Design and Evaluation of a Renewable Water Pumping System

PREPRINT Design and E alua ion o a Renewable Wa e Pumping Sys em Imene Yahyaoui Dep . Sys ems Eng and Au oma ic Indus ial Eng School Uni . Valladolid, Spain imene@au om.u a.es Giuseppe Tina Dep . Elec ic, Elec onic and Compu e Uni . Ca ania, I aly Giuseppe. [email protected] c .i Mahe Chaabene Dep . Elec ic Eng Na ional Eng School Uni . S ax, Tunisia mahe chaabane@yahoo. com Fe nando Tadeo Dep . Sys ems Eng and Au oma ic Indus ial Eng School Uni . Valladolid, Spain e nando@au om.u a.es Abs ac : The design and e alua ion o a s and-alone pho o ol aic sys em o well pumping in ag icul u al applica ion is p esen ed. Gi en en i onmen al (i adiance and ambien empe a u e), sys em (PV and ba e y echnology, a ay geome y) and load (mon hly daily demand) da a, he op imal size o main componen s o PV sys ems a e ob ained by a sizing algo i hm; speci ically he ou pu a e: he su ace o PV a ay and he ba e y pack capaci y. The decision is made based on he es ima ed powe gene a ion, he equi ed powe o he load, he wa e needed by he c ops and he ba e y equi emen s. The p oposed design is hen e alua ed using yea ly simula ions, on hou ly base, pe o med by a specialized comme cial so wa e, named PVSys , o show ha he p oposed op imal size ensu es also a high eliabili y e alua ed by wo indices: numbe o au onomous days (NAD) and Loss o Load P obabili y (LOLP). Keywo ds: pho o ol aic panels; algo i hm; sizing; pumping; PVSys . INTRODUCTION The main componen s o a s and-alone pho o ol aic (PV) sys em, ha supplies a gi en load a e: a PV a ay, a ba e y pack, an MPPT/ cha ge egula o . The MPPT/cha ge egula o mus be used o ope a e co ec ly wi h bo h PV modules and ba e ies. This appliance pe o ms no only he cha ge and discha ge o he ba e ies in o de o a oid damage o poo ene gy pe o mance bu also he maximum powe poin acking (MPPT) in such a way o ha e an e icien PV con e sion. Sizing he componen s o hese PV ins alla ions a ec hei au onomy and cos [1, 2]. Hence, i is necessa y o ix du ing he design adequa e alues o he componen s sizes, such as he pho o ol aic panel su ace and he ba e y capaci y [3, 4]. In ac , o ag icul u al applica ions, du ing he c ops ege a i e cycle, he pho o ol aic ins alla ion size selec ed mus gua an ee he wa e olume needed o he c ops i iga ion, he sys em au onomy and he ba e y bank sa e ope a ing [3]. Indeed, knowing he wa e olume needed o i iga ing he c ops, he si e cha ac e is ics, he sola adia ion and he pho o ol aic panel ype, sizing aims o p o ide he adequa e alues o he panel su ace, ba e y capaci y and (in some ins ances) he ese oi olume. In his sense, esea che s ha e es ablished a ious me hods o op imize pho o ol aic ins alla ions componen s [5]. Fo ins ance, some wo ks ha e ocused on de eloping analy ic me hods based on a simple calcula ion o he panels su ace and ba e y bank capaci y using he ene ge ic balance [6-8]. O he wo ks ha e concen a ed on he cos e sus eliabili y ques ion [9]. Mo eo e , some esea che s ha e p oposed sizing algo i hms based on he minimiza ion o cos unc ions, using he Loss o Load P obabili y (LLP) concep [10-14]. This LLP app oach has also been combined wi h a i icial neu onal ne wo ks and gene ic algo i hms [9, 10]. Howe e , hese me hods may esul in an o e sized sys em o one loca ion and an unde sized one o ano he loca ion [15]. The o e sized case esul s in high ins alla ion cos s. Whe eas, an unde sized case, he ins alla ion is unable o supply he load wi h he ene gy needed [16, 17]. Mo eo e , he ins alla ion li e ime is sho e , due o he excessi e use o he ba e ies. Fo hese easons, he sizes mus be ca e ully selec ed o each speci ic applica ion and loca ion [15]. In [18], he de e minis ic me hod and he p obabilis ic app oach a e used o analyze he impac on design he PV sys em and he ba e y s o age o h ee geog aphic si es in I aly, cha ac e ized by di e en alues o sun adia ion and ambien empe a u e. To ind he bes comp ise be ween wo con lic ing obiecje i es ( eliabili y and cos s) a uzzy logic based mul i-objec i e op imiza ion app oach was used in [19]. This pape p esen s a con inua ion o p e ious published wo ks by some he au ho s [16, 17], whe e an a sizing algo i hm has been p esen ed. He e, he algo i hm is de ailed and alida ed by means o hou ly p obabilis ic simula ions, ha spans one yea , using a widely used gene al so wa e o he s udy o PV sys ems (G id-connec ed, s and-alone o hyb id), named PVSys [20] (Figu e 1). I.SIZING ALGORITHM PRINCIPLE A good sizing mus ul ill he elec ical demand o he load [15]. Hence, he main objec i e is o ensu e he load supply h oughou he day, while cha ging he ba e y wi h he excees o he ene gy and gua an eeing he wa e olume needed o he i iga ion. The scheme o he p oposed app oach is p esen ed in Figu e 1 [16, 17]. The algo i hm depends on: PREPRINT • he wa e olume needed, • he si e cha ac e is ics, • he panel cha ac e is ics. The algo i hm aims o ind he panels su ace op S and he ba e y capaci y op ba C ha gua an ee he ins alla ion au onomy when supplying he pump. Hence, he idea consis s in sea ching he op imal componen s sizes ha ensu e he balance be ween he cha ged and he ex ac ed ene gies c E and e E , espec i ely. In ac , he ba e y bank supply he load by ba E when he panel does no gene a e he su icien ene gy load E , and is cha ged wi h a pa o he PV ene gy p oduced p E (Figu e 2). The balance be ween he accumula ed and he ex ac ed ene gies does no gua an ee he au onomy, due o he luc ua ion in he sola adia ion and he ene gy losses. Thus, o ensu e he au onomy and p o ide he ene gy demanded by he load, he algo i hm is pe o med by adop ing an e iciency coe icien η (sligh ly g ea e han 1). Hence, he ene gy balance can be exp essed as ollows: ce EE η ≈ (1) The sizing algo i hm is pe o med using wo sub algo i hms du ing he ege a i e cycle: he Algo i hm 1 de e mines he sizes o he panel su ace M S and he ba e y capaci y M ba C o each mon h M. Then, Algo i hm 2 is pe o med o deduce he inal sizes based on he sizes de e mined o each mon h and he a ailable componen s, p o iding he numbe s o panels and ba e ies needed. Algo i hm 1 is de ailed now ollowing Figu e 4. a) Algo i hm 1: De e mina ion a each mon h o he minimum panel su ace and ba e y bank capaci y S ep 1 Es ima ion o he di used and di ec adia ion. S ep 2 Deduc ion o he hou ly daily sola adia ion dis ibu ion ( ) dhH , in a il ed panel [17]. S ep 3 Es ima ion o he hou ly cell empe a u e ( ) h T c [16]. S ep 4 Deduc ion o he hou ly panel e iciency ( ) h p η [17]. Panel cha ac e is ics Inpu s Ou pu s Cons ain s Objec i e Panels su ace Wa e needed SIZING ALGORITHM Ba e y bank capaci y P o ec he ba e ies agains deep discha ge Ensu e he ene gy balance Gua an ee he wa e needed ( ) AM PM Ec E E η +  max 0.78dod∆≤ ese oi VV= op S op ba n Si e cha ac e is ics Fig. 1 Planning o he p oposed sizing algo i hm S ep 5 Calcula ion o he wa e needed V [16, 17]: ( ) ( ) ( ) 11 11 iR c To m iR L V kE L  −− = −+   −  (2) whe e: c k : c op g ow h coe icien o mon h M, To E : e e ence e apo anspi a ion a e age o mon h M, m : a e age ain olume o mon h M, i : leaching e iciency, R L : leaching ac ion gi en in he soil. S ep 6 Calcula ion o he pumping du a ion [17]: V Q ∆= (3) whe e Q is he wa e low (m²/h). S ep 7 Calcula ion o he minimum panel su ace i S and he ini ial ba e y capaci y i ba C using equa ions (4) and (5) espec i ely: 21 pump au i ech p ba l p eg in op he ma ching P d Sd W η ηη η η η η ∆ = +   (4) max dodV dE C ba au d ba i∆ = (5) wi h: pump P : pump powe (W), au d : eques ed days o au onomy, ech d : days needed o echa ge he ba e y, p W : a e age daily adia ion (Wh/ 2 m / day), ba η : elec ical e iciency o he ba e y bank, l η : elec ical e iciency o he es o he ins alla ion (includes ohmic wi ing and misma ching wi ing losses), p η : e iciency o each pho o ol aic panel, eg η : egula o pe o mance, in η : in e e pe o mance, op he η : panel pe o mance - op ical and he mal e ec s (%), ma ching η : panel ma ching pe o mance (%), d E : daily ene gy consump ion (Wh), ba V : ba e y ol age (V), max dod ∆ : maximum pe mi ed a ia ion o he dep h o discha ge dod . S ep 8 Calcula ion o p i P co esponding o he minimum panel su ace i S , using (6) [17]: p i p i P SH η = (6) ce EE η ≈ ba C Cha ge he ene gy demanded by he load PREPRINT Pho o ol aic panels Load p E load E load E Ba e y Fig. 2 Ene gy balance p inciple S ep 9 Calcula ion o he ene gies expec ed o be daily s o ed and ex ac ed om he ba e y c E and e E . S ep 10 I he ex ac ed ene gy is highe han he s o ed ene gy, he algo i hm inc eases he panel su ace by he minimum inc emen o he PVP size comme cially a ailable: he algo i hm looks o he bes con igu a ion o gua an ee he balance be ween he demanded and he p oduced ene gies, by equalizing he ene gies s o ed c E and ex ac ed ene gies e E in he ba e y bank (1). S ep 11 Ba e y capaci y M ba C deduc ion o mon h M [17]: ba c ba V E C M = (7) b) Algo i hm 2: Calcula ion o he minimum panel su ace and ba e y bank capaci y o he whole ege a i e cycle Using Algo i hm 2 (Figu e 3), he inal alues o he panel su ace op S and he ba e y bank capaci y op ba C , a e deduced. op S co esponds o he maximum alue o he panel su ace ob ained du ing he mon hs. The inal ba e y capaci y is he co esponding alue o op S , since i is he mos c i ical. II.APPLICATION TO A CASE STUDY The p oposed algo i hm is applied now o e alua e he componen s sizes o a case s udy: he p oposed algo i hm is es ed du ing he mon hs ha co espond o he ege a i e cycle o oma oes (Ma ch o July), using da a o he a ge a ea (No he n o Tunisia: la i ude: 36.39°, longi ude: 9.6°). Following Algo i hm 2, he Algo i hm 1 was i s e alua ed o all he mon hs in he ege a i e cycle: he sola adia ion accumula ed on a il ed panel is e alua ed; hen, he panel yield is calcula ed o each mon h; in pa allel, he wa e needed V is e alua ed depending on he ege a i e cycle and he si e [17]. Then, i he s o ed ene gy is highe han he ex ac ed ene gy, he su ace is inc eased by he minimum su ace in he ma ke (in ou case, he inc emen is 0.5 2 m ). Algo i hm 1 esul s a e summa ized in Table 1, which shows ha he p oposed s a egy always ensu es he wa e needs, espec s he limi s on he ba e y-bank’ dep h o discha ge and he ene gy balance (1). This has been es ed du ing he mon hs o oma oes ege a i e cycle: he e iciency coe icien η is a ound he ixed alues h oughou all he conside ed mon hs. Fo his alue, max dod∆ is gua an eed o be equal o 0.78. Fo ins ance, in July, he minimum η is 1.46, and he alue ob ained wi h Algo i hm 1 1 η is equal o 1.47. On he o he hand, in Ma ch, he gene a ed pho o ol aic powe du ing he mo ning supplies he pump oge he wi h he ba e y bank du ing he pumping du a ion. A e ha , he pho o ol aic powe gene a ed cha ges he ba e y bank. The quo ien be ween he cumula ed and ex ac ed ene gies is 1.66, which is nea o a ge alue 1.7. We mus poin ou ha o he ene gy balance, an e o coe icien is used o conside he clouds. Hence, in ou s udy, we ake in o accoun he possibili y o ha ing cloudy days. Fo example, in Ap il he loss o ene gy each day is 23.23 %. The ob ained esul s (Table 1) p o e ha he panels su ace and ba e y bank capaci y ob ained using he p oposed Algo i hm 1 sa is y he ene gy balance. This is possible hanks o he calcula ion o he ba e y capaci y, which is done by conside ing he same max dod∆ alue ha can be eached. Since July is he mos c i ical mon h o i iga ion, he sys em componen s sizing o July is selec ed. The ob ained size allows he load o be supplied du ing he eques ed pumping du a ion ∆ , and also p o ides he ene gy c E needed o cha ge he ba e y bank. III.VALIDATION USING PVSYST The ins alla ion size has been also es ed using PVSys , since he sola adia ion, he ambien empe a u e and he load equi emen s o he a ge ci e can be manually choosed. This ool allows de e mine and alida e ins alla ions componen s sizes. In addi ion, i akes in o accoun a ios losses ela ed wi h componen s o clima ic pa ame e s. Hence, PVSys e alua e he size e icicency using he sola ac ion (SF), which de e mines whe he he panel su ace is able o supply he load wi h he needed ene gy. Mo eo e , i pe o ms a mo e de ailed e alua ion o he ins alla ion size: he sys em losses (Ls), he unused ene gy (Lu) and he ene gy supplied o he use (Y ). The PVSys simula ion shows ha he adop ed size (S= 101.5 2 m and ba C =1680 A.h) gi es good esul s. In ac , Figu es 5 and 6 show ha du ing he c ops ege a i e cycle, he sola ac ion (SF), which de e mine whe he he sola ins alla ion p o ides he load wi h he su icien ene gy, is p a ically equal o one, excep in June and July, in which i is equal o 0.962 and 0.934, espec i ely (Table 2). This leak o ene gy can be co e ed by conside ing an addi ional wa e olume in he ese oi . Fig. 3 Sizing Algo i hm 2 Fo M=1: max M , e alua e M S and M ba C using Algo i hm 1 . { } Mop ba ba op SSwhenCC SSSS Mop == =max21 ,..,,max M ba M CS , ba E PREPRINT Mo eo e , he ob ained size is es ed du ing he yea (Figu e 7), by aking in o accoun all he possible losses ela ed o he componen s o clima ic pa ame e s. The esul s show ha he choosen size allows supplying he load and ha ing no mo e han 3.4 % o load losses, which ep esen s a good esul . Fig. 4 Sizing Algo i hm 1 o each mon h M Fig. 5 No malized p oduc ion using PVSys Fig. 6 Pe o mance Ra ion and sola ac ion using PVSys PR: Pe o mance Ra io (Y /Y ):0.211 1.2 1.0 Pe o mance Ra io PR Jan Feb Ma Ap May Jun Jul Aug Sep Oc No Dec 0.8 0.6 0.4 0.2 0 SF: Sola F ac ion (ESol/ELoad):0.966 Jan Feb Ma Ap May Jun Jul Aug Sep Oc No Dec Lu: Unused ene gy ( ull ba e y) Lc: Collec ion loss (PV-a ay losses) Ls: sys em losses and ba e y cha ging Y : Ene gy supplied o he use 1.91 kWh/kWp/day 1.07 kWh/kWp/day 1.34 kWh/kWp/day 1.15 kWh/kWp/day 10 8 6 4 2 0 No m alized ene gy [kWh/kWp/day] Si e cha ac e is ics S ep 1: Es ima ion o he di used and he di ec adia ions d H and b H [17]. S ep 2: Deduc ion o he sola adia ion in a il ed panel H [16, 17]. S ep 3: Es ima ion o he empe a u e c T [17]. S ep 4: Deduc ion o he panel’ e icieny p η S ep 5: Calcula ion o he needed wa e olume V [16, 17]. S ep 6: Calcula ion o he pumping du a ion ∆ [16, 17]. p η H S ep 7: Calcula ion o he ini ial panel’ su ace and he ini ial ba e y bank capaci y i ba C and ini ialize . S ep 8: Calcula ion o he pho o ol aic powe p i P co esponding o he ini ial su ace [17]. i S i SS= V ∆ p i P S ep 9: Calcula ion o he ene gies cha ged and ex ac ed om he ba e y bank and e E . c E c E e E S ep 10: Yes M SS= S ep 11 : Deduc ion o M ba C (7) No Dec emen S Yes Inc emen S No and e cEE η ≈ ec E E η > PREPRINT Fig. 7 Loss diag am o e he whole yea IV.CONCLUSION A sizing algo i hm o decide on he sizing o he ins alla ion elemen s was p esen ed and alida ed using PVSys ool. The algo i hm is es ed o a 10 ha land su ace in he no he n o Tunisia. The sizing esul s ensu es supplying he pump du ing he pumping pe iod, he ene gy needed by he load and he needed wa e olume o c ops i iga ion. ACKNOWLEDGEMENTS This wo k was unded by Minis e io de Ciencia e Inno ación (Spain) unde g an DPI2014-54530-R and FEDER unds.Miss Yahyaoui is unded by a g an MICInn BES-2011-047807. V. REFRENCES [1] Kaldellis, J. K; Za i akis, D; & Kondili, E. (2010). “Op imum sizing o pho o ol aic-ene gy s o age sys ems o au onomous small islands”. In e na ional Jou nal o Elec ical Powe & Ene gy Sys ems, 32(1), 24-36. [2] Sid ach-de-Ca dona, M; & Mo a López, Llanos. (1998). “A simple model o sizing s and-alone pho o ol aic sys ems”.Sola Ene gy Ma e ials and Sola Cells,55(3), 199-214. [3] Kha ib, Tame ; Mohamed, Zah.A; & Sopian, K. (2013). “A e iew o pho o ol aic sys ems size op imiza ion echniques”. Renewable and Sus ainable Ene gy Re iews, 22, 454-465. [4] Jakh ani, Abdul Qayoom; O hman, Al-Khalid; Rigi , And ew; Ragai. Hen y; Samo, Saleem. Raza; & Kamboh, Shakeel. Ahmed. (2012). “A no el analy ical model o op imal sizing o s andalone pho o ol aic sys ems”. Ene gy, 46(1), 675-682. [5] Acakpo i, Ame i; Xa ie , Fi a in. F ancois; & Awuah-Ba ou , Robe . (2012). “Analy ical me hod o sizing pho o ol aic wa e pumping sys em”. In he p oceedings o he 4 h IEEE In e na ional Con e ence on Adap i e Science & Technology, 65-69. [6] Sh es ha, G. B; & Goel, L. (1998). “A s udy on op imal sizing o s and-alone pho o ol aic s a ions”. IEEE T ansac ions on Ene gy Con e sion, 13(4), 373-378. [7] Ba a, Luciano; Ca alano i, Se gio; Fon ana, F; & La o an e, F. (1984). “An analy ical me hod o de e mine he op imal size o a pho o ol aic plan ”. Sola Ene gy, 33(6), 509-514. [8] G oumpos, P. P; & Papageo giou, G. (1987). “An op imal sizing me hod o s and-alone pho o ol aic powe sys ems”. Sola Ene gy, 38(5), 341-351. [9] Melli , A; Benghanem, M; Hadj A ab, A; & Guessoum, A. (2003). “Modelling o sizing he pho o ol aic sys em pa ame e s using a i icial neu al ne wo k”. In he p oceedings o he IEEE Con e ence on Con ol Applica ions, 353-357. [10] Yang, Hongxing; Zhou, Wei; Lu, Lin; & Fang, Zhaohong. (2008). “Op imal sizing me hod o s and-alone hyb id sola –wind sys em wi h LPSP echnology by using gene ic algo i hm”. Sola Ene gy, 82(4), 354-367. [11] Kha ib, Tame ; Mohamed, Azah; Sopian, K; & Mahmoud, M. (2012). “A new app oach o op imal sizing o s andalone pho o ol aic sys ems”. In e na ional Jou nal o Pho o Ene gy. [12] Klein, S. A; & Beckman, W. A. (1987). “Loss-o -load p obabili ies o s and-alone pho o ol aic sys ems”. Sola Ene gy, 39(6), 499-512. Ho izon al global i adia ion Global inciden in coll. plane IAM ac o on global E ec i e i adiancy on cells PV con e sion A ay nominal ene gy (a STC e ic) PV loss due o i adiance le el PV loss due o empe a u e Module quali y loss Module a ay misma ch loss Ohmic wi ing loss Loss by espec o he MPP unning Unused ene gy a he ou pu o he a ay E ec i e ene gy a he ou pu o he a ay Con e e loss du ing ope a ion (e iciency) Con e e loss due o powe h eshold Con e e loss o e nominal con . ol age Con e e loss due o ol age h eshold Ba e y s o ed ene gy balance Ba e y e iciency loss Ba e y e iciency loss Ba e y sel -discha ge cu en Ene gy need o he use Ene gy supplied o he use 1808 kWh/m² 1939 kWh/m² * 81 m² Cells -2.9 % +10.5 % E iciency a STC= 15.80 % -3.5 % -10.9 % -1.2 % -1.1 % -0.6 % 0.0 % -42.7 % 11.52 MWh -5.5 % 0.0 % 0.0 % 0.0 % +0.2 % -35.7 % -23.8 % 0.0 % 5.51 MWh 10.89 MWh S o ed 70.6% Di ec use 29.4% Missing ene gy 3.4 % 0.2 MWh 24.68 MWh 5.33 MWh Loss diag am o e he whole yea PREPRINT [13] Abouzah , I; & Ramakuma , R. (1991). “Loss o powe supply p obabili y o s and-alone pho o ol aic sys ems: a closed o m solu ion app oach”. IEEE T ansac ions on Ene gy Con e sion, 6(1), 1-11. [14] Magh aby, H. A. M; Shwehdi, M. H; & Al-Bassam, G. K. (2002). “P obabilis ic assessmen o pho o ol aic (PV) gene a ion sys ems”. IEEE T ansac ions on Powe Sys ems, 17(1), 205-208. [15] Melli , A; Benghanem, M; & Kalogi ou, S. A. (2007). “Modeling and simula ion o a s and-alone pho o ol aic sys em using an adap i e a i icial neu al ne wo k: P oposi ion o a new sizing p ocedu e”. Renewable Ene gy, 32(2), 285-313. [16] Yahyaoui, Imene; Chaabene, Mahe ;& Tadeo, Fe nando.(2013). “An algo i hm o sizing pho o ol aic pumping sys ems o oma oes i iga ion”. In he p oceedings o he IEEE In e na ional Con e ence on Renewable Ene gy Resea ch and Applica ions (ICRERA), 1089- 1095. [17] Yahyaoui, Imene; Ammous, Mahmoud; & Tadeo, Fe nando. (2015). “Algo i hm o op imum sizing o a pho o ol aic wa e pumping sys em”.In . Jou nal o Compu e Applica ions (IJCA), 11(6), 21-28. [18] Capizzi, Giacomo.; Bonanno, F ancesco; & Tina, Giuseppe. Ma co. “Expe iences on he Design o S and-Alone Pho o ol aic Sys em by De e minis ic and P obabilis ic Me hods”. P oceedings o he In e na ional Con e ence on Clean Elec ical Powe (ICCEP), 328– 335. [19] Giuseppe. Tina; Con i, S e ania; & Ragusa, Ca lo (2002). “Op imal sizing p ocedu e o s and-alone pho o ol aic sys ems by Fuzzy Logic” Jou nal o Sola Ene gy Enginee ing, 124, 77-82. [20] PVsys so wa e. CUEPE, Uni e si y o Gene a. www.p sys .com. Table 1 Panel su ace and ba e y capaci y o each mon h M Table 2 Ene gy balance and main PVSys esul s Resul s Mon hs GlobHo (kWh/ 2 m ) GlobE (kWh/ 2 m ) E A ail (MWh) E Unused (MWh) E Miss (MWh) E Use (MWh) E Load (MWh) Sol F ac Janua y 78.0 117.8 1.221 0.630 0.000 0.000 0.000 1.000 Feb ua y 89.1 116.0 1.185 0.608 0.000 0.000 0.000 1.000 Ma ch 140.0 161.3 1.587 0.959 0.000 0.419 0.419 1.000 Ap il 164.1 161.8 1.708 1.089 0.000 0.540 0.540 1.000 May 208.1 183.6 2.055 0.855 0.000 1.116 1.116 1.000 June 225.0 187.0 1.966 0.466 0.056 1.429 1.485 0.962 July 237.0 201.8 1.973 0.099 0.129 1.824 1.953 0.934 Augus 208.0 196.7 1.891 0.968 0.000 0.000 0.000 1.000 Sep embe 166.0 184.2 1.767 0.905 0.000 0.000 0.000 1.000 Oc obe 128.0 166.7 1.661 0.851 0.000 0.000 0.000 1.000 No embe 89.9 138.6 1.412 0.724 0.000 0.000 0.000 1.000 Decembe 75.1 123.8 1.266 0.649 0.000 0.000 0.000 1.000 Yea 1808.3 1939.5 19.692 8.801 0.185 5.328 5.513 0.966 Ma ch Ap il May June July Cloud co e age c A (%) 30.15 23.23 28.38 13.03 14.11 e o η 1.30 1.23 1.28 1.13 1.14 ( ) 2 M Sm 37.5 41.5 54.5 61.5 101.5 M ba C (Ah) 840 1050 840 1050 1680 1c e AM ePM E EE η =+ 1.66 1.57 1.64 1.44 1.46 Resul s Mon hs