Hindawi Publishing Corporation

Journal of Renewable Energy

Volume 2013, Article ID 307547, 12 pages
http://dx.doi.org/10.1155/2013/307547

Research Article
An Algorithm to Determine the Optimum Tilt Angle of
a Solar Panel from Global Horizontal Solar Radiation

Emanuele Calabrò
Department of Physics, University of Messina, Viale D’Alcontres 31, 98166 Messina, Italy

Correspondence should be addressed to Emanuele Calabrò; e.calabro@yahoo.com

Received 19 December 2012; Revised 24 February 2013; Accepted 1 March 2013

Academic Editor: Jayanta Deb Mondol

Copyright © 2013 Emanuele Calabrò. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper proposes an algorithm to calculate the optimum tilt angle of solar panels by means of global horizontal solar radiation
data, provided from Earth-based meteorological stations. This mathematical modeling is based on the maximization of the
theoretical expression of the global solar irradiation impinging on an inclined surface, with respect to the slope and orientation
of the panel and to the solar hour angle. A set of transcendent equations resulted, whose solutions give the optimum tilt and
orientation of a solar panel. A simulation was carried out using global horizontal solar radiation data from the European Solar
Radiation Atlas and some empirical models of diffuse solar radiation. The optimum tilt angle resulted was related to latitude by
a linear regression with significant correlation coefficients. The standard error of the mean values resulted increased significantly
with latitude, suggesting that unreliable values can be provided at high latitudes.

1. Introduction photovoltaic effect, whereas solar thermal systems generate

heat.
Most countries in the world have realized the need for reduc- The amount of electrical power produced from PV
tion of gases emission to contrast the adverse global climatic systems is related to the amount of solar irradiation projecting
change, encouraging the use of renewable and sustainable on the modules. Hence, the global solar irradiation on tilted
sources of energy. Indeed, large quantities of carbon dioxide, surfaces facing in different directions should be considered
nitrogen, and sulfur oxides are emitted in the world by to estimate thermal and electrical power obtained in archi-
conventional energy sources, which are released to the earth’s tectural planning.
atmosphere contributing to climate change. The literature provides that solar power supplied by the
Furthermore, the world will soon run out of its conven- modules depends on many extrinsic factors, such as insola-
tional energy resources because of the rapid depletion of fossil tion levels, temperature, load conditions, and orientation of
fuel reserves. This future scenario and the risks associated the panel.
with CO2 emissions and global warming have increased the The solar radiation is also a function of the nature and
interest in renewable energy. extent of cloud cover and of the atmosphere’s water vapour
The major renewable energy systems include photo- content, because solar radiation entering Earth’s atmosphere
voltaics (PVs), solar thermal, wind, biomass, hydroelectric, is scattered by atmospheric gases, aerosols, and clouds.
and geothermal. However, among various renewable energy Indeed, meteorological parameters used as predictors
sources, the photovoltaic technology for power generation include the amount and distribution of clouds or other obser-
is considered well-suited technology, particularly for dis- vations such as the fractional sunshine and water content
tributed power generation. [1, 2].
Solar panel is the energy conversion fundamental com- Aerosols can either absorb or scatter the radiation and
ponent of PV systems or solar collectors. Solar panels use alter the energy balance of Earth, especially under clear skies
light energy from the sun to generate electricity through the [3].
2 Journal of Renewable Energy

These parameters, obviously, cannot be modified, The diffuse solar component is the most difficult quantity
whereas other variables can be changed to maximize the to determine, because the distribution of the sky-diffuse
solar energy acquired by the panel. In fact, the design of radiance strictly depends on the local condition of the sky.
a solar energy module involves complex tradeoffs due to One way to estimate the diffuse solar component is to
the interaction of several factors such as the characteristics study the regression between the global and the diffuse solar
of the solar cells, power supply requirements and power radiation at locations where appropriate data are available,
management features of the embedded system, application establishing models which may be used to predict the diffuse
behavior, inclination, and orientation of the panel. solar radiation.
Hence, it is essential to understand and exploit these Liu and Jordan performed the first studies on this subject
factors in order to maximize the energy efficiency of a solar determining a relationship between daily diffuse and global
module. radiation on a horizontal surface, assuming an isotropic
In particular, for fixed absorber surfaces, solar energy diffusion of solar radiance in the sky [9].
gain is strictly related to the slope and azimuth angles of a Erbs et al. used a database acquired from four US weather
solar panel. stations, composed of hourly direct normal radiation and
global radiation, to develop an estimation of the diffuse
The global solar radiation for inclined surfaces can be
fraction of hourly, daily, and monthly average global radiation
calculated by the values of direct and diffuse solar radiation
[10].
on the corresponding horizontal surface. Meteorological data
Moreover, [11, 12] used also two predictors for their
from all parts of the world are needed to know the horizontal
correlations: the clearness index and the solar elevation.
global solar radiation, and, for some regions, measured
Garrison proposed a model to represent the dependency
data may only be applied within a radius of about 50 km
of the diffuse fraction on the surface albedo, atmospheric
from weather stations. This circumstance leads to interpolate
precipitable water, atmospheric turbidity, solar elevation, and
parameters between stations.
global horizontal radiation [13].
Furthermore, spectral irradiance is usually not measured Reindl et al. considered two more significant predictors,
routinely, so that the energy instantaneous production of a the ambient temperature and relative humidity [14], reducing
solar plant is to rely on appropriate models. the standard error of Liu and Jordan-type models.
Otherwise, the users need accurate computations of the On the other side, [15] developed an exponential model
slope and orientation of solar panels in order to maximize the for the estimation of the direct normal beam radiation from
solar energy that is collected by fixed solar panels. the global radiation, named the Disc model. This model was
Advantages of fixed solar panels are mainly related to then improved by [16].
their tolerance to misalignment, as approximately 20% of the However, accurate mathematical modeling of global and
incident solar radiation is diffuse light, available at any angle diffuse solar radiation that was used for the simulation in this
of misalignment with the direct sun. study is proposed in the following section.
In contrast, the main advantage of tracking systems is to Various optimum tilt angle values were provided in the
collect solar energy for the longest period of the day with literature for fixed solar panels.
the most accurate alignment as sun’s position shifts with the For instance, Qiu and Riffat suggested the tilt angle of the
seasons. Indeed, daily solar energy collected was calculated solar collector set within the optimum tilt angle of ±10∘ as an
to be 19%–24% higher by a solar PV panel with one axis east- acceptable practice [17].
west tracking system than by a fixed system [4]. Other, computations led to the values of 𝛽 = 𝜑 ± 20∘ [18],
Nevertheless, since solar tracking systems have high oper- 𝛽 = 𝜑 ± 8∘ [19], 𝛽 = 𝜑 ± 5∘ [20], and 𝛽 = 𝜑 ± 15∘ [21], where
ation and maintenance costs and are not always applicable, it 𝜑 represents the geographical latitude and the signs + and −
is often convenient to set the solar collector at a fixed value of refer to winter and summer months, respectively.
an optimum tilt angle [5]. The disagreement among these values may be due to two
However, some algorithms for the minimization of the main reasons:
energy loss generated by the driving actuator were recently
(1) firstly, the different methods of calculation that were
proposed [6, 7].
used for the determination of the optimum slope
Otherwise, the increase of diffuse solar radiation at low
value of a solar panel;
latitudes at some locations with respect to high latitudes
makes fixed solar panels a competitive alternative to other (2) secondly, the different empirical models that were
energy sources mainly in those locations [8]. considered for the determination of diffuse solar
Hence, the following main question regarding the design radiance and its link with the amount of global solar
of fixed solar panels arises. radiation.
What tilt and orientation of a solar panel have to be The aim of this study was to propose an algorithm for the
chosen to maximize the solar radiation? determination of the optimum tilt and orientation of a solar
Solar radiation impinging on an inclined surface can be panel using a mathematical model based on the orientation
divided into direct, diffuse, and ground reflected radiation. of a generic surface with respect to the position of the sun in
Hence, a method to determine these quantities for each the sky. The other physical variables and the empirical model
latitude has to be studied to answer to that fundamental of diffuse solar radiation are considered successively in the
question. algorithm, linking it to a specific location.
Journal of Renewable Energy 3

2. Mathematical Modeling of Global and where 𝐻𝑔 is the monthly average of daily solar radiation on a
Diffuse Solar Radiation horizontal surface.
The global horizontal solar irradiation 𝐻𝑔 , provided from
The literature provides several methods to estimate the global meteorological stations, includes the horizontal direct beam
and diffuse solar radiation using climatologic parameters. irradiation 𝐻𝑏 and the horizontal diffuse sky irradiation 𝐻𝑑 .
A typical empirical model is the regression equation of Some models can convert global horizontal irradiation
the Angstrom type [22] to direct beam irradiation and diffuse sky irradiation on the
𝐻𝑔 horizontal plane by means of empirical relationships.
𝑛
= 𝑎 + 𝑏( ), (1) Heliosat is an algorithm which was developed to estimate
𝐻o 𝑁 ground level global horizontal irradiance by using Meteosat
where 𝐻𝑔 is the monthly average of daily global solar satellites images taken in the visible band.
radiation impinging on a horizontal surface at a location, 𝐻𝑜 Some predictive models for estimating global solar irra-
is the monthly mean of daily radiation on a horizontal surface diation have to be used to obtain the solar radiation on a
in the absence of atmosphere, 𝑛 is the monthly mean of daily tilted surface, so that the relationship between the global solar
number hour of observed sunshine, 𝑁 is the monthly mean irradiation on horizontal planes and that on tilted planes can
value of day length of the interested location, and “𝑎” and “𝑏” be evaluated [9, 28–31].
are the regression constants, determined from climatological Unfortunately, no theoretic relationship between the hor-
data. The ratio 𝑛/𝑁 is often called the “possible sunshine hour izontal sky diffuse irradiation 𝐻𝑑 and the horizontal global
percentage.” solar irradiation 𝐻𝑔 can be determined rigorously. Indeed,
Regression coefficients “𝑎” and “𝑏” can be obtained from a double integral like equation (5) of [32] should be solved,
some relationships as proposed by [23, 24]. which cannot be computed even in the simplest case of
A statistically worldwide 5.3% average decrease of global uniform and isotropic sky diffuse solar radiation, because the
solar radiation, with largest decline between 45∘ and 30∘ N, distribution of solar irradiance through the sky is difficult to
was found by analyzing the data collected in 45 actinometric be represented adequately.
stations during the years 1958, 1965, 1975, and 1985 [25]. A The direct and diffuse components of solar radiation can
significant decrease in mean yearly global solar radiation be estimated using empirical relationships 𝐻𝑑 = 𝑓(𝐻𝑔 ) by
between the years 1964 and 1990 under completely overcast means of the clearness index 𝐾𝑇 .
skies was found in some locations in Germany [26]. Some of these algorithms, requiring the direct normal and
The observed changes in clear-sky radiation could be diffuse radiation on a horizontal surface as input, can provide
related to the recovery by the volcanic eruptions effects, very different estimated results in different locations [33].
submicron aerosol particles with simultaneous reduction of Indeed, the most considerable cause of error in the
aerosol mass concentration, and increasing absorption by computation of the optimum tilt and orientation of a solar
urban aerosol. panel depends on the model of diffuse solar radiation which
Hence, global horizontal solar irradiation data should be is used.
updated and have to be preferred with respect to empirical Muneer [34] recommends the model proposed by [35] for
expression such as (1). the desert and tropical locations:
Otherwise, daily extraterrestrial radiation on a horizontal 𝐻𝑑
surface, named 𝐻𝑜 , can be computed for the day 𝑛 from the = 1.35 − 1.61𝐾𝑇 . (4)
following equation [27]: 𝐻𝑔

𝐺𝑠𝑐 For temperate climates and locations out with the tropics,
𝐻𝑜 = 86400 ∗ equation (5) given by [36] may be used:
𝜋
𝑛 𝐻𝑑
× (1 + 0.033 cos (2𝜋 )) cos 𝜑 cos 𝛿 = 1.00 − 1.13𝐾𝑇 . (5)
365 (2) 𝐻𝑔

∗ (±√(1 − tan2 𝜑 tan2 𝛿)) The daily horizontal diffuse irradiation 𝐻𝑑 and direct beam
irradiation 𝐻𝑏 = 𝐻𝑔 − 𝐻𝑑 can be obtained by (4) and (5), as
well.
+ cos−1 (− tan 𝜑 tan 𝛿) ∗ sin 𝜑 sin 𝛿,
where 𝐺sc is the solar constant (1367 W/m2 ), 𝑛 is the number 3. The Algorithm to Maximize the Global Solar
of the day, 𝛿 is the solar declination, and 𝜑 is the geographical Radiation on a Fixed Sloped Surface
latitude. The hour angle of sunrise 𝜔𝑠 has been expressed as a
function of 𝜑 and 𝛿 by Cooper’s equation. 3.1. The Case of Isotropic Diffusion. Global solar radiation on
The ratio of solar radiation at the surface of the Earth to a tilted surface 𝐼𝑇 consists of daily direct solar radiation 𝐼𝑏 ,
extraterrestrial radiation is called the clearness index, with diffuse solar radiation 𝐼𝑑 , and ground reflected radiation 𝐼𝑟 .
the monthly average clearness index, 𝐾𝑇 , defined as Daily solar radiation on a tilted surface for a given month
𝐻𝑔 can be estimated as follows [9]:
𝐾𝑇 = , (3)
𝐻𝑜 𝐼𝑇 = 𝐼𝑏 + 𝐼𝑑 + 𝐼𝑟 . (6)
4 Journal of Renewable Energy

The daily direct radiation on a tilted surface 𝐼𝑏 can be obtained This physical condition for the maximization of the solar
by means of 𝑅𝑏 , the ratio of the average daily direct radiation radiation acquired by a solar panel can be represented by the
on a tilted plane to that on a horizontal plane and the mathematical expressions [8]:
parameters to it correlated [37]:

𝜕𝐼𝑇
cos 𝜃 = 0,
𝐼𝑏 = = 𝐻𝑏 𝑅𝑏 , (7) 𝜕𝛽
cos 𝜃𝑧
𝜕𝐼𝑇
𝑅𝑏 = cos 𝛽 = 0, (11)
𝜕𝛾
− sin 𝛽 (sin 𝛿 cos 𝜑 cos 𝛾 𝜕𝐼𝑇
= 0.
− cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾 − cos 𝛿 sin 𝛾 sin 𝜔) 𝜕𝜔

−1
× (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ,
(8) Their application to the isotropic model of [9] provides the
following expressions [8]:

where 𝛽 is the slope of the panel as to the horizontal plane, 𝛾

is the azimuth, 𝜔 is the solar hour angle, 𝜑 is the latitude, and 𝜕𝐼𝑇 𝐻 𝐻𝑔 𝜌
𝜃 and 𝜃𝑧 are the solar incidence angle on the considered plane = sin 𝛽 (𝐻𝑜 + 𝑑 − )
𝜕𝛽 2 2
and the solar zenith angle, respectively.
If we consider a uniform and isotropic distribution of − cos 𝛽𝐻𝑜
diffuse solar radiation over the sky hemisphere, 𝐼𝑑 would be
easily obtained from the simple approximation of [9]: × (cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾 − sin 𝛿 cos 𝜑 cos 𝛾

+ cos 𝛿 sin 𝛾 sin 𝜔)

𝐻𝑑 (1 + cos 𝛽) × (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔)
−1
= 0,
𝐼𝑑 (iso) = . (9) (12)
2
𝜕𝐼𝑇
= sin 𝛾 (cos 𝛿 sin 𝜑 cos 𝜔 − sin 𝛿 cos 𝜑)
𝜕𝛾
The evaluation of the ground-reflected diffuse radiation
depends on 𝐼𝑟 . Most studies consider that the ground reflec- − cos 𝛾 cos 𝛿 sin 𝜔 = 0,
tion process is ideally isotropic, in which a specific case 𝐼𝑟 can
𝜕𝐼𝑇
be simplified as follows: = sin 𝜔 sin 𝛿 cos 𝛾 − cos 𝛿 cos 𝜑 sin 𝛾
𝜕𝜔
− cos 𝜔 sin 𝛿 sin 𝜑 sin 𝛾 = 0.
𝐻𝑔 𝜌 (1 − cos 𝛽)
𝐼𝑟 = , (10)
2
This set of equations can be solved with respect to the angles
𝛽, 𝛾, and 𝜔 [8]:
where 𝜌 represents the diffuse reflectance of the ground (also
called ground albedo).
Finally, the daily global solar irradiation on slopes 𝐼𝑇 can
be expressed as the sum of (7), (9), and (10).
𝛽 = tan−1 {𝐻𝑜
The global solar radiation incident on a sloped surface
depends on the position of the sun along its daily trajectory
(represented by the solar angle 𝜔) and on the orientation of × (cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾 − sin 𝛿 cos 𝜑 cos 𝛾
the panel (represented by the slope 𝛽 and the azimuth 𝛾).
The other quantities appearing in (6), (7), (8), (9), and (10) + cos 𝛿 sin 𝛾 sin 𝜔)
depend on the local conditions and can be considered fixed
parameters. 𝐻𝑑 𝐻𝑔 𝜌
× [(𝐻𝑜 + − )
The algorithm proposed here is based on the assumption 2 2
that the daily solar irradiation impinging on a collecting −1
surface is maximum with respect to the angles 𝛽, 𝛾, and 𝜔, × (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ] } ,
where 𝛽 is the slope of the panel as to the horizontal plane, 𝛾
is the azimuth, and 𝜔 is the solar hour angle, respectively. (13)
Journal of Renewable Energy 5

cos 𝛿 sin 𝜔 Klucher [40] proposed a model in which the isotropic

𝛾 = tan−1 { }, (14)
(cos 𝛿 sin 𝜑 cos 𝜔 − sin 𝛿 cos 𝜑) component 𝐼𝑑 (iso) is multiplied by two factors that represent
both circumsolar and horizon brightening:
𝜔 = sin−1 { (sin 𝛿 cos 𝛿 sin 𝛾 cos 𝛾 cos 𝜑
𝛽
𝐼𝑑 = 𝐼𝑑 (iso) (1 + 𝐾sin3 ( )) (1 + 𝐾cos2 𝜃sin3 𝜃𝑧 ) , (18)
± [sin4 𝛿 (sin2 𝜑sin2 𝛾cos2 𝛾 + sin4 𝜑sin4 𝛾) 2
1/2
(15)
−sin2 𝛿 cos2 𝛿sin2 𝜑cos2 𝜑sin4 𝛾] ) where the parameter 𝐾 expresses the degree of anisotropy
as a modulating function of the amount of direct radiation
×(sin2 𝛿cos2 𝛾 + sin2 𝛿sin2 𝜑sin2 𝛾) } .
−1 received by the surface, 𝐾 = 1−𝐻𝑑 /𝐻𝑔 (this model reduces to
the Liu-Jordan isotropic model if the ratio of diffuse to global
radiation 𝐻𝑑 /𝐻𝑔 is close to the unity).
The transcendent equations (13), (14), and (15) can be solved Further models have been proposed in the literature
by iterative methods with respect to the angles 𝛽, 𝛾, and 𝜔, to perform the anisotropy of diffuse solar irradiance, by
which provide, respectively, the optimum tilt and orientation means of some coefficients derived from statistical analyses
of the solar panel and the angular position of the sun in the of empirical data for specific locations [41–43].
sky where the maximization of the solar radiation on the Nevertheless, the anisotropic models of [38–40] have
panel occurs. been taken into account here, because of their easy applica-
bility to every location. In addition, they are extensively vali-
3.2. The Case of Anisotropic Diffusion. The diffuse solar radi- dated models that convert hemispherical data on a horizontal
ance influences the performance of most of the solar energy surface to hemispherical data on a tilted surface, computing
technologies. In fact, the scattering atmospheric processes diffuse solar radiation.
redistribute solar energy out of the direct beam into the The mathematical conditions (11) applied to the models
diffuse radiation. The PV systems, like flat-plate collectors, of [38, 39] provided the following solutions:
have the peculiarity to use both direct and diffuse forms of
radiation, whereas solar concentrator systems can use only
direct radiation. 𝛽 = tan−1 {[ (𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 )
Unfortunately, the planetary network of diffuse solar
× (cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾
radiation measurements stations is poor, whereas global solar
radiation data are available for many locations.
Nevertheless, empirical correlations between diffuse and − sin 𝛿 cos 𝜑 cos 𝛾 + cos 𝛿 sin 𝛾 sin 𝜔) ]
global ratios and diffuse and direct solar radiation can be used
to reduce errors in computing hemispherical radiation from 𝐹𝐻𝑑 𝐻𝑔 𝜌
× [(𝐻𝑜 + + (1 − 𝐹) 𝐻𝑑 − )
estimates of direct and diffuse radiation, especially regarding 2 2
nonclear sky conditions. −1
Besides the isotropic model of diffuse solar radiation × (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ] } ,
described in [9], several models have been proposed to repre-
sent the anisotropy of the diffuse component 𝐼𝑑 by means of (19)
empirical relations which should modify the expressions (12)
and the relative solutions (13), (14), and (15). 𝛾 = tan−1 {[ (𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) cos 𝛿 sin 𝜔]
The assumption that the diffuse radiation originates × [(𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) (20)
entirely from the solar disk gives the relation −1
× (cos 𝛿 sin 𝜑 cos 𝜔 − sin 𝛿 cos 𝜑)] } ,
𝐻𝑑 cos 𝜃
𝐼𝑑 (disk) = , (16) 𝜔 = sin−1 {[ (𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) sin 𝛿 cos 𝛿 sin 𝛾 cos 𝛾 cos 𝜑
cos 𝜃𝑧
4
± [[(𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) sin 𝛿]
which is the modeling opposite to the isotropic one of [9].
Le Quere [38] and Hay and Davies [39] proposed a model ∗ (sin2 𝜑sin2 𝛾cos2 𝛾 + sin4 𝜑sin4 𝛾)
for the diffuse solar component as a combination of the two
2
components, that is, (9) and (16), − [(𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) sin 𝛿 cos 𝛿]
1/2
𝐼𝑑 = 𝐹 ∗ 𝐼𝑑 (iso) + (1 − 𝐹) ∗ 𝐼𝑑 (disk) , (17) ×sin2 𝜑cos2 𝜑sin4 𝛾] ]
2
× [[(𝐻𝑜 + (1 − 𝐹) 𝐻𝑑 ) sin 𝛿]
where (1 − 𝐹) expresses the anisotropy degree.
−1
The value 𝐹 = 0.8 was suggested by [38], while [39] × (cos2 𝛾 + sin2 𝜑sin2 𝛾)] } .
assumed the ratio of terrestrial direct radiation to extrater-
restrial radiation as the degree of anisotropy (1 − 𝐹). (21)
6 Journal of Renewable Energy

Table 1: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of Tripoli
(latitude +32∘ 57󸀠 N; longitude +13∘ 12󸀠 E).

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 50.7 12.44 55.3 13.05 53.3 12.73 57.0 18.54 1.17 1.26
𝑑2 = −13.289 43.6 17.13 47.5 17.96 46.0 17.59 48.7 21.56 0.95 0.87
𝑑3 = −2.819 33.7 22.54 37.0 23.61 35.8 23.18 37.4 24.06 0.72 0.28
𝑑4 = 9.415 21.9 27.08 24.3 28.24 23.5 27.79 23.8 26.01 0.45 0.42
𝑑5 = 18.792 12.9 29.26 14.3 30.16 13.8 29.83 13.6 27.56 0.25 0.50
𝑑6 = 23.314 8.6 30.92 9.5 31.53 9.2 31.34 8.9 29.31 0.17 0.44
𝑑7 = 21.517 10.6 31.92 11.4 32.55 11.2 32.39 10.9 30.00 0.15 0.50
𝑑8 = 13.784 18.4 30.71 19.8 31.58 19.4 31.35 19.4 28.77 0.25 0.55
𝑑9 = 2.217 29.5 25.74 31.9 26.77 31.2 26.43 32.0 25.90 0.50 0.22
𝑑10 = −9.599 40.9 20.16 44.0 20.99 43.0 20.69 44.9 23.73 0.75 0.69
𝑑11 = −19.148 48.5 13.26 53.2 13.94 51.1 13.58 54.8 18.79 1.18 1.13
𝑑12 = −23.335 51.8 10.64 57.2 11.21 54.5 10.87 58.9 16.54 1.35 1.23

Table 2: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of Athens
(latitude +37∘ 58󸀠 N; longitude +23∘ 43󸀠 E).

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 50.1 6.58 60.5 7.42 53.5 6.73 60.2 9.89 2.22 0.66
𝑑2 = −13.289 42.3 9.42 52.7 10.66 45.8 9.68 52.2 12.16 2.19 0.54
𝑑3 = −2.819 33.0 13.16 42.4 14.94 36.2 13.59 40.9 14.65 1.87 0.37
𝑑4 = 9.415 24.7 20.47 30.2 22.59 27.1 21.24 28.8 20.13 1.03 0.47
𝑑5 = 18.792 17.1 24.75 20.5 26.65 18.7 25.56 18.9 23.23 0.61 0.62
𝑑6 = 23.314 13.6 28.03 15.8 29.58 14.8 28.82 14.6 26.03 0.39 0.66
𝑑7 = 21.517 15.7 29.27 17.7 30.82 16.9 30.15 16.8 26.97 0.36 0.73
𝑑8 = 13.784 22.9 27.03 25.9 28.83 24.7 28.03 25.3 25.33 0.55 0.65
𝑑9 = 2.217 33.6 21.73 37.8 23.49 36.0 22.62 38.2 22.39 0.90 0.31
𝑑10 = −9.599 43.2 14.21 49.5 15.64 46.2 14.77 50.6 17.48 1.45 0.62
𝑑11 = −19.148 48.2 7.38 58.5 8.32 51.6 7.56 58.2 10.63 2.19 0.65
𝑑12 = −23.335 52.3 5.92 62.5 6.67 55.6 6.06 62.2 9.35 2.18 0.69

Table 3: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of Rome
(latitude +41∘ 54󸀠 N; longitude +12∘ 27󸀠 E).

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 57.5 6.93 64.6 7.93 60.4 7.21 65.8 11.47 1.66 0.91
𝑑2 = −13.289 49.5 10.56 56.9 12.04 52.7 11.02 58.1 14.38 1.71 0.74
𝑑3 = −2.819 40.1 16.05 46.7 18.15 43.1 16.82 47.4 18.12 1.47 0.44
𝑑4 = 9.415 28 20.02 34.3 22.67 30.7 20.95 33.3 19.77 1.22 0.56
𝑑5 = 18.792 20.2 24.07 24.6 26.66 22.2 25.07 23 22.46 0.79 0.76
𝑑6 = 23.314 16.8 27.14 20 29.45 18.4 28.17 18.5 24.90 0.57 0.83
𝑑7 = 21.517 19 28.82 21.9 31.08 20.7 29.98 20.9 26.21 0.52 0.90
𝑑8 = 13.784 26.5 26.82 30.1 29.19 28.6 28.07 29.7 24.99 0.70 0.78
𝑑9 = 2.217 36.5 19.81 41.8 22.06 39.2 20.80 42.3 20.75 1.15 0.40
𝑑10 = −9.599 49 15.29 53.6 16.83 51.6 16.04 55.3 19.35 1.17 0.76
𝑑11 = −19.148 55.7 8.01 62.6 9.13 58.6 8.35 63.9 12.56 1.63 0.90
𝑑12 = −23.335 61.5 7.05 66.7 7.94 64 7.38 68.4 12.76 1.31 1.15
Journal of Renewable Energy 7

Table 4: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of Paris
(latitude +48∘ 51󸀠 N; longitude +2∘ 20󸀠 E).

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 56.7 3.98 68.1 4.94 59.9 4.10 66.5 6.54 2.33 0.51
𝑑2 = −13.289 46.9 6.05 60.4 7.53 50.4 6.25 58 8.21 2.74 0.44
𝑑3 = −2.819 38.3 10.30 50.2 12.66 41.7 10.73 48.3 11.94 2.41 0.47
𝑑4 = 9.415 28.7 15.74 38.1 18.95 31.7 16.47 35.7 15.92 1.80 0.64
𝑑5 = 18.792 20.4 17.66 28.4 21.00 22.8 18.36 24.5 16.94 1.45 0.76
𝑑6 = 23.314 18.3 22.16 23.9 25.46 20.4 23.11 21.2 20.59 1.00 0.88
𝑑7 = 21.517 21.1 24.83 25.9 28.13 23.3 26.05 24.2 22.76 0.87 0.97
𝑑8 = 13.784 27.8 21.73 33.9 25.06 30.5 22.92 32.9 20.72 1.17 0.80
𝑑9 = 2.217 39 17.48 45.7 20.24 42 18.50 46.2 18.67 1.46 0.49
𝑑10 = −9.599 48.7 10.22 57.1 12.14 52 10.75 57.9 13.32 1.88 0.60
𝑑11 = −19.148 58.1 5.93 66.3 7.12 61.1 6.21 66.9 9.67 1.83 0.73
𝑑12 = −23.335 62.6 4.31 70.3 5.22 65.3 4.50 70.6 8.02 1.70 0.74

Equations (19), (20), and (21) can be solved by iterative meth- 𝜕𝐼𝑇 𝐻𝑜 sin 𝛽 (cos 𝛿 cos 𝜔 sin 𝛾 − cos 𝛿 sin 𝜑 sin 𝜔 cos 𝛾)
=
ods with respect to the angles 𝛽, 𝛾, and 𝜔. The mathematical 𝜕𝜔 sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔
conditions (11) applied to the model of [40], represented by
(18), provided the following expressions: 𝛽
+ (1 + 𝐹sin3 ( )) 𝐻𝑑 𝐹 (1 + cos 𝛽) 𝑞2 √𝑞1
2

𝜕𝐼𝑇 × [sin 𝛽 (cos 𝜔 cos 𝛿 sin 𝛾 − sin 𝜔 cos 𝛿 sin 𝜑 cos 𝛾)

= − 𝐻𝑜 sin 𝛽
𝜕𝛽 − cos 𝛽 cos 𝛿 cos 𝜑 sin 𝜔]
− 𝐻𝑜 cos 𝛽 [ (sin 𝛿 cos 𝜑 cos 𝛾 − cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾 𝛽 3
+ (1 + 𝐹sin3 ( )) 𝐻𝑑 𝐹 (1 + cos 𝛽) ( ) 𝑞1
− cos 𝛿 sin 𝛾 sin 𝜔) 2 4
−1 2 1/2
× (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ] × [1 − (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ]
(1 + 𝐹sin3 (𝛽/2)) × (sin 𝜔 cos 𝛿 cos 𝜑) = 0,
−𝐻𝑑 sin 𝛽
2 (24)
𝛽 (1 + cos 𝛽) 𝐻𝑑 𝐹 sin 𝛽𝑞1 𝑞2 𝑞3
+ 3sin2 ( ) 𝐻𝑑 𝐹 − where
2 2 2
𝛽 (1 + cos 𝛽) 𝑞1 = [cos 𝛽 (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔)
+ 3𝐹sin2 ( ) 𝑞1 𝑞2 𝐻𝑑 𝐹
2 2 + sin 𝛽 (cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾
𝐻𝑔 𝜌 sin 𝛽 2
+ + 𝐻𝑑 𝐹 (1 + cos 𝛽) (√𝑞1 ) + cos 𝛿 sin 𝜔 sin 𝛾 − sin 𝛿 cos 𝜑 cos 𝛾)] ,
2
2 3/2
∗ [cos 𝛽 (cos 𝛿 sin 𝜑 cos 𝜔 cos 𝛾 + cos 𝛿 sin 𝛾 sin 𝜔 𝑞2 = [1 − (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) ] ,
− sin 𝛿 cos 𝜑 cos 𝛾) 𝛽
𝑞3 = [1 + 𝐹sin3 ( )] . (25)
− sin 𝛽 (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔)] 𝑞2 𝑞3 = 0, 2
(22)
𝜕𝐼𝑇 This set of equation can be solved only using iterative
= sin 𝛽 (𝐻𝑜 sin 𝛿 cos 𝜑 sin 𝛾 − 𝐻𝑜 cos 𝛿 sin 𝜑 cos 𝜔 sin 𝛾
𝜕𝛾 methods.
+𝐻𝑜 cos 𝛿 cos 𝛾 sin 𝜔)
−1
× (sin 𝛿 sin 𝜑 + cos 𝛿 cos 𝜑 cos 𝜔) 4. Applying the Maximization Algorithm
𝛽 to a Data Set of Global Horizontal Solar
+ (1 + 𝐹sin3 ( )) 𝐻𝑑 𝐹 (1 + cos 𝛽) 𝑞2 (√𝑞1 ) Radiation
2
× [sin 𝛽 (sin 𝛿 cos 𝜑 sin 𝛾 + cos 𝛿 cos 𝛾 sin 𝜔 The data collected at the European Solar Radiation Atlas
− cos 𝛿 sin 𝜑 cos 𝜔 sin 𝛾) ] = 0, (available at the Internet site HelioClim) were used to test the
(23) algorithm proposed here.
8 Journal of Renewable Energy

Table 5: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of London
(latitude +51∘ 32󸀠 N; longitude 0∘ 05󸀠 W).

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 61.8 2.24 73.8 3.10 64.7 2.32 70.6 4.00 2.36 0.36
𝑑2 = −13.289 52.1 4.39 66.1 5.95 55.5 4.56 63 6.26 2.80 0.41
𝑑3 = −2.819 41.4 7.79 55.9 10.46 44.9 8.14 52.6 9.26 2.90 0.52
𝑑4 = 9.415 32 13.31 44 17.36 35.3 14.02 40.8 13.70 2.33 0.80
𝑑5 = 18.792 25.8 18.88 34.6 23.65 28.6 19.96 31.7 17.86 1.45 1.09
𝑑6 = 23.314 20.7 17.87 29.9 22.55 23.1 18.71 25.2 16.84 1.69 1.07
𝑑7 = 21.517 21.7 16.61 31.7 21.20 24.2 17.37 26.7 15.83 1.84 1.03
𝑑8 = 13.784 28.2 14.54 39.6 18.85 31.2 15.29 35.8 14.44 2.17 0.90
𝑑9 = 2.217 39.1 11.02 51.1 14.37 42.6 11.63 49.3 12.24 2.43 1.13
𝑑10 = −9.599 54.9 8.65 62.9 10.84 58 9.23 63.8 11.86 1.82 1.13
𝑑11 = −19.148 63.1 3.68 72.0 4.84 65.9 3.87 71.3 6.53 1.85 1.06
𝑑12 = −23.335 64.2 1.73 75.8 2.41 66.8 1.78 72.2 3.31 2.26 0.32

Table 6: Optimum tilt angle values and monthly mean of the daily global solar radiation values computed by means of the algorithm proposed
in Section 2 applied to the monthly averages of the daily global horizontal solar irradiation data acquired from 1985 to 1989 at the site of
Stockholm (latitude +59∘ 17󸀠 N; longitude +18∘ 03󸀠 E). The monthly mean of the daily global solar radiation was not acquired in the periods
relative to the solar declinations 𝑑1 , 𝑑11 , and 𝑑12 .

Isotropic model Le Quere’s model Hay and Davies’ model Klucher’s model SEM values
Solar declin. (deg)
𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 ) 𝛽 (deg) 𝐸 (MJ/m2 )
𝑑1 = −21.269 n.c. n.c. n.c. n.c. n.c. n.c.
𝑑2 = −13.289 67.1 5.18 73.2 6.99 69.5 5.65 73.8 8.79 1.37 0.70
𝑑3 = −2.819 54.7 10.25 63.1 13.77 57.9 11.15 63.8 12.56 1.88 0.68
𝑑4 = 9.415 43.1 18.64 51.3 24.32 46.4 20.37 51.8 18.32 1.80 1.19
𝑑5 = 18.792 33.0 21.99 41.8 29.10 36.2 23.88 40.7 19.95 1.77 1.70
𝑑6 = 23.314 29.5 25.39 37.3 32.92 32.4 27.54 35.9 22.19 1.52 1.95
𝑑7 = 21.517 29.8 21.84 39.0 29.12 32.9 23.59 37.0 19.68 1.78 1.75
𝑑8 = 13.784 33.8 14.17 46.5 20.12 37.2 15.15 43.4 13.98 2.50 1.25
𝑑9 = 2.217 46.3 10.25 58.0 14.45 49.8 11.02 56.9 11.52 2.44 0.79
𝑑10 = −9.599 60.5 5.65 69.5 7.93 63.5 6.09 69.3 8.25 1.92 0.56
𝑑11 = −19.148 n.c. n.c. n.c. n.c. n.c. n.c.
𝑑12 = −23.335 n.c. n.c. n.c. n.c. n.c. n.c.

In particular, the global horizontal solar radiation 𝐻𝑔 The isotropic model of [9] and the anisotropic models of
relative to Tripoli, Athens, Rome, Paris, London, and Stock- [38–40] were used to calculate the diffuse solar component
holm, located from 32∘ to 59∘ north latitude, was considered 𝐻𝑑 .
by averaging the values on a monthly basis from 1985 to 1989 The solutions of (14), (15), (20), (21), (23), and (24) gave
and was used for this simulation. the values 𝛾 = 0∘ and 𝜔 = 0∘ , confirming that the optimum
The calculations of 𝛽, 𝛾, and 𝜔 were carried out at every orientation of a solar panel is toward South, where in fact the
month because their time variation cannot be neglected [44]. solar hour angle at solar noon must be zero, as it is defined.
The solar declination values 𝛿 were computed at the Nevertheless, other models of the solar diffuse compo-
middle of every month by Cooper’s equation, which was nent could provide different values.
also used to calculate the sunset hour angle 𝜔𝑠 [37]. Daily The optimum tilt angle values, provided by the solutions
extraterrestrial radiation on a horizontal surface, 𝐻𝑜 , was of (13), (19), (22), were reported on Tables 1, 2, 3, 4, 5, and 6.
computed at the same day by [27]; equation (5) from [39] was The corresponding monthly average of the daily global
used as a relationship between 𝐻𝑑 and 𝐻𝑔 . Ground albedo solar radiation (measured in MJ/m2 ) collected by a surface
was fixed at the typical value 𝜌 = 0.20. inclined at each 𝛽 value was indicated in Tables 1–6, as
Journal of Renewable Energy 9

Table 7: The average 𝛽 of the optimum tilt angle values computed using the diffuse solar radiation models, relative to the latitude values
used above. A linear regression with 95% confidence interval provided the coefficients 𝑎1 and 𝑎2 and the correlation coefficients ranging from
0.944 to 0.993, reported in the last column.

Coefficients
𝛽 (𝑙1 ) 𝛽 (𝑙2 ) 𝛽 (𝑙3 ) 𝛽 (𝑙4 ) 𝛽 (𝑙5 ) 𝛽 (𝑙6 ) 𝑎1 𝑎2 Correlation 𝑟

54.07 56.08 62.07 62.80 67.72 — 31.33 0.68 0.953

46.45 48.25 54.30 53.93 59.17 70.90 16.25 0.86 0.944
35.97 38.13 44.32 44.62 48.70 59.87 6.80 0.84 0.957
23.37 27.70 31.57 33.55 38.03 48.15 −6.07 0.87 0.975
13.65 18.80 22.50 24.03 30.17 37.92 −14.95 0.87 0.978
9.05 14.70 18.42 20.95 24.73 33.77 −19.27 0.87 0.985
11.03 16.78 20.62 23.62 26.08 34.67 −15.65 0.83 0.988
19.25 24.70 28.72 31.27 33.70 40.22 −4.23 0.75 0.989
31.15 36.40 39.95 43.22 45.53 52.75 6.42 0.77 0.993
43.20 47.37 52.37 53.93 59.90 65.70 15.84 0.83 0.983
51.90 54.13 60.20 63.10 68.07 — 23.61 0.84 0.977
55.60 58.15 65.15 67.20 69.75 — 30.56 0.76 0.968

Table 8: The average standard error of the mean (SEM) of the of Tables 1–6) and resulted in increase with increasing
tilt angles and of the monthly average of the daily global solar latitude.
radiation values of diffuse solar radiation, computed separately for
summer months (solar declinations 𝑑4 –𝑑9 ) and for winter months
Furthermore, the optimum tilt angle values were averaged
(solar declinations 𝑑1 –𝑑3 and 𝑑10 –𝑑12 ) as a function of latitude (the
correlation coefficient 𝑟 relative to winter months was computed
for each monthly solar declination and were reported as a
excluding the values relative to the latitude of Stockholm). function of latitude in Table 7. The average tilt angles resulted
were significantly related to latitudes. A correlation between
Average values of SEM values these two sets of variables was studied, applying a linear
Latitude (deg) Summer months Winter months relationship to predict the value of the optimum tilt angle for
a given geographical latitude.
𝛽 𝐸 𝛽 𝐸
In particular, the regression line was performed by the
32.95 0.30 0.44 1.01 0.91 method of least squares, to make the sum of the squares
37.96 0.64 0.57 2.01 0.59 of the differences between the ordinates of the points and
41.90 0.82 0.71 1.50 0.82 those on the straight line as small as possible. The correlation
48.85 1.26 0.75 2.15 0.58 coefficient “𝑟” of the variables was calculated, defined as their
51.53 2.02 0.92 2.33 0.46
covariance (that measures how the two variables are linearly
related) divided by the product of their individual standard
59.28 1.96 1.44 n.c. n.c.
deviations. In addition, a confidence interval was considered
Correlation to know how accurate is the regression.
0.956 0.935 0.835 −0.789
coefficient 𝑟 Linear regressions with 95% confidence interval provided
correlation coefficients ranging from 0.944 to 0.993, reported
in the last column of Table 7 (larger correlation values were
well. Such a value can be considered representative of the
found for summer months). Two typical fits for two solar
conversion efficiency of a PV module, because of a linear
declinations are shown in Figures 1(a) and 1(b).
relationship with solar radiation rate [45].
This result shows that the optimum tilt angle of a solar
The results led to the following considerations.
panel could be easily obtained as a function of geographical
(1) The optimum tilt angle values of solar panels deter- latitude by means of the coefficients 𝑎1 and 𝑎2 of a linear
mined for winter months were confirmed to be very regression 𝛽 = 𝑎1 +𝑎2 ∗𝜑. The coefficients 𝑎1 and 𝑎2 , obtained
different from the values recommended for summer by the regression, depend on the solar declination that was
months. used.
Furthermore, the application of the algorithm showed
(2) The value 𝛾 = 0∘ was confirmed to be the optimum that the differences among the tilt angle values, computed
orientation value for a solar panel. using various models of diffuse solar irradiance, increase with
(3) The disagreement among the models of diffuse solar increasing of the geographical latitude, suggesting that fur-
irradiance resulted was relevant for the winter months ther empirical correlations between diffuse and global ratios
(represented by the solar declinations 𝑑1–3 and 𝑑10–12 , and diffuse and direct solar radiation should be investigated,
corresponding to the first three and the last three rows especially at high latitudes.
10 Journal of Renewable Energy

55
60
50
Optimum tilt angle (deg)

55

Optimum tilt angle (deg)

45
50

40
45

35 40

30 35

30 35 40 45 50 55 60 30 35 40 45 50 55 60
Latitude (deg) Latitude (deg)
(a) (b)

Figure 1: Average values of optimum tilt angles, computed using diffuse solar radiation models, as a function of latitude, relative to the solar
declination 𝑑9 (a) and to the solar declination 𝑑3 (b). Linear regressions with 95% confidence interval provided the correlation coefficients
𝑟 = 0.993 and 𝑟 = 0.957, respectively.

2.2 1.6
SEM of monthly average of daily solar radiation

2
1.4
1.8
SEM of tilt angle values

1.6 1
1.4
values

1.2 1.2

1
0.8
0.8
0.6 0.6
0.4
0.4
0.2

30 35 40 45 50 55 60 30 35 40 45 50 55 60
Latitude (deg) Latitude (deg)
(a) (b)

Figure 2: The standard error of the mean (SEM) of the tilt angle values (a) and of the monthly average of the daily global solar radiation values
(b), computed for the summer months (solar declinations 𝑑4 –𝑑9 ) as a function of latitude. The linear fits provided the correlation coefficients
𝑟 = 0.956 and 𝑟 = 0.935, respectively.

Finally, the standard error of the mean (SEM) of the provided the correlation coefficients 𝑟 = 0.956 and 𝑟 = 0.935
tilt angles and of the monthly average of the daily global for the tilt angles and the monthly average of the daily global
solar radiation values was computed and reported in the last solar radiation values, respectively. These linear regressions
column of Tables 1–6, to quantify the disagreement among were represented in Figures 2(a) and 2(b).
the various models of diffuse solar irradiance used here. Regarding the winter months, lower correlations were
The average values of these SEM values were computed found (see Table 8). Nevertheless, the values obtained for
separately for summer months and for winter months and the winter months can be considered less reliable because
reported in Table 8. the decrease in intensity of direct and diffuse solar radiation
A significant correlation between the averages of these during winter months produces an increase in the relative
SEM values and the latitudes was found for summer months. error for estimating the tilt angles and the monthly average
In fact, a linear regression with 95% confidence interval of the daily global solar radiation values.
Journal of Renewable Energy 11

5. Conclusions [7] S. Seme and G. Štumberger, “A novel prediction algorithm for

solar angles using solar radiation and Differential Evolution for
The method of maximization of global solar radiation pro- dual-axis sun tracking purposes,” Solar Energy, vol. 85, no. 11,
posed here led to take into account a set of equations, which pp. 2757–2770, 2011.
can be solved with respect to the slope 𝛽 and the azimuth [8] E. Calabrò, “The disagreement between anisotropic-isotropic
𝛾 of a collecting surface and to the solar hour angle 𝜔. The diffuse solar radiation models as a function of solar declination:
proposed algorithm was applied to some models of diffuse computing the optimum tilt angle of solar panels in the area of
solar irradiance, using global horizontal solar radiation data southern-Italy,” Smart Grid and Renewable Energy, vol. 3, no. 4,
from the European Solar Radiation Atlas. The relative equa- pp. 253–259, 2012.
tions provided a set of angles 𝛽, 𝛾, and 𝜔, as a function of the [9] B. Y. H. Liu and R. C. Jordan, “The interrelationship and
other physical parameters. The better solar panel orientation characteristic distribution of direct, diffuse and total solar
toward South and the large time dependence of the optimum radiation,” Solar Energy, vol. 4, no. 3, pp. 1–19, 1960.
tilt angle were confirmed, suggesting the use of semifixed [10] D. G. Erbs, S. A. Klein, and J. A. Duffie, “Estimation of the
panels on building structures. diffuse radiation fraction for hourly, daily and monthly-average
The solutions of the set of equations provided tilt angles global radiation,” Solar Energy, vol. 28, no. 4, pp. 293–302, 1982.
and monthly average of daily global solar radiation values [11] M. Iqbal, “Prediction of hourly diffuse solar radiation from
strictly related to the model of diffuse solar irradiation which measured hourly global radiation on a horizontal surface,” Solar
was used. Energy, vol. 24, no. 5, pp. 491–503, 1980.
The computed tilt angles resulted were significantly [12] A. Skartveit and J. A. Olseth, “A model for the diffuse fraction of
related to the latitude value, because linear regressions with hourly global radiation,” Solar Energy, vol. 38, no. 4, pp. 271–274,
95% confidence interval provided correlation coefficients 1987.
ranging from 0.944 to 0.993. This result allows to estimate the [13] J. D. Garrison, “A study of the division of global irradiance into
optimum tilt angle of a solar panel as a function of latitude by direct and diffuse irradiance at thirty-three U.S. sites,” Solar
Energy, vol. 35, no. 4, pp. 341–351, 1985.
means of a linear regression, whose coefficients depend on
global horizontal solar radiation data. [14] D. T. Reindl, W. A. Beckman, and J. A. Duffie, “Diffuse fraction
Nevertheless, relevant differences among the monthly correlations,” Solar Energy, vol. 45, no. 1, pp. 1–7, 1990.
average of the daily global solar radiation and the tilt angle [15] E. L. Maxwell, “A quasi-physical model for converting hourly
values, obtained using different models of diffuse solar global horizontal to direct normal insolation,” Tech. Rep.
SERI/TR-215-3087, Solar Energy Institute, Golden, Colo, USA,
radiation, were highlighted by the simulations performed
1987.
separately for summer and winter months.
The standard error of the mean of such values resulted [16] R. R. Perez, P. Ineichen, E. L. Maxwell, R. D. Seals, and
A. Zelenka, “Dynamic global-to-direct irradiance conversion
in significant increase with increasing latitude for summer
models,” in Proceedings of ISES Solar World Congress, pp. 951–
months, leading to the conclusion that values unreliable for 956, Denver, Colo, USA, 1992.
the optimum tilt angle of a solar panel can be provided by the
[17] G. Qiu and S. B. Riffat, “Optimum tilt angle of solar collectors
models of diffuse solar radiation at high latitudes. and its impact on performance,” International Journal of Ambi-
Hence, further research is needed to better estimate the ent Energy, vol. 24, no. 1, pp. 13–20, 2003.
diffuse solar radiation.
[18] J. I. Yellott, “Utilization of sun and sky radiation for heating
cooling of buildings,” ASHRAE Journal, vol. 15, no. 12, pp. 31–
References 42, 1973.
[19] G. Lewis, “Optimum tilt of a solar collector,” Solar and Wind
[1] ASHRAE Handbook: HVAC Applications, ASHRAE, Atlanta, Technology, vol. 4, no. 3, pp. 407–410, 1987.
Ga, USA, 1999. [20] H. P. Garg, Advances in Solar Energy Technology, Reidel, Boston,
[2] M. Iqbal, “Estimation of the monthly average of the diffuse Mass, USA, 1987.
component of total insolation on a horizontal surface,” Solar [21] P. J. Lunde, Solar Thermal Engineering, John Wiley & Sons, New
Energy, vol. 20, no. 1, pp. 101–105, 1978. York, NY, USA, 1980.
[3] C. A. Gueymard, “Importance of atmospheric turbidity and [22] A. Angstrom, “Solar and terrestrial radiation,” Quarterly Journal
associated uncertainties in solar radiation and luminous effi- of the Royal Meteorological Society, vol. 50, pp. 121–126, 1924.
cacy modelling,” Energy, vol. 30, no. 9, pp. 1603–1621, 2005. [23] G. N. Tiwari and S. Suneja, Solar Thermal Engineering System,
[4] O. C. Vilela, N. Fraidenraich, and C. Tiba, “Photovoltaic Narosa Publishing, New Dehli, India, 1997.
pumping systems driven by tracking collectors. Experiments [24] I. Ulfat, F. Ahmad, and I. Siddiqui, “Determination of Angstrom
and simulation,” Solar Energy, vol. 74, no. 1, pp. 45–52, 2003. coefficient for the prediction of monthly average daily Gobal
[5] H. K. Elminir, A. E. Ghitas, F. El-Hussainy, R. Hamid, M. Solar Radiation Horizontal surface at Karachi, Pakistan,”
M. Beheary, and K. M. Abdel-Moneim, “Optimum solar flat- Karachi University Journal of Science, vol. 33, pp. 7–11, 2005.
plate collector slope: case study for Helwan, Egypt,” Energy [25] G. Stanhill and S. Moreshet, “Global radiation climate changes:
Conversion and Management, vol. 47, no. 5, pp. 624–637, 2006. the world network,” Climatic Change, vol. 21, no. 1, pp. 57–75,
[6] M. A. Ionită and C. Alexandru, “Dynamic optimization of the 1992.
tracking system for a pseudo-azimuthal photovoltaic platform,” [26] B. G. Liepert and G. J. Kukla, “Decline in global solar radiation
Journal of Renewable and Sustainable Energy, vol. 4, Article ID with increased horizontal visibility in Germany between 1964
053117, 15 pages, 2012. and 1990,” Journal of Climate, vol. 10, no. 9, pp. 2391–2401, 1997.
12 Journal of Renewable Energy

[27] S. A. Klein, “Calculation of monthly average insolation on tilted

surfaces,” Solar Energy, vol. 19, no. 4, pp. 325–329, 1977.
[28] R. Perez, P. Ineichen, R. Seals, J. Michalsky, and R. Stewart,
“Modeling daylight availability and irradiance components
from direct and global irradiance,” Solar Energy, vol. 44, no. 5,
pp. 271–289, 1990.
[29] L. Robledo and A. Soler, “Modelling irradiance on inclined
planes with an anisotropic model,” Energy, vol. 23, no. 3, pp. 193–
201, 1998.
[30] F. J. Olmo, J. Vida, I. Foyo, Y. Castro-Diez, and L. Alados-
Arboledas, “Prediction of global irradiance on inclined surfaces
from horizontal global irradiance,” Energy, vol. 24, no. 8, pp.
689–704, 1999.
[31] L. Wenxian, G. Wenfeng, P. Shaoxuan, and L. Enrong, “Ratios
of global radiation on a tilted to horizontal surface for Yunnan
Province, China,” Energy, vol. 20, no. 8, pp. 723–728, 1995.
[32] D. H. W. Li and T. N. T. Lam, “Determining the optimum
tilt angle and orientation for solar energy collection based
on measured solar radiance data,” International Journal of
Photoenergy, vol. 2007, Article ID 85402, 2007.
[33] H. D. Kambezidis, B. E. Psiloglou, and C. Gueymard, “Measure-
ments and models for total solar irradiance on inclined surface
in Athens, Greece,” Solar Energy, vol. 53, no. 2, pp. 177–185, 1994.
[34] T. Muneer, Solar Radiation and Daylight Models, Elsevier,
Berlin, Germany, 2nd edition, 2004.
[35] T. Muneer and M. M. Hawas, “Correlation between daily
diffuse and global radiation for India,” Energy Conversion and
Management, vol. 24, no. 2, pp. 151–154, 1984.
[36] J. K. Page, The Estimation of Monthly Mean Value of Daily
Short Wave Irradiation on Vertical and Inclined Surfaces from
Sunshine Records for Latitudes 60∘ N to 40∘ S. BS32, Department
of Building Science, University of Sheffield, Sheffield, UK, 1977.
[37] J. A. Duffie and W. A. Beckman, Solar Engineering of Thermal
Process, John Wiley & Sons, New York, NY, USA, 1991.
[38] J. LeQuere, “Rapport sur la comparaison des methods de calcul
des besoins de chauffage des logements,” Tech. Rep. CSTB TEA-
S 87, CSTB, Paris, France, 1980.
[39] J. E. Hay and J. A. Davies, “Calculation of the solar radiation
incident on an inclined surface,” in Proceedings of the 1st
Canadian Solar Radiation Data Workshop, J. E. Hay and T. K.
Won, Eds., pp. 59–72, Toronto, Canada, 1980.
[40] T. M. Klucher, “Evaluation of models to predict insolation on
tilted surfaces,” Solar Energy, vol. 23, no. 2, pp. 111–114, 1979.
[41] R. R. Perez, J. T. Scott, and R. Stewart, “An anisotropic model for
diffuse radiation incident on slopes of different orientations and
possible applications to CPCs,” in Proceedings of American Solar
Energy Society (ASES ’83), pp. 883–888, Minneapolis, Minn,
USA, 1983.
[42] R. Perez, R. Seals, P. Ineichen, R. Stewart, and D. Menicucci, “A
new simplified version of the perez diffuse irradiance model for
tilted surfaces,” Solar Energy, vol. 39, no. 3, pp. 221–231, 1987.
[43] T. Muneer, Solar Radiation and Daylight Models for the Energy
Efficient Design of Buildings, Architectural Press, Oxford, UK,
1997.
[44] E. Calabrò, “Determining optimum tilt angles of photovoltaic
panels at typical north-tropical latitudes,” Journal of Renewable
and Sustainable Energy, vol. 1, no. 3, Article ID 033104, 6 pages,
2009.
[45] S. Naihong, N. Kameda, Y. Kishida, and H. Sonoda, “Experi-
mental and Theoretical study on the optimal tilt of photovoltaic
panels,” Journal of Asian Architecture and Building Engineering,
vol. 5, pp. 399–405, 2006.
Journal of Journal of International Journal of

Rotating
Energy Wind Energy Machinery

The Scientific Journal of

Hindawi Publishing Corporation Hindawi Publishing Corporation

World Journal
Hindawi Publishing Corporation
Structures
Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of Journal of
Industrial Engineering Petroleum Engineering

Hindawi Publishing Corporation Hindawi Publishing Corporation

http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of
Solar Energy
Submit your manuscripts at
http://www.hindawi.com

Journal of

Fuels
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014

Engineering Photoenergy
Journal of  International Journal of
Advances in Advances in Journal of
Power Electronics
Hindawi Publishing Corporation Hindawi Publishing Corporation
High Energy Physics
Hindawi Publishing Corporation Hindawi Publishing Corporation
Combustion
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

International Journal of Journal of

Nuclear Energy Renewable Energy

International Journal of

Aerospace
Advances in
Tribology
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Science and Technology of
Nuclear Installations
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014