Hyd ogen Economy o he Fuel Cell Hyb id Powe Sys em op imized by ai low con ol o
mi iga e he e ec o he unce ain y abou a ailable enewable powe and load dynamics
Nicu Bizon
1) Facul y o Elec onics, Communica ion and Compu e s, Uni e si y o Pi es i, 1 Ta gu din Vale,
110040, Pi es i, Romania
2) Uni e si y Poli ehnica o Bucha es , 313 Splaiul Independen ei, 060042, Bucha es , Romania
[email p o ec ed]om, ORCID: 0000-0001-9311-7598
Jose Manuel Lopez-Guede
Sys ems Enginee ing and Au oma ic Con ol Depa men , Facul y o Enginee ing Vi o ia-Gas eiz,
Uni e si y o he Basque Coun y (UPV/EHU), Vi o ia, Spain, [email protected], ORCID: 0000-
0002-5310-1601
E ol Ku
Gazi Uni e si y, Technology Facul y, Depa men o Elec ical and Elec onics Enginee ing, 06500
Teknikokulla , Anka a, Tu key, [email p o ec ed], ORCID: 0000-0002-3615-6926
1
2
Nomencla u e
Ai F
AV
EMS
ESS
FuelF
FC
Ai Flow a e
A e age alue
Ene gy Managemen S a egy
Ene gy S o age Sys em
Fuel Flow a e
Fuel cell
This is he accep ed manusc ip o he a icle ha appea ed in inal o m in Ene gy Con e sion and Managemen 179 :
152-165 (2019) , which has been published in inal o m a h ps://doi.o g/10.1016/j.enconman.2018.10.058. © 2018 Else ie
unde CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
GES
HPS
LF
MV
PEMFC
RTO
RES
sFF
Global Ex emum Seeking
Hyb id Powe Sys em
Load Following
Mean Value
P o on Exchange Memb ane Fuel
Real-Time Op imiza ion
Renewable Ene gies Sou ce
S a ic Feed-Fo wa d
3
Abs ac : A new Ene gy Managemen S a egy o educe he hyd ogen consump ion is p oposed 4
o Hyb id Powe Sys ems based on P o on Exchange Memb ane Fuel Cell sys em used as backup 5
sou ce. The Ene gy Managemen S a egy uses a Load Following con ol loop o eques ed load 6
demand on DC bus and an op imiza ion con ol loop o imp o e he uel economy based on he 7
Global Ex emum Seeking algo i hm applied o he ai low a e. The pe o mance o p oposed 8
s a egy is compa ed o he one ob ained wi h he S a ic Feed-Fo wa d s a egy conside ing h ee 9
case s udies o he op imiza ion unc ion used in di e en scena ios o powe low on DC bus 10
( a iable o cons an load demand, wi hou o wi h a iable enewable ene gy powe ). 11
Keywo ds: P o on Exchange Memb ane Fuel Cell; Hyd ogen Economy; Powe Va iabili y 12
Mi iga ion; Ai Flow Con ol; Global Ex emum Seeking; Load Following 13
14
1. INTRODUCTION15
In he las yea s Renewable Ene gies Sou ces (RESs) ha e expe ienced a apid de elopmen [1] and 16
hese (especially wind and sola ene gy [2]) a e now ecognized as impo an ene gy sou ces o 17
mic o-g ids [3]. Fu he mo e, wo ldwide laws ha e been adop ed o encou age he use o enewable 18
ene gies [4,5]. RESs used in sma g ids is impo an o eco- iendly de elopmen in all coun ies 19
[6,7]. 20
In addi ion o he mos used enewable ene gies (sola and wind ene gy), o he esou ces a e used as 21
well (such as geo he mal, biomass and biogas ene gies) along wi h backup sou ces (such as he 22
P o on Exchange Memb ane Fuel Cell (PEMFC) o he diesel gene a o ) and Ene gy S o age 23
Sys ems (ESSs) in o de o mi iga e he a iabili y o he RES powe low [8,9]. A Hyb id Powe 24
Sys em (HPS) always uses wo o mo e ene gy sou ces o sus ain he powe demand based on 25
app op ia e Ene gy Managemen S a egy (EMS) designed o manage he powe lows on DC bus 26
[10]. I o each egion he RES a e known and chosen acco ding o he equi emen s o he powe 27
demand, he use o ossil uels will be conside ably educed [7-9,11]. 28
RESs a e e y a ac i e o he in es o s bu he RES powe low (PRES) and load demand (Pload) 29
a e a iable and di icul o be p edic ed [12]. So, HPS op imal ope a ion mus be ensu ed wi h an 30
app op ia e EMS design [10,13]. The RES powe low depends on wea he condi ions and load 31
demand on many ac o s, so ESSs s o e he excess ene gy o supply he needed powe due o lack o 32
RES powe on DC bus [14,15]. The EMS will mo o ize he cha ging ESS s a us in o de o 33
gua an ee he HPS ope a ion unde high dynamic load p o iles [16,17]. The powe low eques ed 34
on DC bus (PDC eq) mus ollow he powe low Pload - PRES using he Load Following (LF) con ol 35
[15,18,19]. Du ing ligh load s ages, when Pload < PRES, he excess o powe PRES - Pload will be 36
s o ed o used o supply an elec olyze [16,17]. Many op imiza ion echniques o he HPS 37
ope a ion we e p oposed in he li e a u e [20-24] using di e en algo i hms [25], including he 38
Global Ex emum Seeking (GES) algo i hm [26-30]. EMSs mus be implemen ed using Real-Time 39
Op imiza ion (RTO) loops [6,15,16,31]. 40
In his pape he RTO1 s a egy is p oposed based on LF con ol o he boos con e e and GES 41
op imiza ion o he Ai Flow a e (Ai F ) o he PEMFC sys em, which is used he e as 42
en i onmen al- iendly backup sou ce. 43
In he las decade, he en i onmen al- iendly backup sou ces such as PEMFC sys ems a e p e e ed 44
o non- enewable ene gy backup sou ces such as diesel gene a o s due o some disad an ages o he 45
las : high p ice o main enance and he pollu an emissions [32]. Fu he mo e, he hyd ogen ha is 46
supplied o he PEMFC could be gene a ed enewably using an elec olyze [33-35], as i was 47
men ioned abo e. 48
When wo o mo e ene gy sou ces and loads a e in ol ed in he HPS opology, he EMS is equi ed 49
o be app op ia ely designed based on s a e- low diag ams and RTO algo i hms. Di e en EMSs 50
ha e been p oposed in he li e a u e ha can ensu e he load demand and sa e ope a ion o he HPS 51
[36-44]. 52
The RTO1-based EMS has supplemen a y only he RTO con ol loop compa ed o he S a ic Feed-53
Fo wa d (sFF) s a egy [44], which is mos know and al eady comme cially implemen ed. No e ha 54
bo h RTO1 and sFF s a egies use he LF con ol loop based on PDC eq = Pload - PRES and he con ol 55
o he Fuel Flow a e (FuelF ) based on FC cu en . The RTO con ol loop uses he GES algo i hm 56
o ind he global maximum o he op imiza ion unc ion due o high pe o mance (high accu acy, 57
speed, and sea ching esolu ion) epo ed in [25-29]. The opological di e ences be ween o he 58
RTO s a egies p oposed in he li e a u e a e men ioned in Table 1. 59
60
Table 1. The opological di e ences be ween he RTO s a egies 61
62
The mi iga ion o he unce ain y o a ailable powe om enewable sou ces and he a iabili y o 63
load demand is analyzed in his pape based on LF con ol o he boos con e e . The hyd ogen 64
economy is ob ained based on op imized con ol applied o ai low egula o . The pe o mance o 65
he RTO1-based EMS is alida ed using di e en scena ios o powe low on DC bus ( a iable o 66
cons an load demand, wi hou o wi h a iable enewable ene gy powe ). 67
Thus, he s uc u e o he pape is as ollows. EMSs a e p esen ed and compa ed in Sec ion 2. The 68
p oposed HPS is modeled in Sec ion 3 and he RTO1-based EMS is p esen ed in Sec ion 4. The 69
esul s a e p esen ed and discussed in Sec ion 5. Las sec ion concludes he pape . 70
71
2. ENERGY MANAGEMENT STRATEGIES FOR HYBRID POWER SYSTEM 72
The necessi y o EMSs is no only o s andalone hyb id sys em bu also o sys ems connec ed o 73
he powe g id. Fo s andalone HPSs, EMSs ensu e he con inuous powe supply o he load 74
ega dless o he condi ions, he use o all enewable esou ces o he ulles educing p oduc ion 75
cos s o minimize he cos o ene gy and inc easing he sys em eliabili y. In addi ion, in he case o 76
HPSs connec ed o he powe g id, EMSs ensu e he con ol o ene gy low om he g id o 77
consume s and ice e sa ( om he hyb id sys em o he g id), in addi ion o measu e pa ame e s o 78
minimize he cos s [45,46]. 79
The e a e in he li e a u e a numbe o e iew pape s abou ene gy managemen s a egies 80
[5,6,10,17,26,38,39,47]. Fo example, in [47] he e a e p esen ed he p incipal con igu a ions, usual 81
sizing me hodologies and main EMSs used o HPSs, which can be classi ied in cen alized (Figu e 82
1), dis ibu ed (Figu e 2) and hyb id (Figu e 3) EMSs. The main goals o any s a egy a e o mee 83
he demands o consume s, ex ac he maximum amoun o ene gy om each enewable sou ce, 84
minimize he cos o ene gy, and educe he numbe o loading and unloading cycles o ESSs. 85
I can be obse ed o cen alized EMSs (Figu e 1) ha wo communica ions a e p esen o each 86
componen o he sys em: one o ans e da a om he cen e o he componen s and ano he o he 87
e e se lux. So, cen alized EMSs a e no e ec i e in sma mic o-g ids, because u u e opologies 88
a e always subjec o changes, esul ing in low s abili y o he sys em [48]. Also, when a sys em has 89
many componen s, his is no a cos e icien solu ion due o he high numbe o communica ions 90
channels be ween componen s. Howe e , cen alized me hods canno be ye eplaced by o he new 91
me hods p oposed in he li e a u e o dis ibu ed o hyb id ypes [49]. 92
In dis ibu ed EMSs (Figu e 2) each componen has i s own local con olle and each componen 93
p o ides and uses an app op ia e se o measu emen s (cu en s and ol ages), ac ha implies 94
mo e co ec decisions aken by he con olle . The ad an age o dis ibu ed s a egies is he 95
inc eased s abili y in compa ison wi h he cen alized ones and he abili y o p o ec da a om he 96
componen s [50]. 97
In compa ison wi h dis ibu ed EMSs (whe e componen s p o ides/uses a se o measu emen s 98
wi hou sha ing in o ma ion wi h o he componen s o he sys em), hyb id EMSs (Figu e 3) allow 99
hei componen s o sha e in o ma ion be ween hem, esul ing in high lexibili y and s abili y o 100
hese mic o-g ids. The ad an age o hese s a egies i is i s use in he u u e sma g ids due o he 101
ac ha new elemen s can be added wi hou a ec ing he EMS [51]. 102
EMSs mus be designed conside ing he opology o he HPSs [52-54]. An algo i hm o p edic 103
sizes o hyb id sou ces and an EMS o sha e powe be ween he PEMFC sys em and ESS is 104
p oposed in [52] o au omo i e applica ions. A e iew ega ding modeling, ene gy low 105
managemen and con ol s a egies o RES HPSs is p esen ed in [53]. The issues ela ed o HPSs 106
con igu a ions and ene gy managemen using PV panels, wind u bines, hyd o-powe s a ions and 107
uel cells a e e ised and some solu ions a e p oposed o o e come hese p oblems. So, in he nex 108
sec ion, he p oposed HPS is p esen ed in he ame o his classi ica ion p oposed o HPSs. 109
110
Figu e 1. Cen alized EMSs.
Figu e 4. The DC bus HPS opology.
Figu e 2. Dis ibu ed EMSs.
Figu e 5. The AC bus HPS opology.
Figu e 3. Hyb id EMSs.
Figu e 6. The hyb id bus HPS opology.
111
3. PROPOSED PEMFC HYBRID POWER SYSTEM TOPOLOGY 112
The classi ica ion o HPSs can be made depending on he ype o common bus used [53-58]. The 113
mos known HPS opologies a e he DC bus HPS opology (Figu e 4), he AC bus HPS opology 114
(Figu e 5) and hyb id bus HPS opology (Figu e 6). In each opology, he DC o AC bus is used o 115
connec he ene gy sou ces and he load (and he g id i he HPS opology is g id connec ed) using 116
app op ia e powe con e e s. The EMSs con ol he powe low o he a ailable ene gy sou ces o 117
ensu e he load demand based on he powe low balance on common bus [57]. The ad an ages and 118
disad an ages o he DC bus opology compa ed wi h he AC bus opology a e e ised in [58]. 119
Anyway, a disad an age o all opologies is clea ly ep esen ed by he inc eased numbe o 120
con e sion s ages, bu his could be educed by he hyb id bus opology ha uses bidi ec ional 121
con e e s [54-56]. Howe e , he EMSs o his hyb id bus opology a e mo e complex due o he 122
ene gy managemen o bo h buses and he con ol o bidi ec ional con e e be ween hem [57,58]. 123
A bidi ec ional hyb id bus HPS opology is p oposed in [58] o simpli y he ene gy managemen 124
be ween he DC and AC buses based on bidi ec ional DC-AC con e e s (Figu e 6). 125
The HPS opology conside ed in his s udy is educed o a DC bus modeling he o he pa s o he 126
HPS wi h an equi alen DC load (Figu e 7). So, he in e e sys em (g id-connec ed o no ) is 127
modeled on DC bus wi h an equi alen load wi h dynamic p o ile as in eali y. Choosing a speci ic 128
HPS opology depends on se e al ac o s such as he RES a iabili y, load demand dynamics and 129
economic cons ain s. 130
The RTO1-based HPS opology is ob ained wi h he swi ch on he GES posi ion, as i is shown in 131
Figu e 7, when he GES op imiza ion con ol is applied o Ai F egula o and he FuelF egula o 132
is con olled by he FC cu en . The sFF s a egy (used as e e ence o compa e he esul s ob ained 133
wi h he RTO1 s a egy) will be ob ained i he swi ch is mo ed o he sFF posi ion and bo h Ai F 134
and FuelF a iables a e con olled by he FC cu en [34]. The LF con ol is applied o boos 135
con e e in bo h RTO1 and sFF s a egies. The 6 kW / 45 V PEMFC model om he lib a y o 136
Ma lab-Simulink (wi h he FC ime cons an se o 0.1 s) is used. The FC ol age is boos ed o 137
VDC VDC( e ) = 200 V using a boos con e e . The boos con olle is o hys e e ic ype, ha ing as 138
inpu s he FC cu en (IFC) and he e e ence cu en (I e LF). The e e ence cu en I e LF is gene a ed 139
by LF con ol block, as will be de ailed in nex sec ion. 140
141
Figu e 8. The LF con ol block. 142
Figu e 7. The RTO1-based HPS opology. 143
Figu e 9. The GES con ol block. 144
4. MODELING OF HYBRID POWER SYSTEM 145
The powe low balance on DC bus HPS shown in Figu e 7 is gi en by (1): 146
CDC
udc
dudc/d =
boos
pFC+ pESS - pLoad
(1)
whe e CDC is he capaci o on DC bus (which is used o il e he ol age on DC bus, udc),
boos is 147
he ene gy e iciency o he boos con e e (se o 95%) and pFC, pESS , and pLoad a e he FC ne 148
powe , he ESS powe , and he load demand eques ed on DC bus espec i ely. 149
The a e age alue (AV) o powe low balance o (1) is gi en by (2): 150
0 =
boos PFC(AV)+ PESS(AV) - PLoad(AV)
(2)
whe e
boos is he ene gy e iciency o he boos con e e , which was se a 95%. 151
So, since ha he LF con ol o he powe eques ed on DC bus (PDC eq(AV)
PDC(AV) = PLoad(AV) – 152
PRES(AV)) is implemen ed o he FC sys em, he FC powe will ollow he a e age alue o needed 153
ene gy on DC bus (PDC eq(AV)): 154
PFC(AV)= PDC eq(AV) /
boos
(3)
The LF e e ence cu en will be: 155
I e (LF) =IFC(AV)= PDC eq(AV) / (VFC(AV
boos )
(4)
Thus, he AV o he ESS powe will be almos ze o (conside ing (1-4)): 156
PESS(AV)
0
(5)
Consequen ly, he ESS size can be educed a minimum. 157
The LF con ol block (Figu e 8) implemen s (3), wi h PDC eq(AV) ob ained by Mean Value (MV) o 158
he pDC eq powe gi en by Eq. (6), bu o he il e ing echniques could be used as well [38]: 159
pDC eq = pload
(6)
The GES con ol block (Figu e 9) implemen s he ela ionships (6) [6]: 160
12
,y
,
N Ny
y k y
(6a)
h h N
y y y
,
HPF N
y y y
,
l BPF l HPF
BPF
y y y
(6b)
, sin( )
DM BPF d d
y y s s
,
(6c)
DM
In
yy
(6d)
1
,
d MV MV BPF
d
G y y y d
T
(6e)
Md
yG
(6 )
1 1 1
,
In sd
p k y k
(6g)
22Md
p k y s
(6h)
3md
p A s
(6i)
1 2 3 e GES Np
I k p p p
,
(6j)
161
whe e (6a) o (6 ) ep esen he op imiza ion unc ion, he inpu no maliza ion gain (kNy), he high-162
pass il e (HPF) and he band-pass il e (BPF), he demodula ion, he in eg a ion, he compu ing o 163
he di he gain (Gd) based on AV o he ybp signal, and he signal ha will modula e he di he [6]. 164
The componen s p1, p2, and p3 o he sea ching signal (p) a e gi en by (6g) o (6i), and he e e ence 165
cu en I e GES by (6j), whe e and kNp is he ou pu no maliza ion gain [6]. 166
The swi ch selec s he e e ence cu en s o I e (O2) as I e GES and I e (H2) as IFC in he RTO1, while 167
I e (H2) = I e (O2) = IFC in sFF s a egy. 168
The no maliza ion gains a e kNy= 1/YMax and kNp= IFC( a ed) / 2, whe e IFC( a ed) and YMax a e he 169
nominal alue o he FC cu en and he es ima ion o maximum alue o he op imiza ion unc ion 170
espec i ely. The uning pa ame e s a e designed based on [59,60]. Apa om o he op imiza ion 171
echniques p oposed in he li e a u e [26-33,61-66], he op imiza ion unc ion used he e is de ined 172
o inc ease he FC sys em ene gy e iciency and educe he o al uel consump ion by maximizing 173
(7a): 174
( , , , )
ne FCne uel e Load
k P k Fuel x Ai F FuelF P
(7a)
subjec o dynamics gi en by Eq. [7b]:
(7b)
6. CONCLUSION 306
In his las sec ion, he main indings o his s udy a e summa ized as ollows: 307
Bo h s a egies (sFF and RTO1) use he LF con ol o he FC boos con e e so ha he FC 308
powe ensu es he load demand on DC bus, ope a ing he ba e y in cha ge sus ained mode; 309
The RTO1-based EMS p oposed he e has be e pe o mance compa ed wi h ha o sFF 310
s a egy by adding only a op imiza ion loop o he Ai F egula o , which is implemen ed in 311
eal- ime by using he GES algo i hm; 312
The pe o mance indica o s such as he FC ene gy e iciency, he uel e iciency, and he 313
uel economy could inc ease up o 2.65 %, 10.35 W/lpm, and 11.8 l o he HPS unde 8 kW 314
load; 315
The uel economy could be inc eased 2- imes using k uel=25 in he op imiza ion unc ion 316
de ined as a mix ela ionship o he FC ne powe and he uel e iciency (see Table 5 o 317
k uel=0 and k uel=25); 318
The uel economy is ob ained on he ull ange o load demand, cons an o a iable, by 319
using he RTO1 s a egy wi h k uel0 (see Table 5 and Table 7); 320
The e ec o a iabili y o he load demand dynamics and unce ain y o a ailable powe 321
om enewable sou ces can be mi iga ed by LF con ol o he FC powe gene a ed, which 322
will ollow he powe demand eques ed on DC bus; 323
The nex wo k will be ocused on ope a ion o he HPS wi h a iable RES powe and dynamic load 324
p o ile ha will use an elec olyze supplied wi h ene gy in excess (PRES(AV) - Pload) in o de o 325
main ain he cha ge sus ained mode o he ba e y and a leas he pe o mances epo ed in his 326
pape . 327
328
ACKNOWLEDGEMENTS 329
This wo k was suppo ed by Resea ch Cen e “Modeling and Simula ion o he Sys ems and 330
P ocesses” based on g an s o he Minis y o Na ional Educa ion and Scien i ic Resea ch, 331
CNCS/CCCDI-UEFISCDI wi hin PNCDI III “Expe imen al alida ion o a p opulsion sys em wi h 332
hyd ogen uel cell o a ligh ehicle - Mobili y wi h Hyd ogen Demons a o ”, 53PED, ID: PN-III 333
P2-2.1-PED-2016-1223, and wi hin RDI P og am o Space Technology and Ad anced Resea ch - 334
STAR, p ojec numbe 167/2017: “Concep De elopmen o an Ene gy S o age Uni Using High 335
Tempe a u e Supe conduc ing Coil o Spacec a Powe Sys ems (SMESinSpace). 336
337
REFERENCES 338
[1] Pe e a ATD, A alage RA, Pe e a KKCK, Dassanayake VPC. Designing s andalone hyb id 339
ene gy sys ems minimizing ini ial in es men , li e cycle cos and pollu an emission. Ene gy 340
2013;54(0):220–30. 341
[2] Banda a K, Swee T, Ekanayake J. Pho o ol aic applica ions o o -g id elec i ica ion using 342
no el mul i-le el in e e echnology wi h ene gy s o age. Renew Ene gy 2011;37(1):82–8. 343
[3] Mogha emi M, Ismail MS, Mu ali B, Yang SS, A a an A, Mogha emi S. De elopmen and 344
op imiza ion o a PV/diesel hyb id supply sys em o emo e con olled comme cial la ge scale 345
FM ansmi e s. Ene gy Con e s Manag 2013;75:542–51. 346
[4] Balcombe P, Rigby D, Azapagic A. Mo i a ions and ba ie s associa ed wi h adop ing 347
mic ogene a ion ene gy echnologies in he UK. Renew Sus ain Ene gy Re 2013;22(0):655–348
66. 349
[5] Meya -Naimi H, Vaez-Zadeh S. Sus ainable de elopmen based ene gy policy making 350
amewo ks, a c i ical e iew. Ene gy Policy 2012;43(0):351–61. 351
[6] Bizon N, Thoun hong P. Real- ime s a egies o op imize he ueling o he uel cell hyb id 352
powe sou ce: A e iew o issues, challenges and a new app oach. Renew Sus ain Ene gy Re 353
2018;91:1089–1102. 354
[7] Ismail MS, Mogha emi M, Mahlia TMI. Ene gy ends in Pales inian e i o ies o Wes Bank 355
and Gaza S ip: possibili ies o educing he eliance on ex e nal ene gy sou ces. Renew 356
Sus ain Ene gy Re 2013;28(0):117–29. 357
[8] Hoicka CE, Rowlands IH. Sola and wind esou ce complemen a i y: ad ancing op ions o 358
enewable elec ici y in eg a ion in On a io, Canada. Renew Ene gy 2011;36(1):97–107. 359
[9] Mekhile S, Fa ama zi SZ, Saidu R, Salam Z. The applica ion o sola echnologies o 360
sus ainable de elopmen o ag icul u al sec o . Renew Sus ain Ene gy Re 2013;18(0):583–94.] 361
[10] Ola omiwa L, Mekhile S, Ismail MS, Mogha emi M. Ene gy managemen s a egies in 362
hyb id enewable ene gy sys ems: A e iew. Renew Sus ain Ene gy Re 2016; 62:821–835. 363
[11] Gup a SK, Pu ohi P. Renewable ene gy ce i ica e mechanism in India: a p elimina y 364
assessmen . Renew Sus ain Ene gy Re 2013;22(0):380–92. 365
[12] Ve ga a-Die ich JD, e al. Ad anced chance-cons ained p edic i e con ol o he e icien 366
ene gy managemen o enewable powe sys ems, J P ocess Con ol (2017), 367
h ps://doi.o g/10.1016/j.jp ocon .2017.11.003] 368
[13] Bizon N. Op imal Ope a ion o Fuel Cell / Wind Tu bine Hyb id Powe Sys em unde 369
Tu bulen Wind and Va iable Load. Appl Ene gy 2018;212:196-209 370
h ps://doi.o g/10.1016/j.apene gy.2017.12.034 371
[14] C is ian N, Bizon N, Alexand u A. Design o hyb id powe sys ems using HOMER 372
simula o o di e en enewable ene gy sou ces. ECAI - 9 h Edi ion o In e na ional 373
Con e ence on Elec onics, Compu e s and A i icial In elligence. Ta go is e, ROMÂNIA, 29 374
June -01 July, 2017 10.1109/ECAI.2017.8166507 375
[15] Bizon N. Real- ime op imiza ion s a egy o uel cell hyb id powe sou ces wi h load-376
ollowing con ol o he uel o ai low. Ene gy Con e s Manage 2018;157:13–27; 377
h ps://doi.o g/10.1016/j.enconman.2017.11.084 378
[16] Bizon N, Maza e AG, Ionescu LM, Enescu FM. Op imiza ion o he P o on Exchange 379
Memb ane Fuel Cell Hyb id Powe Sys em o Residen ial Buildings. Ene gy Con e s Manage 380
2018;163:22–37; 381
[17] Chauhan A, Saini RP. A e iew on In eg a ed Renewable Ene gy Sys em based powe 382
gene a ion o s and-alone applica ions: con igu a ions, s o age op ions, sizing me hodologies 383
and con ol. Renew Sus ain Ene gy Re 2014;38:99–120. 384
[18] Bizon N, Radu M, Op oescu M. Ene gy con ol s a egies o he Fuel Cell Hyb id Powe 385
Sou ce unde unknown load p o ile. Ene gy 2015;86:31-41, 386
h p://dx.doi.o g/10.1016/j.ene gy.2015.03.118 387
[19] Bizon N. Load- ollowing Mode Con ol o a S andalone Renewable/Fuel Cell Hyb id Powe 388
Sou ce. Ene gy Con e s Manage 2014;77:763–772 389
h p://dx.doi.o g/10.1016/j.enconman.2013.10.035 390
[20] T i ko ic M, Sheikhzadeh M, Nigim K, Daou idis P. Modeling and con ol o a enewable 391
hyb id ene gy sys em wi h hyd ogen s o age. IEEE T ans Con ol Sys Technol 2014;22(1): 169 392
- 179. h ps://ieeexplo e.ieee.o g/documen /6477101/ 393
[21] Bizon N. Nonlinea con ol o uel cell hyb id powe sou ces: Pa II –Cu en con ol. Appl 394
Ene g 88(7) (2011) 2574–2591 h p://dx.doi.o g/10.1016/j.apene gy.2011.01.044 395
[22] Ha i M, Meha a A, Tiou si M. Powe managemen s a egy in he al e na i e ene gy 396
pho o ol aic/PEM uel cell hyb id sys em. Renew Sus ain Ene gy Re 2011;15(9):5104–10. 397
[23] Bizon N. T acking he maximum e iciency poin o he FC sys em based on ex emum 398
seeking scheme o con ol he ai low. Appl Ene g 2014;129:147–157 399
h p://dx.doi.o g/10.1016/j.apene gy.2014.05.002 400
[24] Uzunoglu M, Ona OC, Alam MS. Modeling, con ol and simula ion o a PV/FC/UC based 401
hyb id powe gene a ion sys em o s and-alone applica ions. Renew Ene gy 2009;34(3):509–402
20. 403
[25] Sikkabu S, Mungpo n P, Ekka a a odome C, Bizon N, e . al. Con ol o High-Ene gy High-404
Powe Densi ies S o age De ices by Li-ion Ba e y and Supe capaci o o Fuel 405
Cell/Pho o ol aic Hyb id Powe Plan o Au onomous Sys em Applica ions. IEEE T Ind Appl 406
2016 52(5):4395-4407, 10.1109/TIA.2016.2581138. 407
[26] Bizon N. Global Maximum Powe Poin T acking (GMPPT) o Pho o ol aic a ay using he 408
Ex emum Seeking Con ol (ESC): A e iew and a new GMPPT ESC scheme. Renew Sus ain 409
Ene gy Re 2016;57:524–539 doi:10.1016/j. se .2015.12.22 410
[27] Bizon N. Global Maximum Powe Poin T acking based on new Ex emum Seeking Con ol 411
scheme. P og Pho o ol aics: Resea ch and Applica ions 2016;24(5):600-22. DOI: 412
10.1002/pip.2700. 413
[28] Bizon N. Sea ching o he Ex eme Poin s on Pho o ol aic Pa e ns using a new 414
Asymp o ic Pe u bed Ex emum Seeking Con ol scheme. Ene gy Con e s Manage 415
2017;144:286–302; h ps://doi.o g/10.1016/j.enconman.2017.04.065 15 July 2017. 416
[29] Bizon N. Ene gy op imiza ion o Fuel Cell Sys em by using Global Ex emum Seeking 417
algo i hm. Appl Ene g 2017;206:458-474; h ps://doi.o g/10.1016/j.apene gy.2017.08.097 418
[30] Bizon N. Global Ex emum Seeking Con ol o he Powe Gene a ed by a 419
Pho o ol aic A ay unde Pa ially Shaded Condi ions. Ene gy Con e s Manage 2016;109; 420
71-85 (Feb ua y 2016), h p://dx.doi.o g/10.1016/j.enconman.2015.11.046 421
[31] Dagdougui H, Mincia di R, Ouammi A, Robba M, Sacile R. A dynamic decision model o 422
he eal- ime con ol o hyb id enewable ene gy p oduc ion sys ems. IEEE Sys J 423
2010;4(3):323–33. h p://dx.doi.o g/10.1109/JSYST.2010.2059150. 424
[32] Val e de L, Ali D, and Go do R. A echnical e alua ion o Wind-Hyd ogen (WH) 425
demons a ion p ojec s in Eu ope. In: P oceedings o powe enginee ing, ene gy and elec ical 426
d i es (POWERENG) 2013 10.1109/Powe Eng.2013.6635764 427
[33] Du o-López R, Be nal-Agus ín JL, Con e as J. Op imiza ion o con ol s a egies o s and-428
alone enewable ene gy sys ems wi h hyd ogen s o age. Renew Ene gy 2007;32(7):1102–26. 429
[34] Vasallo MJ, B a o JM, Andúja JM. Op imal sizing o UPS sys ems based on ba e ies 430
and/o uel cell. Appl Ene gy 2013;105:170–81. 431
[35] Segu a F, Du án E, Andúja JM. Design, building and es ing o a s andalone uel cell hyb id 432
sys em. J Powe Sou ces 2009;193(1):276–84. 433
[36] Semaoui S, Hadj A ab a, Bacha S, Azoui B. The new s a egy o ene gy managemen o a 434
pho o ol aic sys em wi hou ex a in ended o emo e-housing. Sol Ene gy 2013;94:71–85. 435
[37] Bizon N, Op oescu M, Raceanu M. E icien Ene gy Con ol S a egies o a S andalone 436
Renewable/Fuel Cell Hyb id Powe Sou ce. Ene gy Con e s Manage 2015;90:93-110 437
doi:10.1016/j.enconman.2014.11.002. 438
[38] Bizon N, Taba abaei NM, Blaabje g F, Ku E (Ed.), Ene gy Ha es ing and Ene gy 439
E iciency: Technology, Me hods and Applica ions, Sp inge Ve lag London Limi ed, 2017. 440
[39] Bizon N, Taba abaei NM, Shayeghi H (Ed.), Analysis, Con ol and Op imal Ope a ions in 441
Hyb id Powe Sys ems - Ad anced Techniques and Applica ions o Linea and Nonlinea 442
Sys ems, Sp inge UK, 2013. 443
[40] Bizon N, Culce M, Iliescu, Maza e AG, Lau en iu IM, Beloiu R. Real- ime s a egy o 444
op imize he Fuel Flow a e o Fuel Cell Hyb id Powe Sou ce unde a iable load cycle. ECAI 445
- 9 h Edi ion o In e na ional Con e ence on Elec onics, Compu e s and A i icial In elligence. 446
Ta go is e, ROMÂNIA, 2017 10.1109/ECAI.2017.8166513 447
[41] Bizon N, Culce M, Op oescu M, Iana G, Lau en iu IM, Maza e AG, Iliescu M. Real- ime 448
s a egy o op imize he Ai low a e o Fuel Cell Hyb id Powe Sou ce unde a iable load 449
cycle, The 22 d In e na ional Con e ence on Applied Elec onics - APPEL 2017, Pilsen, 2017 450
10.23919/AE.2017.8053577 451
[42] Mehne HH. E alua ion o Pa allelism in An Colony Op imiza ion Me hod o Nume ical 452
Solu ion o Op imal Con ol P oblems. Jou nal o Elec ical Enginee ing, Elec onics, Con ol 453
and Compu e Science - JEEECCS 2015 1(2):15-20; 454
h p://jeeeccs.ne /index.php/jou nal/a icle/ iew/17 455
[43] Hoa ca IC, Raducu M. On he mic o-in e e pe o mance based on h ee MPPT con olle s. 456
Jou nal o Elec ical Enginee ing, Elec onics, Con ol and Compu e Science - JEEECCS 2015 457
1(1):7-14; a ailable on h p://jeeeccs.ne /index.php/jou nal/a icle/ iew/9 458
[44] Puk ushpan JT, S e anopoulou AG, Peng H. Con ol o uel cell powe sys ems. New Yo k: 459
Sp inge ; 2004. 460
[45] Valenciaga F. and Pablo F. Pules on and Ped o E. Ba aio o. Supe iso Con ol o a 461
S and-Alone Hyb id Gene a ion Sys em Using Wind and Pho o ol aic Ene gy. IEEE T ans 462
Ene gy Con e s 2005;20(2):398-405. 463
[46] Mo az Ame M, Naamane A, M’Si di KN. Op imiza ion o Hyb id Renewable Ene gy 464
Sys ems (HRES) Using PSO o Cos Reduc ion. Ene gy P ocedia 2013;42:318-327. 465
[47] Chauhan A, Saini RP. A e iew on in eg a ed enewable ene gy sys em based powe 466
gene a ion o s and-alone applica ions: con igu a ions, s o age op ions, sizing me hodologies 467
and con ol. Renew Sus ain Ene gy Re 2014;38:99–120 468
[48] Yang S, Tan S, Xu J-X. Consensus Based App oach o Economic Dispa ch P oblem in a 469
Sma G id. IEEE T ansac ions on Powe Sys ems 2013;28(4):4416 – 4426. 470
[49] Bine i G, Naso D, Tu chiano B, Da oudi A, F ank L. Lewis.Consensus-based App oach o 471
he Economic Dispa ch P oblem. IFAC P oceedings Volumes 2014;47(3):3140-3145. 472
[50] He Y, Venka esh B, Guan L. Op imal scheduling o cha ging and discha ging o elec ic 473
ehicles. IEEE T ans Sma G id 2012;3(3):1095–1105. 474
[51] Guo Y, Xiong J, Xu S, Su W. Two-S age Economic Ope a ion o Mic og id-Like Elec ic 475
Vehicle Pa king Deck. IEEE T ansac ions on Sma G id 2016;7(3):1703 – 1712. 476
[52] Bendjedia B, Rizoug N, Boukhni e M, Boucha aa F. Hyb id Fuel Cell/Ba e y Sou ce 477
Sizing and Ene gy Managemen o Au omo i e Applica ions. IFAC Pape sOnLine 2017; 478
50(1):4745–475. 479
[53] Neh i MH, Wang C, S unz K, Aki H, Ramakuma R, Bing J, e al. A e iew o hyb id 480
enewable/al e na i e ene gy sys ems o elec ic powe gene a ion: con igu a ions, con ol, and 481
applica ions. IEEE T ans Sus ain Ene gy 2011;2(4):392–403. 482
[54] Bizon N. Ene gy e iciency o he mul ipo powe con e e s a chi ec u es o se ies and 483
pa allel hyb id powe sou ce ype used in plug-in/V2G uel cell ehicles. Appl Ene g 484
2013;102:726-734 h p://dx.doi.o g/10.1016/j.apene gy.2012.08.021 485
[55] Bizon N. Ene gy E iciency o Mul ipo Powe Con e e s used in Plug-In/V2G Fuel Cell 486
Vehicles, Appl Ene g 2012;96: 431-443 h p://dx.doi.o g/10.1016/j.apene gy.2012.02.075 487
[56] Mihaescu M. Applica ions o mul ipo con e e s. Jou nal o Elec ical Enginee ing, 488
Elec onics, Con ol and Compu e Science - JEEECCS 2016 2(1):13-18; a ailable on 489
h p://jeeeccs.ne /index.php/jou nal/a icle/ iew/25 490
[57] Taba abaei NM, Bizon N, Aghbolaghi AJ, Blaabje g F (Ed.), Fundamen als and 491
Con empo a y Issues o Reac i e Powe Con ol in AC Powe Sys ems, Sp inge Ve lag 492
London Limi ed, 2017. 493
[58] Chauhan A, Saini RP. A e iew on In eg a ed Renewable Ene gy Sys em based powe 494
gene a ion o s and-alone applica ions: con igu a ions, s o age op ions, sizing me hodologies 495
and con ol. Renew Sus ain Ene gy Re 2014;38:99–120. 496
[59] Bizon N, Ku E. Pe o mance Analysis o T acking o he Global Ex eme on Mul imodal 497
Pa e ns using he Asymp o ic Pe u bed Ex emum Seeking Con ol Scheme. In J Hyd ogen 498
Ene g 2017; 42(28):17645-54, h p://dx.doi.o g/10.1016/j.ijhydene.2016.11.173 499
[60] Bizon N, Thoun hong P, Raducu M, Cons an inescu LM. Designing and Modelling o he 500
Asymp o ic Pe u bed Ex emum Seeking Con ol Scheme o T acking he Global Ex eme. In 501
J Hyd ogen Ene g 2017; 42(28):17632-44, h p://dx.doi.o g/10.1016/j.ijhydene.2017.01.086 502
[61] opez- uede M, Ramos-He nanz A, a a aga M, Ga mendia A, Ionescu VM. S udy on 503
he in luence o Lambda pa ame e on se e al pe o mance indexes in Dynamic Ma ix Con ol. 504
Jou nal o Elec ical Enginee ing, Elec onics, Con ol and Compu e Science - JEEECCS 2015 505
2(1):1-8; a ailable on h ps://jeeeccs.ne /index.php/jou nal/a icle/ iew/15 506
[62] E dinç O, Vu al B, Uzunoglu M. A wa ele - uzzy logic based ene gy managemen s a egy 507
o a uel cell/ba e y/ul a-capaci o hyb id ehicula powe sys em, J Powe Sou ces 508
2009;194:369-80. 509
[63] Naci i M, Aggou M, Ahmed WA. Wind ene gy s o age by pumped hyd o s a ion / Pages: 510
32-42 Jou nal o Ene gy Sys ems (JES) 2017;1(1):32-42, 511
h p://de gipa k.go . /jes/issue/30882/329315 512
[64] Uğu lu A, ökçöl C. A case s udy o PV-wind-diesel-ba e y hyb id sys em. Jou nal o 513
Ene gy Sys ems (JES) 2017;1(4):138-147, h p://de gipa k.go . /jes/issue/33528/348335 514
[65] San os-He e o JM, Lopez-Guede JM, I an Flo es I. A Sho e iew on he use o enewable 515
ene gies and model p edic i e con ol in buildings. Jou nal o Ene gy Sys ems (JES) 2017;1(3) 516
112-119, h p://de gipa k.go . /jes/issue/32837/346653 517
[66] Ramos-Paja CA, Spagnuolo G, Pe one G, Emilio Mama elis M. A pe u ba ion s a egy o 518
uel consump ion minimiza ion in polyme elec oly e memb ane uel cells: Analysis, Design 519
and FPGA implemen a ion. Appl Ene g 2014;119:21-32. 520
521
522
Table 1. The opological di e ences be ween he RTO s a egies
No.
I e (Boos )
I e (Ai )
I e (Fuel)
RTO s a egy
Re
0
ILF
IFC
IFC
sFF
[44]
1
ILF
IGES+IFC
IFC
RTO1
[41]
2
ILF
IFC
IGES+IFC
RTO2
[40]
3
IGES
ILF
IFC
RTO3
[29]
4
IGES
IFC
ILF
RTO4
[15]
Table 2. The sFF s a egy applied o FCHPS a di e en Pload
Pload
PFCne 0
sys0
Fuele 0
FuelT0
[kW]
[W]
[%]
[W/lpm]
[l]
2
1942
93.26
137.3
34.02
3
2884
91.85
129.5
56.3
4
3786
90.43
121.6
74.88
5
4650
88.75
113.4
98.6
6
5467
86.89
104.7
125.58
7
6229
84.78
95.16
158.34
8
6912
82.3
83.75
176
Table 3. The RTO1 s a egy applied o FCHPS a di e en Pload
Pload
PFCne 1
sys1
Fuele 1
FuelT1
[kW]
[W]
[%]
[W/lpm]
[l]
2
2009
93.53
136.3
35.24
3
2960
92.27
128.8
56.43
4
3868
90.96
122.3
74.75
5
4751
89.36
115.2
98.22
6
5595
87.58
108.2
124.2
7
6504
85.69
101.37
154
8
7240
84.95
94.1
164.2
Table 4. The gaps in FC ene gy e iciency, uel e iciency, and uel economy o he RTO1
s a egy compa ed o he sFF s a egy
Pload
sys
Fuele
FuelT
[kW]
[%]
[W/lpm]
[l]
2
0.27
-1
1.22
3
0.42
-0.7
0.13
4
0.53
0.7
-0.13
5
0.61
1.8
-0.38
6
0.69
3.5
-1.38
7
0.91
6.21
-4.34
8
2.65
10.35
-11.8
Table(s)
Figu e 13. The gap in o al uel consump ion o Pload cons an and di e en k uel
Figu e 14. The gap in o al uel consump ion o Pload a iable and di e en k uel
Figu e 15. The beha io o RTO1 s a egy wi h PRES=0, Pload(AV)=4kW, and kne =0.5 W-1 and
k uel=50 W-1 lpm
Figu e 16. The beha io o gaps in pe o mance indica o s o he RTO1 s a egy compa ed o
sFF s a egy
Figu e 17. The beha io o he RTO1-based HPS wi h PRES a iable, Pload=6kW, and kne =0.5 W-1
and k uel=50 W-1 lpm
Figu e 18. The beha io o RTO1-based HPS wi h a iable p o iles o PRES and Pload, and
kne =0.5 W-1 and k uel=50 W-1 lpm
