Direct normal irradiance predictions using broadband models for Indian stations

Rohilla, Laxmi Na ain; Singh, Dinesh Kuma A icle Di ec no mal i adiance p edic ions using b oadband models o Indian s a ions Ene gy Repo s P o ided in Coope a ion wi h: Else ie Sugges ed Ci a ion: Rohilla, Laxmi Na ain; Singh, Dinesh Kuma (2020) : Di ec no mal i adiance p edic ions using b oadband models o Indian s a ions, Ene gy Repo s, ISSN 2352-4847, Else ie , Ams e dam, Vol. 6, Iss. 2, pp. 572-576, h ps://doi.o g/10.1016/j.egy .2019.11.121 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/243934 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/ A ailable online a www.sciencedi ec .com ScienceDi ec Ene gy Repo s 6 (2020) 572–576 www.else ie .com/loca e/egy The 6 h In e na ional Con e ence on Powe and Ene gy Sys ems Enginee ing (CPESE 2019), 20–23 Sep embe 2019, Okinawa, Japan Di ec no mal i adiance p edic ions using b oadband models o Indian s a ions Laxmi Na ain Rohilla∗, Dinesh Kuma Singh Ne aji Subhas Uni e si y o Technology, Sec o -3, Dwa ka, New Delhi 110078, India Recei ed 10 Oc obe 2019; accep ed 23 No embe 2019 A ailable online 14 Decembe 2019 Abs ac In his pape h ee pa ame ic models CPCR2, MLWT2 and REST ha e been used o compu e di ec no mal i adiance (DNI) o i e Indian s a ions, namely New Delhi, Pune, Jaipu , Kolka a and Mumbai. Compu ed alues o DNI ha e been compa ed wi h measu ed alues in e ms o pe cen age oo mean squa e e o (RMSE) and pe cen age mean bias e o (MBE). I is obse ed ha he a e age pe cen age RMSE o he yea is he minimum o MLWT2 model o Jaipu ollowed by New Delhi and Pune and he alues a e 1.67%, 2.33% and 2.49%, espec i ely. Bu o coas al s a ions Kolka a and Mumbai he co esponding minimum alues occu o CPCR2 model and REST model espec i ely, and he alues a e 2.62% and 2.76%, espec i ely. I shows ha MLWT2 model pe o ms be e o mos o he Indian s a ions excep he coas al s a ions. c 2019 Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). Pee - e iew unde esponsibili y o he scien i ic commi ee o he 6 h In e na ional Con e ence on Powe and Ene gy Sys ems Enginee ing (CPESE 2019). Keywo ds: Di ec no mal i adiance; Roo mean squa e e o ; Mean bias e o 1. In oduc ion Se e al au ho s including [1] has explained in hei wo k ha he sola ene gy sys ems a e mo e ene gy e icien unde cloudless condi ions. Unde such condi ions, mos o he sola adia ion is in he o m o di ec adia ion. Many de ices, like sola concen a o s, use only he di ec componen o he sola adia ion. Ne wo k o di ec i adiance measu ing s a ions is a he sca ce h oughou he wo ld and he same is ue o India as well. One way o knowing di ec adia ion a no mal incidence, also called di ec no mal i adiance (DNI), is o measu e global and di use adia ion on ho izon al and hen con e hese alues in o DNI, by using zeni h angle. Ano he way is o use a py heliome e and measu e DNI di ec ly. In India, DNI is measu ed by using py heliome e only a ou synop ic hou s: 9:30, 11:30, 13:30 and 15:30, and his oo a ew s a ions only. The e is oo much pauci y o DNI da a in India and his necessi a es he de elopmen o heo e ical models o p edic DNI unde di e en a mosphe ic condi ions. In he p esen wo k as explained in he ea lie wo k by Na ain [1], i has been ∗Co esponding au ho . E-mail add ess: [email p o ec ed] (L.N. Rohilla). h ps://doi.o g/10.1016/j.egy .2019.11.121 2352-4847/ c 2019 Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/ licenses/by-nc-nd/4.0/). Pee - e iew unde esponsibili y o he scien i ic commi ee o he 6 h In e na ional Con e ence on Powe and Ene gy Sys ems Enginee ing (CPESE 2019). L.N. Rohilla and D.K. Singh / Ene gy Repo s 6 (2020) 572–576 573 assumed ha he e is no cloud co e and p edic ions a e made unde di e en amoun o wa e apo , ozone con en and humidi y e c. Many au ho s ha e wo ked along hese lines. Sola adia ion models o p edic ing he a e age daily and hou ly global adia ion, beam adia ion and di use adia ion ha e been e iewed by Powe [2]. Pa ame e iza ion Model C has been discussed by Iqbal [3]. King and Buckius [4] ha e explained di ec sola ansmi ance o a clea sky, ha accu a ely p edic s he a ia ion wi h undamen al quan i ies. Di ec i adiance model, has been explained by Bi d and Huls om [5]. The pa ame e ized model o global adia ion unde pa ially cloudy skies has been explained by Choudha y [6]. As explained by Ko i e al. [7] ha he di ec sola i adiance can also be accu a ely calcula ed om global and di use ho izon al i adiance measu emen s a e co ec ing he di use i adiance alues. In he s udy o Munee e al. [8], Me eo ological Radia ion Model (MRM) enables compu a ion o ho izon al beam and di use sola adia ion using basic me eo ological pa ame e s — hou ly d y and we bulb empe a u e, a mosphe ic p essu e and b igh sunshine du a ion. Eigh een clea sky b oadband models o es ima ing he DNI in Alge ia ha e been in es iga ed by Benkaciali e al. [9]. Sola adia ion and illuminance was es ima ed om he me eo ological pa ame e s by Munee e al. [10]. Powe [2] employed he mon hly-a e aged clima e u bidi y da a o No h Ame ica and Eu ope o e alua e he ela ionships among mon hly-a e aged obse ed beam adia ion, clea -sky beam i adiance, sunshine du a ion and, day-leng h. A hyb id model was designed by Yang e al. [11] o es ima ing mon hly mean daily global adia ion by employing hou ly- eco ded b igh sunshine ime in Japan. Gueyma d [12,13,14] has explained abou he p edic ion o di ec sola ansmi ance and i adiance by using a ious b oadband models and concluded ha ou models, i.e. mul i-laye -weigh ed ansmi ance model, e sion 2 (MLWT2), Yang’s model, CPCR2 and e e ence e alua ion o sola ansmi ance (REST) a e showing good ag eemen wi h measu ed alues. In he p esen s udy, h ee models, CPCR2, MLWT2 and REST as p oposed by Gueyma d [12,13,14] ha e been selec ed o compu e DNI a di e en hou s and di e en mon hs o he yea , o Indian opical condi ions. 2. Models desc ip ion 2.1. CPCR2 model This is he wo bands model. The limi s o he model ha e been chosen as 0.29 µm and 2.7µm. These alues app oxima ely co espond o he a e age sensi i i y limi s o wo widely used py anome e s, he Eppley PSP (0.285– 2.8µm) and he Kipp & Zonen CM6 (0.3–2.5µm). The cu o wa eleng h be ween he UV/ isible band, B1, and he in a ed (IR) band, B2, has been se o 0.7µm. This model has been ound o be he bes model in ea lie s udies made by Gueyma d [15] and Ba le e al. [16]. I is comple ely desc ibed by Gueyma d [17]. E en hough his model depends on Angs om’s wa e exponen α, i has been ixed he e a 1.3. 2.2. MLWT2 model This model is modi ied e sion o MLWT1, which was o iginally de eloped as bo h a adia ion and u bidi y p edic ion me hod by Gueyma d [12,13,18]. MLWT2 model was based on he ela i ely new concep o mul ilaye spec al weigh ing, which is gi en by Gueyma d [19] and Molineaux and Ineichen [20]. This me hod a oids he limi a ions o he Bee –Bougue –Lambe law when applied o la ge spec al o b oad bands. I equi es in ica e pa ame e iza ions because he ansmi ance o an a mosphe ic laye depends on he spec al cha ac e is ics o all he laye s. This echnique gua an ees be e o e all accu acy. I also akes in o accoun he a iable e ec o ci cumsola adia ion, which inc eases wi h u bidi y and ai mass. In he cu en wo k, all he basic unc ional o m is simila o he o iginal MLWT2 model as explained by Gueyma d [12,13], excep ha he o al NO2abso p ion is no aken in o accoun . The equa ion o MLWT2 model is: Ebn =EonTRgToTWTa. He e Ebn, is di ec no mal i adiance; Eon, he ex a- e es ial i adiance (i.e. he sola cons an imes he sun–ea h dis ance co ec ion ac o ); TRg, is he Rayleigh and uni o mly mixed gas ansmi ance; To, he ozone ansmi ance; TW, he wa e apo ansmi ance and Ta, he ae osol ansmi ance. 574 L.N. Rohilla and D.K. Singh / Ene gy Repo s 6 (2020) 572–576 2.3. REST model This model has been p oposed by Gueyma d [12,13,14] and i s o m is as ollows: Ebn =EonTRTgToTWTa. He e Ebn , is di ec no mal i adiance; Eon , he ex a- e es ial i adiance (i.e. he sola cons an imes he sun–ea h dis ance co ec ion ac o ); TR, is he Rayleigh ansmi ance; Tg, he uni o mly mixed gas ansmi ance; To, he ozone ansmi ance; TW, he wa e apo ansmi ance and Ta, he ae osol ansmi ance. 3. Me eo ological da a Fi e s a ions o di e en clima ic condi ions ha e been chosen. The s a ions chosen a e — New Delhi ep esen ing composi e clima e; Kolka a, being highly indus ialized, ep esen s highly pollu ed, wa m and humid a mosphe e; Jaipu , being a dese s a ion, ep esen s ho and d y clima e; Mumbai, being coas al ci y, ep esen wa m and humid clima e; Pune ep esen s mode a e clima e. The sola adia ion da a, comp ising o mon hly mean hou ly global and di use sola adia ion o hese Indian s a ions, ha e been collec ed om India Me eo ology Depa men (IMD) Pune, India. Mon hly mean alues a e aged o e he yea o all he i e s a ions, a e shown in Table 1. Table 1. Geog aphical and clima ic da a o i e Indian s a ions used in his s udy. Jan Feb Ma Ap May Jun Jul Aug Sep Oc No Dec Kolka a (2002) Ta20 22 26 29 30 30 29 29 28 27 23 21 [22.65◦N, 88.45◦ERH76 75 73 74 77 80 81 82 80 77 75 74 6 masl] O30.24 0.26 0.26 0.27 0.28 0.27 0.26 0.26 0.26 0.25 0.25 0.25 Igh 11.71 16.34 17.42 19.33 18.81 15.65 14.48 15.02 14.99 15.31 13.49 11.83 Jaipu (1999) Ta16 21 24 30 36 33 30 29 28 28 23 18 [26.93◦N, 75.86◦E RH48 45 26 22 37 41 70 70 52 28 38 43 431 masl] O30.26 0.27 0.28 0.29 0.29 0.29 0.28 0.27 0.26 0.25 0.25 0.26 Igh 15.22 17.26 23.48 26.56 26.20 24.94 18.77 19.94 19.36 17.33 16.02 13.89 Mumbai (2001) Ta21 23 24 26 29 28 27 26 26 25 25 23 [19.12◦N, 72.85◦E RH67 68 72 75 78 81 87 87 86 82 75 71 14 masl] O30.24 0.26 0.26 0.27 0.28 0.27 0.26 0.26 0.26 0.25 0.25 0.25 Igh 17.17 20.71 23.71 24.61 25.02 16.16 14.13 14.32 16.05 15.06 16.17 15.23 New Delhi (1999) Ta15 16 21 27 31 35 34 32 29 26 21 17 [28.63◦N, 77.20◦E RH53 53 54 40 38 56 65 72 54 44 44 54 216 masl] O30.26 0.27 0.28 0.29 0.29 0.30 0.28 0.27 0.27 0.26 0.25 0.26 Igh 10.23 15.05 20.98 23.45 22.71 21.25 18.04 18.88 16.36 14.35 14.49 11.67 Pune (2002) Ta15 16 17 20 22 22 21 20 22 22 19 14 [18.53◦N, 73.85◦E RH48 45 26 22 37 41 70 70 52 28 38 43 559 masl] O30.26 0.27 0.28 0.29 0.29 0.29 0.28 0.27 0.26 0.25 0.25 0.26 Igh 18.72 19.94 23.54 25.92 24.10 16.62 20.37 13.00 20.20 18.81 18.04 16.33 Ta: Mon hly mean daily empe a u e (◦C); RH: Mon hly mean daily ela i e humidi y (%); O3: Ozone amoun (cm). Igh : Mon hly mean daily global adia ion (MJ/m2day); masl: me e abo e sea le el. Global and Di use adia ion a e measu ed by using he moelec ic py anome e s. These py anome e s a e calib a ed once a yea wi h e e ence o he Wo ld Radiome ic Re e ence (WRR). The es ima ed unce ain y in he measu ed da a is abou ±5%. The alues o Angs om u bidi y ac o , β, o hese loca ions we e compu ed by Louche e al. [21], by aking Angs om’s wa e exponen αas 1.3. Es ima ion o a mosphe ic u bidi y o Indian loca ions ha e been s udied by Na ain and Ga g [22] and Ahe and Agashe [23]. The o he me eo ological da a includes mon hly mean hou ly alues o a mosphe ic p essu e, empe a u e and ela i e humidi y. This da a is used o compu e he amoun o p ecipi able wa e in he a mosphe e, by using an equa ion gi en by Leckne [24]. Da a o ozone con en s o he s udied s a ions ha e been aken om [3]. One-yea da a o each o he i e Indian s a ions was used in his s udy and he yea o each s a ion is shown in Table 1. 4. Resul s and discussions A compu e p og am was w i en o compu e hou ly alue o DNI by using h ee di e en models; CPCR2, REST and MLWT2 model, o each o he 5 Indian s a ions men ioned ea lie . The model-compu ed alues ha e L.N. Rohilla and D.K. Singh / Ene gy Repo s 6 (2020) 572–576 575 been compa ed wi h he measu ed alues in e ms o pe cen age oo mean squa e e o , RMSE (%) and pe cen age mean bias e o , MBE (%), and he esul s a e shown in Table 2. The compu a ions we e no pe o med o he monsoon mon hs, June o Sep embe , as du ing hese mon hs i is cloudy. Table 2. RMSE (%) and MBE (%) o compu ed DNI wi h h ee models, in compa ison o measu ed DNI, o New Delhi (1999), Pune (2002), Jaipu (1999), Kolka a (2002) and Mumbai (2001). Loca ion Mon h RMSE (%) MBE (%) REST MLWT2 CPCR2 REST MLWT2 CPCR2 New Delhi Jan 5.76 3.86 3.15 −4.51 −3.26 −2.74 Feb 3.07 1.06 2.06 −2.33 −0.51 −2.00 Ma 2.38 1.00 2.73 −1.86 −0.63 −2.72 Ap 1.46 0.74 2.66 −0.99 −0.50 −2.65 May 0.92 1.59 2.99 −0.28 −1.39 −2.96 Oc 3.63 5.74 4.40 −3.00 −5.55 −4.38 No 5.19 2.33 3.19 −4.52 −2.12 −3.16 Dec 6.63 2.37 2.45 −5.60 −2.09 −2.25 A e age 3.63 2.33 2.95 −2.89 −2.01 −2.86 Pune Jan 3.32 1.90 4.09 −2.82 −1.68 −4.08 Feb 2.53 1.51 3.97 −2.13 −1.22 −3.95 Ma 1.87 1.27 3.71 −1.34 −0.88 −3.68 Ap 1.65 1.52 3.75 −1.10 −1.12 −3.73 May 2.18 3.76 5.39 −1.99 −3.64 −5.39 Oc 3.19 4.32 5.57 −2.95 −4.21 −5.55 No 3.73 2.98 4.71 −3.35 −2.83 −4.71 Dec 3.99 2.67 4.55 −3.62 −2.53 −4.54 A e age 2.81 2.49 4.47 −2.41 −2.26 −4.45 Jaipu Jan 5.72 1.89 3.70 −4.80 −1.61 −3.70 Feb 3.57 1.19 3.38 −2.93 −0.85 −3.35 Ma 2.08 0.92 2.85 −1.33 −0.05 −2.81 Ap 1.47 0.78 2.83 −0.82 −0.04 −2.80 May 1.30 1.04 3.41 −0.89 −0.77 −3.40 Oc 3.45 3.05 4.44 −3.04 −2.89 −4.41 No 4.96 2.35 4.00 −4.20 −2.01 −3.97 Dec 6.65 2.17 3.83 −5.61 −1.88 −3.78 A e age 3.65 1.67 3.55 −2.95 −1.26 −3.53 Kolka a Jan 3.46 3.13 1.43 −1.80 −2.19 −1.19 Feb 2.40 1.69 1.64 −1.43 −1.25 −1.61 Ma 1.58 2.97 2.04 −0.04 −2.46 −2.01 Ap 1.94 5.49 3.55 −0.18 −4.99 −3.43 May 2.18 7.95 5.12 −0.73 −7.48 −5.06 Oc 2.73 7.32 3.73 −1.87 −7.16 −3.72 No 3.44 3.38 2.18 −2.61 −3.12 −2.11 Dec 3.85 2.46 1.25 −2.74 −2.04 −1.17 A e age 2.70 4.30 2.62 −1.43 −3.84 −2.54 Mumbai Jan 3.25 1.83 2.20 −2.57 −1.45 −2.19 Feb 2.82 1.67 2.44 −2.18 −1.23 −2.39 Ma 2.18 2.04 2.63 −1.77 −1.82 −2.60 Ap 1.72 2.86 2.84 −1.37 −2.75 −2.82 May 1.59 3.56 2.99 −1.26 −3.45 −2.98 Oc 2.21 6.50 3.87 −1.54 −6.41 −3.82 No 3.19 2.81 2.57 −2.73 −2.59 −2.55 Dec 5.12 3.50 3.15 −3.70 −2.68 −2.75 A e age 2.76 3.09 2.84 −2.14 −2.80 −2.76 Table 2 shows ha he minimum RMSE is 0.74% o MLWT2 model, in he mon h o Ap il, and o New Delhi s a ion. Simila ly, able shows ha minimum MBE is –0.04% o MLWT2 model, in he mon h Ap il, o Jaipu . 576 L.N. Rohilla and D.K. Singh / Ene gy Repo s 6 (2020) 572–576 The nega i e sign shows ha he compu ed alue is unde es ima ed and he quan i y shows ha DNI compu ed by MLWT2 model in his mon h o he men ioned s a ion is much close o measu ed alue. The a e age o pe cen age RMSE, o he whole yea , a e he minimum o MLWT2 model and o s a ions Jaipu , ollowed by New Delhi and Pune and he minimum alues a e 1.67, 2.33 and 2.49, espec i ely. While o Kolka a he co esponding minimum alue is in CPCR2 model and he alue is 2.62 and o Mumbai he same is in REST model ha is 2.76. This beha io is because o he Angs om’s u bidi y ac o being highe du ing some o he mon hs o hese wo s a ions. He e i is obse ed ha he pe cen age RMSE is lesse in mos o he mon hs o mos o he s a ions in MLWT2 model. This shows he sui abili y o MLWT2 model o mos o he Indian s a ions. 5. Conclusions The pe o mances o h ee models CPCR2, MLWT2 and REST ha e been s udied a i e Indian s a ions iz. New Delhi, Kolka a, Mumbai, Pune and Jaipu . The beam adia ion a no mal incidence was compu ed. The compu ed DNI was compa ed wi h measu ed DNI. Fo mos o he Indian s a ions, RMSE pe cen age, a e aged o e a yea , is he minimum o MLWT2 model as compa ed o, CPCR2 and REST models. This s udy e eals he bes pe o mance o MLWT2 model in Indian condi ions. Re e ences [1] Na ain L. A s udy on sola adia ion compu a ion o e india [Disse a ion]. New Delhi, India: Indian Ins i u e o Technology; 2010. [2] Powe HC. Es ima ing clea -sky beam i adiance om sunshine du a ion. Sol Ene gy 2001;71:217–24. [3] Iqbal M. An in oduc ion o sola adia ion. New Yo k: Academic P ess; 1983. [4] King R, Buckius RO. Di ec sola ansmi ance o a clea sky. Sol Ene gy 1979;22:297–301. [5] Bi d RE, Huls om RL. Re iew, e alua ion and imp o emen o di ec i adiance model. T ans ASME J Sol Ene gy Eng 1981;103(3):182–92. [6] Choudha y B. A pa ame iza ed model o global insola ion unde pa ially cloudy skies. Sol Ene gy 1982;29(6):479–86. [7] Ko i MC, A gi iou AA, Kazan zidis AK. Use o shadowband co ec ion models o p edic ing di ec sola i adiance. P oc AIP Con : Pap 2013;1531:107–10. [8] Munee T, Gul M, Kambezidis H. E alua ion o an all-sky me eo ological adia ion model agains long- e m measu ed hou ly da a. Ene gy Con e s Manage 1998;39(3–4):303–17. [9] Benkaciali S, Haddadi M, Khella A. E alua ion o di ec sola i adiance om 18 b oadband pa ame ic models: case o Alge ia. Renew Ene gy 2018;125(C):694–711. [10] Munee M, Gul MS, Kubie J. Models o es ima ing sola adia ion and illuminance om me eo ological pa ame e s. T ans ASME Sol Ene gy Eng 2000;122(3):146–53. [11] Yang K, Huang GW, Tamai N. A hyb id model o es ima ing global sola adia ion. Sol Ene gy 2001;70(1):13–22. [12] Gueyma d C. Di ec sola ansmi ance and i adiance p edic ions wi h b oadband models. Pa I: de ailed heo e ical pe o mance assessmen . Sol Ene gy 2003;74:355–79. [13] Gueyma d C. Di ec sola ansmi ance and i adiance p edic ions wi h b oadband models. Pa II: alida ion wi h high quali y measu emen s. Sol Ene gy 2003;74:381–95. [14] Gueyma d C. Co igendum o di ec sola ansmi ance and i adiance p edic ions wi h b oadband models. Pa I: de ailed heo e ical pe o mance assessmen . Sol Ene gy 2004;74:513K. [15] Gueyma d C. Spec al models. In: Go don J, edi o . Sola ene gy – he s a e o he a . London: James and James Publise s; 2001, p. 527–31. [16] Ba le FJ, Olmo FJ, To a J, Alados-A boledas L. Compa ison o cloudless sky pa ame e iza ion o sola i adiance a a ious Spanish mid la i ude loca ion. Theo Appl Clima ol 2000;66:81–93. [17] Gueyma d C. Two band model o he calcula ion o clea sky sola i adiance, illuminance, and pho osyn he ically ac i e adia ion a he ea h su ace. Sol Ene gy 1989;43(5):253–65. [18] Gueyma d C. Tu bidi y de e mina ion om b oadband i adiance measu emen s. J Appl Me eo ol 1998;37(4):414–35. [19] Gueyma d C. Mul ilaye -weigh ed ansmi ance unc ions o use in b oadband i adiance and u bidi y calcula ions. In: Campbell- Howe R, Wilkins-C owde B, edi o s. P oceedings o sola 96, annual con e ence o he Ame ican sola ene gy socie y. Ashe ille, NC: ASES; 1996, p. 281–6. [20] Molineaux B, Ineichen P. On he b oad band ansmi ance o di ec i adiance in a cloudless sky and i s applica ion o he pa ame e iza ion o a mosphe ic u bidi y. Sol Ene gy 1996;56(6):553–63. [21] Louche A, Mau el M, Simonno G, Pe i G, Iqbal M. De e mina ion o Angs om’s u bidi y coe icien om di ec o al sola i adiance measu emen s. Sol Ene gy 1987;38:89–96. [22] Na ain L, Ga g SN. Es ima ion o linke u bidi y ac o s o di e en egions o India. In J En i on Was e Manage 2013;12(1):52–64. [23] Ahe GR, Agashe VV. De e mina ion o a mosphe ic u bidi y a Pune. Indian J Radio Space Phys 1998;27(2):53–9. [24] Leckne B. The spec al dis ibu ion o sola adia ion a he ea h’s su ace. Elemen s o a model. Sol Ene gy 1978;20(2):143–50.