t) along system (15)is given as
The notation A2− refers to the second column of the Amatrix.
Thus, the equivalent control input is given as
ϕeqi = − Ri
Hi, i = 1, 2, (21)
where
Ri = ki(a2−xi + b2−ui − ILi ref
)− k3
(A3−xi + B3−ui − Sref
)+ αei,
(22)
Hi = ki(A2−xi + B2−ui). (23)
It is well known that the exact value of the externaldisturbance and the parameter variations of the system, suchas internal resistance and magnetizing current, are difficultto measure in advance for practical applications. Therefore,I propose a nonlinear switching control-input ϕsi to estimatethe upper bound of uncertainties and external disturbance:
ϕsi(t) = − γi(t)Hi(t)
sign(δi(t)). (24)
In the presence of parameter uncertainties and externaldisturbances the dynamic equation (15) becomes
X(t) = F(x) + G(x)ϕ + Δ(t), (25)
where Δ includes the uncertainties and perturbations.In order to reduce chattering, the most common method
is to replace the sign function by the saturation functionsat(δ(t), ζ), where
sat(δ(t), ζ) =⎧⎪⎨⎪⎩
sign(δ), |δ| > ζ > 0,δ
ζ, |δ| ≤ ζ
(26)
and ζ is a small positive constant.Hence, the SMDC rule can be designed as
ϕi(t) = ϕeqi(t)− γi(t)Hi(t)
sat(δi(t), ζ). (27)
The main theorem, stated and proved below, providessufficient conditions to ensure the stability and robustnessof the multisources system.
Theorem 1. Let the nonlinear multisources system given by(15) and sliding surface given by (17), and suppose that theLyapunov function is defined by
V(t) = 12δ1(t)2 +
12δ2(t)2. (28)
ISRN Renewable Energy 7
Table 1: NMSE.
PI SMDC
1.2483 0.0075
Then, the sufficient stability condition for the SMDC inthe sense of Lyapunov should satisfy the following range ofswitching gains:
γi(t) ≥ |Δ(t)|, i = 1, 2, (29)
where |Δ| represent the upper bound of the uncertainties andperturbations.
Proof. If V(t) is a Lyapunov function candidate defined asin (28), the asymptotic stability will be satisfied if V(t) ≤ 0,with
V(t) = δ1(t)δ1(t) + δ2(t)δ2(t). (30)
Using (17), (25), and (27) in (30) I obtain
V(t) = δ1(t)[−γ1(t) sat(δ1(t), ζ) + Δ(t)
]+ δ2(t)
[−γ2(t) sat(δ2(t), ζ) + Δ(t)] ≤ 0.
(31)
As a conclusion, to satisfy the condition V(t) ≤ 0, andto compensate the bounded parametric uncertainties anddisturbances, I must restrict to sufficient condition (29).
3.3. Evaluation of Low Level Control Design. To illustratethe low level control design and performance evaluationof my structure, a hybrid power system coupled to twosources through a similar DC/DC converter is considered.The characteristics of the full bridge converters are
Rmos = 0.005Ω, Rl = 2Ω, n = 12,
Vd = 0.3 v, L = 1e − 3 H,
Lm = 20e − 6 H, rp = 0.05Ω,
rs = 0.05Ω, Ce = 20e − 6 F,
Re = 56e3Ω, CDC = 10e − 6 F.
(32)
The DC bus voltage reference is set at 100 V. Simulationsare obtained with sampling interval Te = 50μs.
We suppose that 40% of the load demands are suppliedwith the first source and 60% with the second one.
Responses obtained with my algorithm are comparedwith ones resulting from the PI algorithm with cascadestructure. The linearization techniques are applied to theconverters in order to deduce linear parameters [28].
The simulation conditions are identical for both con-trollers.
To compare the global performances of my algorithmwith PI one, let us consider the normalised mean square error
(NMSE, Table 1), that is interpreted as the overall deviationsbetween output plant and desired values and is defined as:
NMSE =∑
l (Vbus −Vbus ref)2∑
l (Vbus ref)2 . (33)
In order to test the capacity of SMDC to reject dis-turbances, and to investigate the robustness of the controlscheme, the load and sources dynamics of the hybridsystem are supposed fast. The sources and load profiles arerepresented by Figures 9 and 10.
The closed-loop responses of load current DC busvoltage, tracking voltage error, and load current obtainingwith PI controller and sliding mode controller are shown inFigures 11, 12, and 13, respectively.
Given the fast dynamics imposed on the sources and theload as evident from Figures 9 and 10, the simulation resultsreveal that the SMDC is advantageous in providing negligiblesteady-state errors (NMSE = 0.0075), to adjust the flowsource that meets the load demand and guaranteed stablesystem (Figure 12). It can also be seen that the responsesare satisfactory in terms of overshoot, settling time, and falltime. This shows an excellent behavior of the sliding modecontroller in comparison with PI controllers which does notrun correctly for wide variations of operating point.
4. High Level Controller: IntelligentPower and Energy Management Strategy
Generally, an HPS is a dynamic system of electric powerproduction. While during their operation, fluctuation inwind speed, solar radiation, and load demand will occur,to ensure a reliable power supply under all operatingconditions, the global control strategy is critical to controlefficiently the energy flow in the multisource system.
Indeed, under the operation constraints related to eachtype of sources, the energy flow must be managed suchthat the load demand is satisfied for all time (preserve thereliability of the system), the ESS is managed in an optimalway (the battery and ultracapacitor SOC minimum andmaximum threshold values which must be respected, and thecurrent and voltage should be limited to maximum absolutevalue in order to increase the ESS durability), and the fuel useis optimal.
These requirements are achieved by improving the rate ofrenewable energy penetration, by an appropriate controllerof CE and finally, and by a specific management of thecharge-discharge ESS operation.
Since the desired behavior of the multisource system iswell known and can be described using linguistic variables,the use of an online Takagi-Sugeno fuzzy dynamic supervisorseems appropriate. Indeed, the fuzzy approach has manyadvantages, such as the possibility to use multiple inputvariables without increasing the complexity of the controller,and the desired behavior can be described by simplifieslinguistic variables but it would be difficult to expressmathematically and finally, historical data is stored only fora short time window, which perfectly meets the real-timeapplication requirements.
Figure 11: DC bus voltage plot obtained with PI controller andsliding mode controller.
4.1. FLC’s Structure. Based on reviews of multisource power,the block diagram of the global control strategy is illustratedin Figure 14.
The required behavior was implemented with the help ofsix input variables (the energy sources availability and loadpower demand). These information are useful to determinewhether there is surplus or deficient energy to supply the loaddemand.
The load power estimation (P load) and there variation(ΔPload) are used to evaluate instantaneously the loadrequirements and to decide which ESS operating mode mustbe used (battery with high-specific energy or ultracapacitorwith high-specific power).
The SOCs of ESS are used to prevent an excessivedischarge or an overcharge of the ESS and to know howmuch supply energy is available. It is estimated from theSOC model calculated depending on ESS power (batteriesand ultracapacitors) that is also a fuzzy output.
The online dynamic fuzzy logic controller output vari-ables are a reference power vector (power set points: batterypower (Pbat), ultracapicitor power (Puc), combustion engineproduced power and Pren) sent to the low level control, andCE required power sent to the CE controller.
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
Time (s)
Sliding mode
PI
−2−1
Vol
tage
tra
ckin
ger
ror
(V)
Figure 12: Tracking voltage error plot (logarithmic scale) obtainedwith PI controller and sliding mode controller.
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2040
60
Loa
d cu
rren
t (A
)
0.1 0.102 0.104 0.106 0.108 0.110123456
Time (s)
Loa
d cu
rren
t (A
)
PI
Sliding mode
Reference current
Zoom
−40−20
Figure 13: Load current plot obtained with PI controller andsliding mode controller.
4.2. Membership Functions. Trapezoidal type membershipfunctions were used for the linguistic input variables, asshown in Figures 15(a) to 15(e), because they producesmoother control action due to the flatness at the top of thetrapezoidal shape [5].
The battery state of charge (SOCbat) input was dividedusing the linguistic variables: too low (TL), low (L), normal(N) and high (H), as shown on Figure 15(a). The SOCwindow defines the relationship between the battery’s totalenergy and its usable energy. Fuzzy sets (TL) and (H)represent the ranges where the SOC should not be. Fuzzy set(N) represents the range where the SOC should be, and (L)acts as an energy reserves level necessary to satisfy the loadpower requirement, when the CE is on start-up phase.
The membership function for ultracapacitor state ofcharge (SOCuc) is presented at Figure 15(b). Fuzzy sets (L)and (H) represent the ranges of ultracapacitor operatinglimit value. Fuzzy set (N) represents the optimal SOC range.
Depending on the instantaneous load power demand, themembership function for renewable power (Pren) is dividedof two ranges: positive (+) and negative (−) (Figure 15(c)).
The fuzzy set (−) represents the range where therenewable sources must be assisted by ESS or CE to satisfythe power demand. Fuzzy set (+) represents the autonomoussupply mode ensured by renewable sources (ESS is onrecharge mode and CE is on Off mode).
The demand power variation was divided using the vari-ables positive (+) and negative (−), as shown in Figure 15(d).This information is useful to prevent abusive use of the
ISRN Renewable Energy 9
Loadmodel
RS model
ESSmodel
CE
Low-level controller
ToDynamicfuzzy
controller
Estimated loadprofiles
Extern constraints
IrradiationtemperatureWind speed
PloadΔPloadSOCbatSOCucPrenPce-pro
PbatPucPren
Pce-proPce-req
Figure 14: Block diagram of dynamic fuzzy logic controller for power management strategy.
SOCmaxSOCmin 10
HNTL
0.5
1 L
ReservelevelD
egre
e of
mem
bers
hip
(a)
L
0
HN
0.5
1
1 SOCmax
Deg
ree
ofm
embe
rsh
ip
(b)
Deg
ree
ofm
embe
rsh
ip +−
0
0.5
1
1 Phvd
(c)
Deg
ree
ofm
embe
rsh
ip +−
0
0.5
1
ΔPhvd-limit
(d)
Transient state
Steady state +−
Deg
ree
ofm
embe
rsh
ip
Pce-slo0
0.5
1
(e)
Figure 15: Input membership functions for: (a) SOCbat, (b) SOCuc, (c) renewable energy power, (d) load demand power variation, and (e)CE power (Pce-slo: combustion engine slow down power).
batteries. In fact, when the demand power variation is belowa predetermined level, the ESS discharge-mode must beensured by batteries; otherwise, the ESS discharge modemust be ensured by ultracapacitor (absorption the peakingpower demand).
The membership function for CE power is presentedat the Figure 15(e). The fuzzy sets (+) represent the CEProduction mode (the power from the generator is effectiveto supply the load and, simultaneously, recharge the ESS ifnecessary). The fuzzy sets (−) represent the CE slow-downmode or transient state (the generator do not contribute toprovide load power).
4.3. Rules. The rules were chosen very intuitively accordingto the main following points.
(1) A RES takes over as main energy source.
(2) In the optimal conditions, the battery bank and theultracapacitor have been dimensioned in order tosatisfy the maximum load power demand; that is,100% of the energy demand can be supplied by theESS.
(3) An ESS comes into play in order to balance therenewable power fluctuations and to allow for a CEOff mode (zero emission).
(4) The CE is turned on when the ESS state of chargedrops below a minimum value and renewable poweris insufficient to meet the load.
(5) The minimum period of CE operation must cover theESS time of recharge.
The latter (4 and 5) are important to avoid running the CEfor very short period of time.
Using these linguistic directives, it makes easy to obtaina set of fuzzy rules. Thus, the fuzzy rule basis consists ofa collection of fuzzy if-then rules type. For example, oneof the energy management rules (rule1) can be stated as(IF the renewable power exceeds load demand power ANDbattery state of charge is high AND ultracapacitor state ofcharge is high THEN the renewable produced power is equalto load demand power). Using these linguistic variables, aset of fuzzy rules was developed. For the sake of simplicity,the rules number was minimized to 16 rules. These rulesare shown in (Table 2). Where Pbat max is a maximumbattery power, Puc max is a maximum ultracapacitor power,
10 ISRN Renewable Energy
1200
800
400
0
200
300
400
0
Time (s) Time (s)
150
50
5.5 7
150
110
7.2 7.5
SOCbat15
10
5
−5
−15
−250 5 10 15 20
1
0.8
0.6
0.4
0.2
1
0.8
0.6
0.4
0.2
00
5 15 20
SOCsc
0
−10
−20
Pce-proPce-con
Pce-proPce-con
Pload-conPload-demPren
Pload-conPload-demPren
ZoomZoom
0 5 10 15 20 0 5 10 15 20
10(a) (b)
(c) (d)
(e) (f)
Ibat ISC
Figure 16: Simulation results for the power balance of the HES: (a) and (b) present the SOC and currant (A), respectively, of battery andultracapacitor, (c) shows the generated and consumed power (W) output from the combustion engine, and (d) shows the generated power(W) output from the renewable sources, the demanded and consumed power by the load.
Pbat rech is a recharge battery power, and Puc rech is a rechargeultracapicitor power.
The logical AND (resp.) has been implemented with theminimum (resp. maximum) operator.
While the fuzzy input variables are obtained either froma prediction method based on the model of each source orelse using local measurements, to obtain the output of thecontroller, the degrees of membership functions of the if-parts of all rules are evaluated, and the THEN-parts of allrules are averaged, weighted by these degrees of membershipfunctions.
5. Application to the Multisource System
The objective of this simulation is not to produce a realisticpractical scenario of HES, but to validate the developedenergy management strategy. For this reason, the simulationsare made with simplified models of renewable sources, CE,
battery and ultracapacitor. Furthermore, the simulations aimto show that the developed energy management strategycan operate in different system configurations with differentcharacteristics. Therefore, any HES state can be controlledwithout redesign the energy management strategy.
In the presented simulations, to illustrate all possibleoperating modes of multisource system, I suppose that thebattery time constant is reduced to few seconds (5 s). Thesimulation results for the power balance of the HES arepresented in Figure 16.
It can be observed from these results that the optimaloperation objective of HES as mentioned below has beenachieved.
(1) The renewable sources take over as main energysources.
(2) An energy storage system comes into play in orderto balance the renewable power fluctuations and toallow for a CE Off mode (zero emission).
ISRN Renewable Energy 11
Ta
ble
2:R
ule
s.
Ru
leIfP
ren
isA
nd
SOC
uc
isA
nd
SOC
bat
isA
nd
ΔP
hvd
isA
nd
Pce
pro
isT
hen
Pba
tA
ndP
uc
An
dP
ren
An
dP
cere
qA
ndP
cepr
o
1+
HH
//
00
Phv
d0
02
+H
N/
−P
bat
rech
0M
in(P
ren,(P
hvd
+P
bat
max
))0
03
+N
OR
HL
/−
Pba
tre
ch0
Min
(Pre
n,(P
hvd
+P
bat
max
))0
04
+/
TL
/−
Pba
tre
ch0
Min
(Pre
n,(P
hvd
+P
bat
max
))0
05
+N
NO
RH
/−
0P
uc
rech
Min
(Pre
n,(P
hvd
+P
uc
max
))0
06
+L
L,N
OR
H/
−0
Pu
cre
chM
in(P
ren,(P
hvd
+P
uc
max
))0
0
7−
/N
OR
H−
−P
hvd−
Pre
nm
ax0
Pre
nm
ax0
0
8−
LN
OR
H+
−P
hvd−
Pre
nm
ax0
Pre
nm
ax0
0
9−
NO
RH
LO
RT
L/
−0
Phv
d−
Pre
nm
axP
ren
max
00
10−
NO
RH
NO
RH
+−
Pba
tm
ax
Phv
d−
Pre
nm
ax−
Pba
tm
ax
Pre
nm
ax0
0
11−
LL
/−
Phv
d−
Pre
nm
ax0
Pre
nm
ax0
Phv
d(s
tart
up)
12−
TL
L/
−0
0P
ren
max
0P
hvd
(sta
rtu
p)
13/
L,N
TL,
L,N
/+
Pba
tre
chP
uc
rech
Min
(Pre
n,(P
hvd
+P
bat
max
+P
uc
max
))M
ax(0
,min
(Pce
pro,P
hvd−P
ren
+P
bat
rech
+P
uc
rech
))M
ax(P
cesl
o,P
hvd−P
ren
+P
bat
max
+P
uc
max
)
14/
L,N
H/
+0
Pu
cre
chM
in(P
ren,(P
hvd
+P
uc
max
))M
ax(0
,min
(Pce
pro,P
hvd−P
ren
+P
uc
rech
))M
ax(P
cesl
o,P
hvd−P
ren
+P
uc
max
)
15/
HT
L,L,
N/
+P
bat
rech
0M
in(P
ren,(P
hvd
+P
bat
max
))M
ax(0
,min
(Pce
pro,P
hvd−P
ren
+P
bat
rech
))M
ax(P
cesl
o,P
hvd−P
ren
+P
bat
max
)
16−
HH
/+
00
Min
(Pre
n,P
hvd)
Max
(0,m
in(P
cepr
o,P
hvd−P
ren))
Max
(Pce
slo,P
hvd−P
ren)
12 ISRN Renewable Energy
(3) The battery and ultracapacitor SOCs were stable basi-cally and decreased (current is negative)/increased(current is positive) slightly from 0.3 to 0.9 (theoperation limits).
(4) The CE turns on when the ESS state of chargedrops below a minimum value, and renewablepower is insufficient to meet the load (as shown inFigure 16(c) at 2 s and 17 s).
(5) The minimum period of CE operation covers the ESStime of recharge (as shown in Figure 16(c) between2 s–7 s and 17 s–20 s).
(6) Load demand is satisfied for all time which preservethe reliability of the system (Figure 16(d)).
6. Conclusion and Prospects
This paper presented the basis for the development of anintelligent and autonomous energy management strategy forHPS. Two hierarchical levels are presented to control andmanage the HPS. The low level is performed by a localsliding mode control unit (DC-DC converters controller)of the different power sources. This control techniqueprovides good overall performances compared to standardcurrent control and good robustness against load and inputvoltage variations. The high level is performed by the onlinesupervisor unit. This unit is designed by applying onlinedynamic Takagi-Sugeno fuzzy logic principles. As a resultthe robust control system gets rid of the limits of the HPS,which has the imprecision, uncertainty, strong coupling, andnonlinearity.
Under the operation constraints related to each type ofsources, it can be observed from the simulation results thatthe optimal operation objective of HES has been achieved(the load demands are satisfied for all time, the energystorage system is managed in an optimal way, and the fueluse is optimal).
The application of the proposed control scheme tolarge-scale renewable multi-source power systems will beconsidered in my future work.
The neural-based adaptive observer, in order to identifyunknown functions in the multi-source system and toestimate the unmeasured states, will be also studied in myfurther works.
References
[1] Y. Chen and W. Jie, “Agent-based energy management andcontrol of a grid-connected wind/solar hybrid power system,”in Proceedings of the 11th International Conference on ElectricalMachines and Systems (ICEMS ’08), pp. 2362–2365, October2008.
[2] E. Ortjohann, O. Omari, M. Lingemann et al., “An onlinecontrol strategy for a modular DC coupled hybrid powersystem,” in Proceedings of the European Conference on PowerElectronics and Applications (EPE ’07), September 2007.
[3] D. Lu, T. Zhou, H. Fakham, and B. Francois, “Design of apower management system for an active PV station includingvarious storage technologies,” in Proceedings of the 13th
International Power Electronics and Motion Control Conference(EPE-PEMC ’08), Poznan, Poland, September 2008.
[4] M. E. Torres-Hernandez and M. Velez-Reyes, “Hierarchicalcontrol of Hybrid Power Systems,” in Proceedings of the 11thIEEE International Conference on Power Electronics Congress(CIEP ’08), pp. 169–176, August 2008.
[5] A. Bilodeau and K. Agbossou, “Control analysis of renewableenergy system with hydrogen storage for residential applica-tions,” Journal of Power Sources, vol. 162, no. 2, pp. 757–764,2006.
[6] S. Kermani, S. Delprat, R. Trigui, and T. M. Guerra, “Predictiveenergy management of hybrid vehicle,” in Proceedings of theIEEE Vehicle Power and Propulsion Conference (VPPC ’08),September 2008.
[7] M. Mohebbi, M. Charkhgard, and M. Farrokhi, “Optimalneuro-fuzzy control of parallel hybrid electric vehicles,” in Pro-ceedings of the IEEE Vehicle Power and Propulsion Conference(VPPC ’05), pp. 252–256, September 2005.
[8] F. Soltani and N. Debbache, “Integration of converter lossesin the modelling of hybrid photovoltaic-wind generatingsystem,” European Journal of Scientific Research, vol. 21, no. 4,pp. 707–718, 2008.
[9] R. Belfkira, R. Reghem, J. Raharijaona, G. Barakat, and C.Nichita, “Non linear optimization based design methodologyof wind/PV hybrid stand alone system,” in Proceedings of theEuropean Association for Vision and Eye Research (EVER ’09),Monaco, France, March 2009.
[10] X. Lu and S. H. Yang, “Solar energy harvesting for ZigBeeelectronics,” in International Conference on Sustainability inEnergy and Buildings, pp. 19–27, Brighton, UK, 2009.
[11] B. Ai, H. Yang, H. Shen, and X. Liao, “Computer-aided designof PV/wind hybrid system,” Renewable Energy, vol. 28, no. 10,pp. 1491–1512, 2003.
[12] M. El Mokadem, C. Nichita, B. Dakyo, and W. Koczara,“Control strategy for a variable load supplied by a wind dieselsystem,” Electromotion Journal, vol. 10, pp. 635–640, 2003.
[13] W. X. Shen, “State of available capacity estimation for lead-acid batteries in electric vehicles using neural network,” EnergyConversion and Management, vol. 48, no. 2, pp. 433–442, 2007.
[14] M. B. Camara, H. Gualous, F. Gustin, and A. Berthon,“Design and new control of DC/DC converters to share energybetween supercapacitors and batteries in hybrid vehicles,”IEEE Transactions on Vehicular Technology, vol. 57, no. 5, pp.2721–2735, 2008.
[15] J. K. Kaldellis and D. Zafirakis, “Optimum energy storagetechniques for the improvement of renewable energy sources-based electricity generation economic efficiency,” Energy, vol.32, no. 12, pp. 2295–2305, 2007.
[16] J. B. Copetti, E. Lorenzo, and F. Chenlo, “A general batterymodel for PV system simulation,” Progress in Photovoltaics:Research and Applications, vol. 1, pp. 283–292, 1993.
[17] O. Gergaud, G. Robin, B. Multon, and H. Ben Ahmed,Energy Modeling of a Lead-Acid Battery within HybridWind/Photovoltic Systems, EPE Toulouse, 2003.
[18] A. Sripakagorn and N. Limwuthigraijirat, “Experimentalassessment of fuel cell/supercapacitor hybrid system forscooters,” International Journal of Hydrogen Energy, vol. 34, no.15, pp. 6036–6044, 2009.
[19] S. Zerkaoui, A. B. Mboup, D. Lefebvre, F. Guerin, J. Bosche,and A. El Hajjaji, “Sliding mode based control strategy formulti-sources renewable energy system,” in Proceedings ofthe International Conference on Electric Power and EnergyConversion Systems (EPECS ’09), November 2009.
ISRN Renewable Energy 13
[20] J. A. Sabate, V. Vlatkovic, R. B. Ridley, F. C. Lee, and B. H.Cho, “Design considerations for high-voltage high-power full-bridge zerovoltage-switched PWM converter,” in IEEE AppliedPower Electronics Conference and Exposition, pp. 275–284,March 1990.
[21] S. J. Jeon and G. H. Cho, “A zero-voltage and zero-currentswitching full bridge DC-DC converter with transform isola-tion,” IEEE Transactions on Power Electronics, vol. 16, no. 5, pp.573–580, 2001.
[22] P. Z. Lin, C. M. Lin, C. F. Hsu, and T. T. Lee, “Type-2fuzzy controller design using a sliding-mode approach forapplication to DC-DC converters,” IEE Proceedings, ElectricPower Applications, vol. 1526, pp. 1482–1488, 2005.
[23] J. Sun and H. Grotstollen, “Symbolic analysis methods foraveraged modeling of switching power converters,” IEEETransactions on Power Electronics, vol. 12, no. 3, pp. 537–546,1997.
[24] A. B. Mboup, F. Guerin, P. A. Ndiaye, and D. Lefebvre,“Control design for hybrid system based on dc /dc convertersduty cycle value,” COMPEL, vol. 30, no. 1, pp. 310–335, 2011.
[25] V. I. Utkin, “Variable structure systems with sliding modes,”IEEE Transactions on Automatic Control, vol. 22, no. 2, pp.212–222, 1977.
[26] A. Hijazi, M. Di Loreto, E. Bideaux, P. Venet, G. Clerc, and G.Rojat, “Sliding mode control of boost converter: applicationto energy storage system via supercapacitors,” in Proceedingsof the 13th European Conference on Power Electronics andApplications (EPE ’09), September 2009.
[27] I. S. Kim, “Sliding mode controller for the single-phase grid-connected photovoltaic system,” Applied Energy, vol. 83, no.10, pp. 1101–1115, 2006.
[28] A. J. Forsyth and S. V. Mollov, “Modelling and control of DC-DC converters,” Power Engineering Journal, vol. 12, no. 5, pp.229–236, 1998.
