Download as pdf or txt
Download as pdf or txt
You are on page 1of 37

MECHANICAL ENGINEERING DEPT.

ME439: SOLAR ENERGY CONVERSION

Chapter 2 – Available Solar Radiation (2)


Course instructor: Prof. Mohammed A. Antar
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

In this and the following two sections we review methods for estimation of
the fractions of total horizontal radiation that are diffuse and beam.

The questions of the best methods for doing these calculations are not fully
settled.

A broader database and improved understanding of the data will probably


lead to improved methods.

In each section we review methods that have been published and then
suggest one for use.

The suggested correlations are in substantial agreement with other


correlations, and the set is mutually consistent.

2
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

The split of total solar radiation on a horizontal surface into its diffuse and
beam components is of interest in two contexts.

First, methods for calculating total radiation on surfaces of other orientation


from data on a horizontal surface require separate treatments of beam and
diffuse radiation (see Section 2.15).

Second, estimates of the longtime performance of most concentrating


collectors must be based on estimates of availability of beam radiation.

The present methods for estimating the distribution are based on studies of
available measured data; they are adequate for the first purpose but less than
adequate for the second.

3
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

The usual approach is to correlate Id/I, the fraction of the hourly radiation on a
horizontal plane which is diffuse, with kT = I / Io , the hourly clearness index.

Figure 2.10.1 shows a plot of diffuse fraction Id/I versus kT for Cape
Canaveral, Florida.

In order to obtain Id/I -versus- kT correlations, data from many locations


similar to that shown in Figure 2.10.1 are divided into “bins,” or ranges of
values of kT, and the data in each bin are averaged to obtain a point on the
plot.

4
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

A set of these points then is the basis of the correlation.

Within each of the bins there is a distribution of points; a kT of 0.5 may be


produced by skies with thin cloud cover, resulting in a high diffuse fraction, or
by skies that are clear for part of the hour and heavily clouded for part of the
hour, leading to a low diffuse fraction.

Thus the correlation may not represent a particular hour very closely, but over
a large number of hours it adequately represents the diffuse fraction.

Orgill and Hollands (1977) have used data of this type from Canadian
stations,
Erbs et al. (1982) have used data from four U.S. and one Australian station,
and
Reindl et al. (1990a), have used an independent data set from the United
States and Europe.

5
The three correlations are shown in Figure 2.10.2.
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

They are essentially identical, although they were derived from three separate
databases.

The Erbs et al. correlation (Figure 2.10.3) is11

6
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

For values of kT greater than 0.8, there are very few data. Some of the data
that are available show increasing diffuse fraction as kT increases above
0.8.

This apparent rise in the diffuse fraction is probably due to reflection of


radiation from clouds during times when the sun is unobscured but when
there are clouds near the path from the sun to the observer.
7

The use of a diffuse fraction of 0.165 is recommended in this region.


2.10 BEAM AND DIFFUSE COMPONENTS OF DAILY RADIATION

Studies of available daily radiation data have shown that the average fraction
which is diffuse, Hd/H, is a function of KT, the day’s clearness index.

The original correlation of Liu and Jordan (1960) is shown in Figure 2.11.1;
the data were for Blue Hill, Massachusetts.

Also shown on the graph are plots


of data
for a part of MENA from Stanhill
(1966),
for New Delhi from Choudhury
(1963),
for Canadian stations from Ruth and
Chant (1976) and Tuller (1976),
for Highett (Melbourne), Australia,
from Bannister (1969), and
for four U.S. stations from Collares-
8 Pereira and Rabl (1979a).
2.10 BEAM AND DIFFUSE COMPONENTS OF DAILY RADIATION

There is some disagreement, with differences probably due in part to


instrumental difficulties such as shading ring corrections and possibly in part
due to air mass and/or seasonal effects.

The correlation by Erbs (based on the same data set as is Figure 2.10.2) is
shown in Figure 2.11.2.

9
2.10 BEAM AND DIFFUSE COMPONENTS OF HOURLY RADIATION

A seasonal dependence is shown; the spring, summer, and fall data are
essentially the same, while the winter data show somewhat lower diffuse
fractions for high values of KT.

The season is indicated by the sunset hour angle ωs.

Equations representing this set of correlations are as follows:

For winter ωs ≤ 8 1.4°

For spring, summer, and fall ωs > 8 1.4°

10
Example 2.11.1

The day’s total radiation on a horizontal surface for St. Louis, Missouri (latitude
38.6°), on September 3 is 23.0 MJ/m2. Estimate the fraction and amount that is
diffuse.

Solution
For September 3, the declination is δ = 7°.

From Equation 1.6.10, the sunset hour angle is ωs = 95.6°.

From Equation 1.10.3, the day’s extraterrestrial radiation is Ho =33.3 MJ/m2.


Then

KT = H/Ho = 23.0/33.3 = 0.69

From Figure 2.11.2 or Equation 2.11.1b, Hd/H = 0.26, so an estimated 26%


of the day’s radiation is diffuse.

11 The diffuse energy is 0.26 × 23.0 = 6.0 MJ/m2.


2.10 BEAM AND DIFFUSE COMPONENTS OF MONTHLY
RADIATION
Charts similar to Figures 2.11.1 and 2.11.2 have been derived to show the
distribution of monthly average daily radiation into its beam and diffuse
components.

In this case, the monthly fraction that is diffuse, 𝐻ഥ 𝑑 /𝐻,


ഥ is plotted as a function
of monthly average clearness index, 𝐾 ഥ T ( = 𝐻/
ഥ 𝐻ഥ o).

The data for these plots can be obtained from daily data in either of two ways.

First, monthly data can be plotted by summing the daily diffuse and total
radiation data.

Second, as shown by Liu and Jordan (1960), a generalized daily Hd/H -versus-
KT curve can be used with a knowledge of the distribution of good and bad
days (the cumulative distribution curves of Figure 2.9.2) to develop the
monthly average relationships.
12
2.10 BEAM AND DIFFUSE COMPONENTS OF MONTHLY
RADIATION
ഥ 𝑑/𝐻
Figure 2.12.1 shows several correlations to 𝐻 ഥ versus 𝐾
ഥT .

13
2.10 BEAM AND DIFFUSE COMPONENTS OF MONTHLY
RADIATION
Erbs et al. (1982) developed monthly average diffuse fraction correlations
from the daily diffuse correlations of Figure 2.11.2.

As with the daily correlations, there is a seasonal dependence: the winter


curve lies below the other, indicating lower moisture and dust in the winter
sky with resulting lower fractions of diffuse.

14
2.10 BEAM AND DIFFUSE COMPONENTS OF MONTHLY
RADIATION

ഥ 𝑑/𝐻
The dependence of 𝐻 ഥ on 𝐾
ഥ T is shown for winter and for the other seasons
in Figure 2.12.2.

Equations for these correlations are as follows:



For ωs ≤ 8 1.4° and 0.3 ≤ KT ≤ 0.8


For ωs > 8 1.4° and 0.3 ≤ KT ≤ 0.8

15
Example 2.12.1

Estimate the fraction of the average June radiation on a horizontal surface that is
diffuse in Madison, Wisconsin.

Solution
ഥ for Madison is 23.0
From Appendix G, the June average daily radiation 𝐻
MJ/m2.

From Equation 1.10.3, for June 11 (the mean day of the month, n = 162,
from Table 1.6.1), when the declination is 23.1°, Ho = 41.8 MJ/m2.

ഥ T = 23.0/41.8 = 0.55.
Thus 𝐾

From Equation 1.6.10, ωs = 113.4°.

Then, using either Equation 2.12.lb or the upper curve from Figure 2.12.2,
ഥ 𝑑/𝐻
𝐻 ഥ = 0.38.
16
2.13 ESTIMATION OF HOURLY RADIATION FROM DAILY DATA

It may be necessary to start with daily data and estimate hourly values from
daily numbers.

Statistical studies of the time distribution of total radiation on horizontal


surfaces through the day using monthly average data for a number of stations
have led to generalized charts of rt the ratio of hourly total to daily total
radiation, as a function of day length and the hour in question:

Figure 2.13.1 shows such a chart, adapted from Liu and Jordan (1960) and
based on Whillier (1956, 1965) and Hottel and Whillier (1958).

The hours are designated by the time for the midpoint of the hour, and days
are assumed to be symmetrical about solar noon.

A curve for the hour centered at noon is also shown.


17
2.13 ESTIMATION OF HOURLY RADIATION FROM DAILY DATA

Day length N can be


calculated from Equation
1.6.11 or it can be
estimated from Figure
1.6.3.

Thus, knowing day length


(a function of latitude φ
and declination δ) and
daily total radiation, the
hourly total radiation for
symmetrical days can be
estimated.

18
2.13 ESTIMATION OF HOURLY RADIATION FROM DAILY DATA

The figure is based on long-term averages and is intended for use in


determining averages of hourly radiation.

The curves of Figure 2.13.1 are represented by the following equation from
Collares Pereira and Rabl (1979a):

I/H =

The coefficients a and b are given by

In these equations ω is the hour angle in degrees for the time in question (i.e.,
the midpoint of the hour for which the calculation is made) and ωs is the
19
sunset hour angle.
Example 2.13.1

What is the fraction of the average January daily radiation that is received at
Melbourne, Australia, in the hour between 8:00 and 9:00?

Solution
For Melbourne, φ = 38°.

From Table 1.6.1 the mean day for January is the 17th.

From Equation 1.6.1 the declination is δ = -20.9°.

From Equation 1.6.11 the day length is 14.3 h.

From Figure 2.13.1, using the curve for 3.5 h from solar noon, at a day
length of 14.3 h, approximately 7.8% of the day’s radiation will be in that
hour.

Or Equation 2.13.2 can be used; with ωs = 107° and ω = -52.5°, the result is
20
rt = 0.076.
Example 2.13.2

The total radiation for Madison on August 23 was 31.4 MJ/m2. Estimate the
radiation received between 1 and 2 PM.

Solution
For August 23, δ = 11° and φ for Madison is 43°.

From Figure 1.6.3, sunset is at 6:45 PM and day length is 13.4 h.

Then from Figure 2.13.1, at day length of 13.4 h and mean of 1.5 h from
solar noon, the ratio hourly total to daily total rt = 0.118.

The estimated radiation in the hour from 1 to 2 PM is then 3.7 MJ/m2.

(The measured value for that hour was 3.47 MJ/m2.)

21
2.13 ESTIMATION OF HOURLY RADIATION FROM DAILY DATA

Figure 2.13.2 shows a related set of


curves for rd, the ratio of hourly diffuse
to daily diffuse radiation, as a function of
time and day length.

In conjunction with Figure 2.11.2, it can


be used to estimate hourly averages of
diffuse radiation if the average daily total
radiation is known:

These curves are based on the


assumption that Id/Hd is the same as Io/Ho
and are represented by the following
equation from Liu and Jordan (1960):

22
Example 2.13.3

From Appendix G, the average daily June total radiation on a horizontal plane in
Madison is 23.0 MJ/m2. Estimate the average diffuse, the average beam, and the
average total radiation for the hours 10 to 11 and 1 to 2.

Solution
The mean daily extraterrestrial radiation Ho for June for Madison is 41.7
MJ/m2 (from Table 1.10.1 or Equation 1.10.3 with n = 162), ωs = 113°, and
the day length is 15.1 h (from Equation 1.6.11).

Then (as in Example 2.12.1), 𝐾ഥ T =𝐻/H


ഥ o= 0.55. From Equation 2.12.1,
ഥ 𝑑/𝐻=
𝐻 ഥ 0.38, and the average daily diffuse radiation is 0.38 × 23.0 = 8.74
MJ/m2.

Entering Figure 2.13.2 for an average day length of 15.1 h and for 1.5 h
from solar noon, we find rd = 0.102.

(Or Equation 2.13.4 can be used with ω = 22.5° and ωs = 113° to obtain rd =
23 0.102.)
Thus the average diffuse for those hours is 0.102 × 8.74 = 0.89 MJ/m2.
Example 2.13.3

From Appendix G, the average daily June total radiation on a horizontal plane in
Madison is 23.0 MJ/m2. Estimate the average diffuse, the average beam, and the
average total radiation for the hours 10 to 11 and 1 to 2.

Solution

From Figure 2.13.1 (or from Equations 2.13.1 and 2.13.2) from the curve for
1.5 h from solar noon, for an average day length of 15.1 h, rt = 0.108 and
average hourly total radiation is 0.108 × 23.0 = 2.48 MJ/m2.

The average beam radiation is the difference between the total and diffuse,
or 2.48 - 0.89 = 1.59 MJ/m2.

24
2.14 RADIATION ON SLOPED SURFACES

We turn next to the general problem of calculation of radiation on tilted


surfaces when only the total radiation on a horizontal surface is known.

For this we need the directions from which the beam and diffuse components
reach the surface in question.

Section 1.8 (Rb = Gb,T / Gb) dealt with the geometric problem of the direction
of beam radiation.

The direction from which diffuse radiation is received, that is, its distribution
over the sky dome, is a function of conditions of cloudiness and atmospheric
clarity, which are highly variable.

Clear-day data suggest a diffuse radiation model as being composed of three


parts.

25
2.14 RADIATION ON SLOPED SURFACES

The first is an isotropic part, received uniformly from the entire sky dome.

The second is circumsolar diffuse, resulting from forward scattering of


solar radiation and concentrated in the part of the sky around the sun.

The third, referred to as horizon brightening, is concentrated near the


horizon and is most pronounced in clear skies.

Figure 2.14.2 shows schematically these three parts of the diffuse radiation

26
2.14 RADIATION ON SLOPED SURFACES

It is necessary to know or to be able to estimate the solar radiation incident


on tilted surfaces such as solar collectors, windows, or other passive system
receivers.

The incident solar radiation is the sum of a set of radiation streams including
beam radiation, the three components of diffuse radiation from the sky, and
radiation reflected from the various surfaces “seen” by the tilted surface.

The total incident radiation on this surface can be written as

where the subscripts iso, cs, hz, and refl refer to the isotropic, circumsolar,
horizon, and reflected radiation streams.

For a surface (a collector) of area Ac, the total incident radiation can be
expressed in terms of the beam and diffuse radiation on the horizontal surface
27
and the total radiation on the surfaces that reflect to the tilted surface.
2.14 RADIATION ON SLOPED SURFACES

The terms in Equation 2.14.1 become

The first term is the beam contribution.


The second is the isotropic diffuse term, which includes the product of sky
area As (an undefined area) and the radiation view factor from the sky to the
collector Fs-c.
The third is the circumsolar diffuse, which is treated as coming from the
same direction as the beam.
The fourth term is the contribution of the diffuse from the horizon from a
band with another undefined area Ahz.
The fifth term is the set of reflected radiation streams from the buildings,
fields, and so on, to which the tilted surface is exposed.
The symbol i refers to each of the reflected streams: Ii is the solar radiation
incident on the ith surface, ρi is the diffuse reflectance of that surface, and Fi-c
is the view factor from the ith surface to the tilted surface.
28 It is assumed that the reflecting surfaces are diffuse reflectors; specular

(mirror) reflectors require a different treatment.


View Factor Relations

An important view factor relation is suggested by the above equations.

In particular, equating the integrals appearing in these equations, it follows


that
Ai Fi j  Aj Fj i
This expression, termed the reciprocity relation, is useful in determining one
view factor from knowledge of the other.

Another important view factor relation pertains


to the surfaces of an enclosure (see Figure).

From the definition of the view factor, the


summation rule
N

F
j 1
ij 1

29

may be applied to each of the N surfaces in the enclosure.


View Factor Relations

To illustrate the foregoing procedure, consider a simple, two-surface


enclosure involving the spherical surfaces of this Figure.

Although the enclosure is characterized by N 2 = 4 view factors (F11, F12, F21,


F22), only N(N-1)/2 = 1 view factor need be determined directly.

In this case such a determination may be made by inspection.

In particular, since all radiation leaving the inner surface must reach the outer
surface, it follows that F12 = 1.

The same may not be said of radiation leaving


the outer surface, since this surface sees itself.

However, from the reciprocity relation and


summation rule, we obtain

30
2.14 RADIATION ON SLOPED SURFACES

In general, it is not possible to calculate the reflected energy term in detail, to


account for buildings, trees, and so on, the changing solar radiation incident
on them, and their changing reflectances.

Standard practice is to assume that there is one surface, a horizontal, diffusely


reflecting ground, large in extent, contributing to this term.

In this case, Ii is simply I and ρi becomes ρg, a composite “ground”


reflectance.

Equation 2.14.2 can be rewritten in a useful form by interchanging areas and


view factors (since the view factor reciprocity relation requires that, e.g.,
AsFs-c = AcFc-s).

31

(AsFs-c = AcFc-s)
2.14 RADIATION ON SLOPED SURFACES

𝐼𝜌𝑔 𝐴𝑔 𝐹𝑔−𝑐
(AsFs-c = AcFc-s)
is (AhzFhz-c = AcFc-hz)
𝐼𝜌𝑔 𝐴𝑐 𝐹𝑐−𝑔

32
2.14 RADIATION ON SLOPED SURFACES

This eliminates the undefined areas As and Ahz.

The area Ac appears in each term in the equation and cancels.

The result is an equation that gives IT in terms of parameters that can be


determined either theoretically or empirically:

This equation, with variations, is the basis for methods of calculating IT that
are presented in the following sections.

When IT has been determined, the ratio of total radiation on the tilted surface
to that on the horizontal surface can be determined. By definition,

33
2.14 RADIATION ON SLOPED SURFACES- Isotropic Sky

An improvement on this model, the isotropic diffuse model, was derived by


Liu and Jordan (1963).

The radiation on the tilted surface was considered to include three


components: beam, isotropic diffuse, and solar radiation diffusely reflected
from the ground.

The third and fourth terms in Equation 2.14.3 are taken as zero as all diffuse
radiation is assumed to be isotropic.

A surface tilted at slope β from the horizontal has a view factor to the sky
Fc-s = (1 + cos β)/2.

(If the diffuse radiation is isotropic, this is also Rd, the ratio of diffuse on the
34
tilted surface to that on the horizontal surface.)
2.14 RADIATION ON SLOPED SURFACES

The tilted surface has a view factor to the ground Fc-g = (1 - cos β)/2, and if the
surroundings have a diffuse reflectance of ρg for the total solar radiation, the
reflected radiation from the surroundings on the surface will be I ρg (1 - cos β)/2.

Thus Equation 2.14.3 is modified to give the total solar radiation on the tilted
surface for an hour as the sum of three terms:

and by the definition of R, R = IT/I

35
Example 2.15.1

Using the isotropic diffuse model, estimate the beam, diffuse, and ground-
reflected components of solar radiation and the total radiation on a surface
sloped 60° toward the south at a latitude of 40° N for the hour 9 to 10 AM on
February 20. Here I = 1.04 MJ/m2 and ρg = 0.60.
Solution
For this hour, Io = 2.31 MJ/m2, so kT = 1.04/2.31 = 0.45.
From the Erbs correlation (Equation 2.10.1) Id/I = 0.757. Thus

Id = 0.757 × 1.04 = 0.787 MJ/m2


1-0.757 = 0.243
Ib = 0.243 × 1.04 = 0.253 MJ/m2

The hour angle ω for the midpoint of the hour is -37.5°.


The declination δ = -11.6°.
Then for this south-facing surface

36
Example 2.15.1
Solution
Equation 2.15.1 gives the three radiation streams and the total:

Thus the beam contribution is 0.433 MJ/m2, the diffuse is 0.590 MJ/m2, and
the ground reflected is 0.156 MJ/m2.

The total radiation on the surface for the hour is 1.18 MJ/m2.

There are uncertainties in these numbers.

This example is for a surface with a surface azimuth angle of zero.

The model (Equation 2.5.1) is applicable for surfaces of any orientation,


provided the correct relationship for Rb is used.
37

You might also like