Anda di halaman 1dari 8

Renewable Energy 63 (2014) 90e97

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Articial neural networks for the performance prediction of large solar


systems
S.A. Kalogirou a, *, E. Mathioulakis b, V. Belessiotis b
a
b

Department of Mechanical Engineering and Materials Science and Engineering, Cyprus University of Technology, P.O. Box 50329, 3603 Limassol, Cyprus
Solar & Other Energy Systems Laboratory, DEMOKRITOS National Center for Scientic Research, 15310 Agia Paraskevi, Attikis, Greece

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 15 May 2013
Accepted 30 August 2013
Available online 27 September 2013

In this paper, articial neural networks (ANNs) are used for the performance prediction of large solar
systems. The ANN method is used to predict the expected daily energy output for typical operating
conditions, as well as the temperature level the storage tank can reach by the end of the daily operation
cycle. These are considered as the most important parameters for the user. Experimental measurements
from almost one year (226 days) have been used to investigate the ability of ANN to model the energy
behavior of a typical large solar system. From the results, it can be concluded that the ANN effectively
predicts the daily energy performance of the system; the statistical R2-value obtained for the training and
validation data sets was better than 0.95 and 0.96 for the two performance parameters respectively. The
data used in the validation were completely unknown to the ANN, which proves the ability of the ANN to
give good predictions on completely unknown data. The results obtained from the method were also
compared to the inputeoutput model predictions with good accuracy whereas multiple linear regression
could not give as accurate results. Additionally, the network was used with various combinations of input
parameters and gave results of the same order of magnitude as the suggested method, which prove the
robustness of the method. The advantages of the proposed approach include the simplicity in the
implementation, even when the characteristics of the system components are not known, as well as the
potential to improve the capability of the ANN to predict the performance of the solar system, through
the continuous addition of new data collected during the operation of the system.
2013 Elsevier Ltd. All rights reserved.

Keywords:
Articial neural networks
Large solar thermal systems
Daily energy output
Maximum temperature of storage tank

1. Introduction
The development of large capacity installations for the exploitation of solar thermal energy can be characterized as critical for
the broadening of solar thermal energy applications [1]. Such an
expansion though, seems to come up against the suspicion of the
potential users regarding the expected energy benet, and consequently the reciprocity of the required investment. From this point
of view, the realistic evaluation of the energy performance of large
solar thermal systems (LSTS) constitutes a critical parameter, not
only for the candidate investor, but also for any type of supporting
actions provided by the state.
In the case of small, standardized, factory-made domestic
systems, the demand for reliable information regarding the expected energy output is satised by national or regional

* Corresponding author. Tel.: 357 2500 2621; fax: 357 2500 2637.
E-mail addresses: Soteris.kalogirou@cut.ac.cy, skalogir@spidernet.com.cy
(S.A. Kalogirou).
0960-1481/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.renene.2013.08.049

certication systems, the implementation of which is based on the


existing testing and requirements standards (EN12976-1, Solar
Keymark). A LSTS though is one of a kind, designed in accordance
with the satisfaction of the particular requirements for a specic
application. The discussion thus, concerns custom-made, nonstandardized systems, where the only available information for
the energy performance comes out from the initial long-term
performance prediction made in the study phase. Often, this information is characterized by signicant uncertainties, which can
be attributed to the simplied working hypotheses of the calculation methods, the lack of information regarding the characteristics of the subsystems employed (collectors, storage tank,
controller, hydraulics), or even more to the insufcient knowledge
of the actual load prole [2,3].
Even though each LSTS is unique, the demand for reliable information regarding its actual energy performance is of critical
importance, mainly for the following reasons:
- The comparison of performance prediction, made in the study
phase, with the actual performance of the system, will assist the

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

user to shape a realistic view for the reciprocity of the investment required, and, eventually, to redesign some elements of
the energy system (e.g. changing of the capacity of the auxiliary
sources).
- The analysis of potential signicant deviations between the
predictions and the actual performance, may lead to the
detection of failures related to the implementation phase, and to
the elaboration of corrective actions.
- The continuous comparison of actual energy performance to the
predicted one, or the one which had been initially veried, can
be proved to be valuable for the on time detection and correction of potential malfunctions which have appeared in due
course. Such a comparison could be implemented on an automated basis by the control system, which would also undertake
the activation of the proper alarms in case of a problem.
Moreover, the need for extending the certication of solar
thermal systems on large capacity custom-made systems has
turned the interest on methodological approaches which allow the
reliable assessment of the expected energy benet [4,5]. Within
this context, the discussion of the methodologies suitable for the
verication of energy performance of LSTS has been intensied in
recent years [6,7].
Thus, the comparison between expected and actual performance presupposes the existence of a simulation tool, which would
allow the reliable estimation of the expected energy benet of a
LSTS, as this system has been nally implemented. Such a tool can
belong in one of the following two main categories:
- Models which allow the dynamic simulation of the system
behavior on the basis of information related to its structure and
the characteristics of the individual subcomponents, e.g.
TRNSYS [2,3]. The quality of the predictions by such models is a
function of the reliability of the available information, noting the
usual signicant deviations between the nal actual conguration of a LSTS and the working hypotheses of the initial study.
These concern mainly the weather data used in the simulation,
usually in the form of a Typical Meteorological Year (TMY), and
the load prole assumed, which strongly depends on the habits
of the user/s and the type of application (hotel, industry, etc.). An
additional difculty is related to the requirement for specialized
knowledge and adequate experience by the personnel responsible for the operation of the specic software.
- Models, which treat the LSTS as a black box, and their application, can be performed on two distinct phases. In the rst phase,
i.e., training, experimental measurements taken through operation of the system are exploited, in order to shape an adequate
picture for the behavior of the system (mapping). The calculation procedure uses experimental information in order to
elaborate a kind of energy identity for the system, usually under
the form of coefcients of one or more characteristic equations.
Advantages of this kind of models include the relatively simple
use and the realistic assessment of the performance of the
system as it has been nally congured. As a disadvantage, one
may consider the inability to perform simulation for congurations or conditions of use signicantly different from the ones
used during the training phase [6].
Another factor, which inuences the quality of results of the
discussed models, is the operating conditions assumed. Given the
stochastic variation of meteorological conditions and their large
inuence on the energy behavior of the solar system, in order the
discussed comparison between expected and actual performance
to be realistic, this comparison has to be performed for certain
reference conditions, the same as the ones used during the

91

planning phase. Actually, this concerns not only the meteorological


conditions, for which the Typical Meteorological Year is in most
cases used, but also the load prole, which has to be as close as
possible to the conditions of use.
The proposed work aims to present a new approach, which
belongs on the second category of models, and it is based on the
availability of short term, relatively simple measurements, as well
as on the exploitation of Articial Neural Network (ANN) method.
The proposed approach is quite simple to use, and can be easily
implemented for the verication of the predictions of the initial
study, as well as for the real-time inspection for potential malfunctions [8].
The structure of the paper is as follows: In Section 2 the concept
of Articial Neural Networks is presented and their potential for the
simulation of energy processes is discussed. In Section 3 the
experimental set up is discussed, as well as the protocol of the
elaborated measurements, while in Section 4 the ANN performance
is analysed and the potential of implementing ANN in actual LSTS
applications is discussed, and nally, in Section 5 the conclusions of
the overall investigation are presented.
2. ANN modeling of energy systems
Although the concept of articial neural network (ANN) analysis
has been discovered nearly 60 years ago, it is only in the last 30
years that application software has been developed to handle
practical problems. ANNs are good for some tasks while lacking in
some others. Specically, they are good for tasks involving incomplete data sets, fuzzy or incomplete information, and for highly
complex and ill-dened problems, where humans usually decide
on an intuitive basis [9].
ANNs have been applied successfully in various elds of mathematics, engineering, medicine, economics, meteorology, psychology, neurology, and many others. Some of the most important ones
are; pattern, sound and speech recognition, analysis of medical
signatures, identication of military targets and of explosives in
passenger suitcases. They have also been used in weather and
market trends forecasting, prediction of mineral exploration sites,
electrical and thermal load prediction, adaptive and robotic control
and many others [10].
Articial neural networks are systems of weight vectors, whose
component values are established through various machinelearning algorithms, which take a linear set of pattern inputs and
produce a numerical pattern representing the actual output. ANNs
mimic somewhat the learning process of the human brain. Instead
of complex rules and mathematical routines, ANNs are able to learn
key information patterns within a multi-information domain. In
addition, inherently noisy data does not seem to present a problem,
as ANNs are tolerant in noise variations [11].
Articial neural networks differ from the traditional modeling
approaches in that they are trained to learn solutions rather than
being programmed to model a specic problem in the normal way.
Table 1
Technical specications of the system employed.
Parameter

Value

Units

Collector aperture area, AC


Maximum collector efciency, no
Overall collector heat loss coefcient, UC
Effective thermal capacity of collector, (MC)C
Water content in collector, WW
Number of collectors in the eld, NC
Heat-exchanger overall heat transfer coefcient, (UA)ex
Storage tank volume, VS
Overall storage tank heat loss coefcient, (AU)S

2.59
0.732
4.455
29
5.8
48
13,300
32,000
150

m2
[e]
Wm2 K1
kJK1
kg
[e]
W K1
l
W K1

92

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

Fig. 1. The solar system considered. (a) Schematic diagram of the solar system. (b) Position of the various thermocouples in the storage tank.

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

93

Such work is never done before, as these systems were not studied
in the appropriate depth.
3. The experimental setup

Fig. 2. Articial neural network architecture employed.

They are usually used to address problems that are intractable or


cumbersome to solve with traditional methods. Neural networks
are widely accepted as a technology offering an alternative way to
tackle complex and ill-dened problems. They can learn from examples, are fault-tolerant in the sense that they are able to handle
noisy and incomplete data, are able to deal with non-linear problems, and once trained, can perform predictions at a very high
speed. ANNs have been used in many engineering applications,
such as in control systems, in classication, and in modeling complex process transformations [9,11].
The neural network method falls under the generic non-linear
analogue techniques and has revived the idea of analogue data
analysis. Neural networks have been used in the past for predicting
the performance of domestic size solar water heaters [12] and for the
prediction of the performance of long-term performance of forced
circulation domestic solar water heaters [13]. They have also been
used in the long-term performance of thermosiphon domestic solar
water heating systems [14] and for the modeling of the starting-up of
a solar steam generator [15]. A review of applications of articial
neural networks (ANNs) in energy systems is presented in Refs. [9e
11]. Morabito et al. [16] used other approaches to ANNs for predictions in energy-related engineering applications with special
reference to the treatment of input selection and uncertainty.
The objective of this work is to train a suitable ANN to estimate
two of the main parameters of large solar water heating systems.

For the needs of the present investigation, experimental data


concerning operation of a typical LSTS for a duration of almost one
year have been used. The characteristics of the solar system
employed are presented in Table 1.
The experimental system is a typical LSTS, with a stratied solar
tank and a ratio of storage volume to the overall collector surface:
Vs =AF z257 l=m2 . The system is shown schematically in Fig. 1a. It is
noted that the experimental data collected from the LSTS have been
used in the past for the validation of the InputeOutput modeling
approach, suggested by the standard [6,7].
For the needs of the present investigation, the temperature Ts of
the storage tank water has been controlled through the use of a
proper water-to-air heat-exchanger (cooling tower, shown in
Fig. 1a). Then, and before the beginning of the daily operation, the
tank water is fully homogenized through a proper mixing pump.
Measurements were implemented on a continuous basis and
included the following quantities, which can be used for the characterization of the system behavior:
- Solar radiation at the collector level, G.
- Ambient temperature, Ta.
- Temperatures at the inlet and outlet of the storage tank, on the
load side, Tin and Tout respectively.
_
- Draw-off ow rate of the heat transfer uid, on the load side, m.
- The temperature of the storage tank water in various points,
distributed within the overall height and width of the tank (see
Fig. 1b).
Based on the above quantities, some derived ones, which characterize the energy behavior of a LSTS have been calculated. The
selection of these quantities has been performed according to the
easiness of the determination of their values, and, mainly, due to
the fact that the correlation of the one to the other recommends a
simple and functional model for a solar thermal system, as it has
been documented in previous investigations [6,17]. The derived
quantities, which have been exploited within the present investigation, are:
- The daily total incident radiation at the collector level, H, which
corresponds to the solar energy the LSTS can turn into exploitable thermal energy for the user.

Fig. 3. Comparison between actual (measured) and ANN predicted values for Tsmax for the training data set.

94

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

Fig. 4. Comparison between actual (measured) and ANN predicted values for Q for the training data set.

- The daytime average ambient temperature, T a , which inuences


to a signicant degree the level of system thermal losses to the
environment.
- The storage tank water temperature at the start of the day, Tsin,
which represents the energy level of the tank by the beginning
of the day.
- The daily measured energy output of the system, Q, a quantity
that can be characterized as the most interesting one for the
user.
- The mean storage tank water temperature at the end of the day,
Ts,f, which represents the energy level of the tank by that time,
and can be directly correlated to the daily energy output Q.
- The maximum temperature of water in the tank at the end of the
day, Ts,max, which, in combination with the water temperature at
the end of the day, Ts,f, provides an indication for the temperature level the tank has reached, as well as for its degree of
stratication.
The average water temperature at the tank has been measured
with the help of 46 calibrated thermocouples, properly distributed
within the overall width and height of the tank. The determination
of the daily measured energy output of the system, Q, has been
based on the draw-off of the stored energy by the end of the day,
through the recording of the water temperatures Tin and Tout at the
inlet and outlet of the tank respectively, and of the volumetric ow
_
rate m.
The accumulated energy in the tank by the end of the day was
drawn off with the help of a cooling tower, simulating the load. In
this way, it has been possible to regulate the level of draw-off and
the level of energy remaining in the tank, and control the temperature of the tank at the beginning of the next day. Thus, a wide
range of system operation temperature level has been included in
the investigation.
From the whole set of available measurements, 226 valid days
have been selected, while the rest ones have been rejected for
various reasons (malfunctions, raining days, etc.). The measurements concerned daily operation of the system, with the beginning
of operation 6 h before solar noon and the end of operation 6 h after
solar noon, according to the usual practice for factory-made thermal solar systems [17,18].
It has been attempted to cover an adequate range of solar radiation, characterized by a satisfactory distribution of high radiation days (w25 MJ/m2), and days presenting a signicant fraction of
cloudiness (w8 MJ/m2). Moreover, days corresponding to different
seasons of the year have been included in the set of measurements,

aiming to the investigation of potential seasonal effects on the nal


result, which mainly is related to the duration of the day and the
altitude angle of the sun.
4. Results e discussion
According to Haykin [19], a neural network is a massively
parallel-distributed processor that has a natural propensity for
storing experiential knowledge and making it available for use. It
resembles the human brain in two respects: the knowledge is acquired by the network through a learning process, and inter-neuron
connection strengths known as synaptic weights are associated
with the knowledge.
ANN models represent a new method in system prediction. An
ANN operates like a black box model, requiring no detailed information about the system. Instead, it learns the relationship between the input parameters and the controlled and uncontrolled
variables by studying previously recorded data, similar to the way
in which a non-linear regression might perform. An advantage of
using ANNs is their ability to handle large and complex systems
with many interrelated parameters [9e11].
A neural network consists of a number of processing elements
called neurons, each of which has many inputs but only one output.
In a typical network, there are three layers of neurons, i.e., an input
layer which receives input from the outside world, a hidden layer or
layers which receive inputs from the input layer neurons, and an
output layer which receives inputs from the hidden layers and
passes its output to the outside world and in some cases back to the
preceding layers. The strength of the network lies in the interconnections between the neurons, which is modied during
training. The training is done by exposing the network to a specic
data set of information and by applying a training algorithm to
enable the network to produce the desired output [9].
The input parameters used in this work on the neural network
are the following:
Table 2
Evaluation of the performance of the trained ANN.
Range of accuracy

Up to 10%
10e20%
20e30%
More than 30%

Percent of patterns belonging in the category


Tsmax

95.45%
4.55%
0%
0%

59.10%
27.25%
9.10%
4.55%

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

95

Fig. 5. Comparison between actual (measured) and ANN predicted values for Tsmax for the validation data set.

- Taav Daytime average ambient temperature ( C)


- H Daily total incident radiation at the collector level (MJ/m2)
- Tsin Storage tank water temperature at the start of the day ( C)
These are easily obtained data, which are based mostly on the
prevailing weather conditions and the initial state of the system.
Other combinations of input parameters will be tried later to prove
the robustness of the ANN prediction method.
Two types of output parameters were predicted, which are the
most important for the evaluation of the performance of a large
solar water heating system. These are:
- Q Daily energy output of the system (MJ).
- Tsmax Maximum water temperature in the storage tank at the
end of the day ( C).
Various network architectures have been investigated to nd
the one that could provide the best overall performance. The architecture, among those tested, that gave the best results and was
adopted for the present work, is shown in Fig. 2. This architecture
has been used in a number of engineering problems for modeling
and prediction, with very good results, and it is a feedforward architecture composed of ve slabs, three of which are hidden. There
are different activation functions in each slab, as shown in Fig. 2.
Different activation functions were applied to the hidden layer
slabs in order to detect different features in a pattern processed

through the network. Three element inputs have been used corresponding to the values of the input parameters listed above and
the output parameters are two, Q and Tsmax. The learning procedure
was implemented by using the back-propagation algorithm. For the
training of the network, a learning rate and a momentum factor
needs to be specied by the user. Both of these constant terms are
specied at the start of the training cycle and determine the speed
and stability of the network. For this purpose, the learning rate was
set to a constant value of 0.1 and the momentum factor to 0.3. The
weights were initialized to a value of 0.3. The back-propagation
learning algorithm was used, described in Refs. [12,13]. The originators of back-propagation are Rumelhart [20] and Werbos [21].
As indicated before a total of 226 test patterns were available.
These concern actual measurements carried out as described in
Section 3 of this paper. From these patterns 184 were used for the
training of the ANN, and the rest 42 (20%) were randomly selected
to be used as test (20) and validation (22) patterns. Test patterns are
used during training to evaluate the training accuracy whereas the
validation patterns are used after the training phase to evaluate the
nal model accuracy.
The training was stopped when the average error, obtained by
comparing the actual and the ANN modeled data, remained constant for 200,000 events; i.e., about 1087 iterations through all data
(epochs) in the training dataset. This is considered a good value,
enabling the network to learn the input patterns satisfactorily and
to give good predictions while avoiding overtraining, which will

Fig. 6. Comparison between actual (measured) and ANN predicted values for Q for the validation data set.

96

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97

Actual energy output

Training

Network predicted energy output

1800

Qsim

Daily energy output (MJ)

1600
1400
1200
1000
800
600
400
200
0
1

21

41

61

81

101

121

141

161

181

201

Patterns
Fig. 7. Comparison between actual (measured), ANN predicted values and inputeoutput method (Qsim) for Q for the training data set.

enable the ANN to map the data more accurately but lose the ability
of the ANN to generalize, i.e., to give good predictions for unknown
data sets.
The correlation coefcient obtained between the predicted and
training data set is 0.9546 and 0.9681 for the maximum water
temperature in the storage tank at the end of the day (Tsmax) and
the daily energy output of the system (Q) respectively. Both values
are very close to 1, indicating a very good mapping of the data.
Comparative plots of the actual (measured) and ANN predicted
values for the two output parameters for the training data set are
shown in Figs. 3 and 4 respectively. It should be noted that multiple
linear regression tried on the same data set give correlations of
0.9232 and 0.9382 for the above two parameters respectively,
which cannot be considered adequate.
Once a satisfactory degree of inputeoutput mapping was achieved, the network training was frozen and a set of completely
unknown test data was applied for verication. The validation of
the network was performed by using the unknown data for 22
cases of the validation data set. The correlation coefcients for the
unknown cases were of the same order of magnitude as in the
training phase and were equal to 0.9625 and 0.9560 for the two
output parameters respectively. The range of prediction accuracy
for the various patterns is shown in Table 2.
Comparative plots of the actual (measured) and ANN predicted
values for the two output parameters for the validation data set,

which includes data that are completely unknown to the ANN, are
shown in Figs. 5 and 6 respectively.
As can be seen from Figs. 5 and 6, the performance for both
output parameters is very satisfactory.
It should be noted that two different neural networks, one for
each output parameter, of the same architecture as shown in Fig. 2
but with one output neuron only give very comparable results, so
for the sake of compactness of the model the one ANN model with
two outputs as described before is preferred.
Various other tests have been performed to prove the
robustness of the method. The rst method concern the inpute
output modeling method described in Belessiotis et al. [7], which
is used both for the training and validation data sets to predict
the daily energy output of the system, Q. The results are shown in
Figs. 7 and 8 respectively. As can be seen from the graphs the
ANN model can give as accurate results as a simulation program
(Qsim) used for the modeling of the data. In fact the inpute
output method give slightly inferior t of the data with R2-values
equal to 0.9273 and 0.9327 for the training and validation data
sets respectively as opposed to values of about 0.96 for the ANN
(as indicated before).
Finally, various other combinations of input data were tried to
investigate the accuracy of daily energy output predictions. These
are shown in Table 3 together with the input data used in each case
and the correlation obtained. For all these models, the same

Validation

Daily energy output (MJ)

Actual energy output

1600

Network predicted energy output

1400

Qsim

1200
1000
800
600
400
200
0
1

10 11 12 13

14 15 16

17 18 19

20 21 22

Patterns
Fig. 8. Comparison between actual (measured), ANN predicted values and the inputeoutput method (Qsim) for Q for the validation data set.

S.A. Kalogirou et al. / Renewable Energy 63 (2014) 90e97


Table 3
ANN prediction accuracy for various combinations of input parameters.
Model

Input data

R2-value

1
2
3
4

Taav, H, Tsin
H, DT
Tsm, H, DT
Tsmax, H, DT

0.9324
0.9310
0.9331
0.9325

Where: DT is temperature difference (Ta  Tsin); Tsm is the mean storage tank
temperature; Other symbols are as described before.

architecture shown in Fig. 2, was used and the output in all cases
was Q, the daily energy output of the solar system.
As shown, all cases investigated give the same order to accuracy as the method suggested in this paper, which is based on
easily obtainable parameters without the need to carry out extra
estimations, like for DT or Tsm, and also gives slightly better accuracy with R2-values of about 0.96. This exercise however
proves the robustness of the method to provide reliable
predictions.
5. Conclusions
In the present paper, the use of ANN for the performance prediction of large solar systems is discussed. The analysis is concentrated on the two efciency indices, which are considered as the
most important for the user. These indices are the expected daily
energy output for typical operation conditions, as well as the
temperature level the storage tank can reach by the end of the daily
operation cycle.
Experimental measurements from almost one year (226 days)
have been used to investigate the ability of ANN to model the energy behavior of a typical large solar system. From the results
presented, it can be concluded that the ANN effectively predicts the
daily energy performance of the system; the statistical R2-value
obtained for the training data set was better than 0.95 and 0.96 for
the two performance parameters respectively. The validation also
included investigation of the sufciency of the trained ANN when
having to cope with unknown data, i.e., data the network has not
seen before. The results obtained from the method were also
compared to the inputeoutput model predictions with good accuracy. Additionally, the network was used with various combinations of input parameters and gave results of the same order of
magnitude as the suggested method, which prove the robustness of
the method.
The advantages of the proposed approach include the simplicity
in the implementation, even when the characteristics of the system
components are not known, as well as the potential to improve the
capability of the ANN to predict the performance of the solar system, through the continuous integration of new data.

97

The proposed approach can be used for the comparison of


design performance predictions with the actual performance of the
system, as well as for the early detection of potential malfunctions.
Finally, it should be noted that for each large solar system a new
ANN need to be trained using data collected for a number of days as
described in this paper.
References
[1] ESTIF. Solar thermal action plan for Europe heating & cooling from the sun,
Brussels; 2007.
[2] Dufe JA, Beckman WA. Solar engineering of thermal processes. 3rd ed.
Hoboken, New Jersey: Wiley; 2006.
[3] Kalogirou SA. Solar energy engineering e processes and systems. Burlington,
Massachussets: Elsevier/Academic Press; 2009.
[4] CEN, FprEN 12977-2. Thermal solar systems and components e custom built
systems e Part 2: Test methods for solar water heaters and combi-systems;
2011.
[5] QAiST Project. Information on quality assurance in solar thermal: recent
changes in the standards and outlook. QAiST Industry Workshop, Warsaw,
Poland, 14 December 2011.
[6] Belessiotis V, Mathioulakis E, Papanicolaou E. Theoretical formulation and
experimental validation of the inputeoutput modeling approach for large
solar thermal systems. Solar Energy 2010;84(2):245e55.
[7] Belessiotis V, Mathioulakis E, Papanicolaou E. Experimental validation of the
inputeoutput modeling approach for large solar thermal systems e accuracy
of the test procedure. Renew Energy 2012;51:197e205.
[8] Kalogirou SA, Lalot S, Florides G, Desmet B. Development of a neural networkbased fault diagnostic system for solar thermal applications. Solar Energy
2008;82(2):164e72.
[9] Kalogirou SA. Applications of articial neural networks for energy systems.
Appl Energy 2000a;67(1e2):17e35.
[10] Kalogirou SA. Articial neural networks in renewable energy systems: a review. Renew Sustain Energy Rev 2001;5(4):373e401.
[11] Kalogirou SA. Articial intelligence for the modelling and control of combustion processes: a review. Prog Energy Combust Sci 2003;29(6):515e66.
[12] Kalogirou SA, Panteliou S, Dentsoras A. Modelling of solar domestic water
heating systems using articial neural networks. Solar Energy 1999;65(6):
335e42.
[13] Kalogirou SA. Long-term performance prediction of forced circulation solar
domestic water heating systems using articial neural networks. Appl Energy
2000b;66(1):63e74.
[14] Kalogirou SA, Panteliou S. Thermosiphon solar domestic water heating systems long-term performance prediction using articial neural networks. Solar
Energy 2000;69(2):163e74.
[15] Kalogirou SA, Neocleous C, Schizas C. Articial neural networks for modelling
the starting-up of a solar steam generator. Appl Energy 1998;60(2):89e100.
[16] Morabito FC, Versaci M, Pautasso G, Tichmann C. ASDEX upgrade team fuzzyneural approaches to the prediction of disruptions in ASDEX upgrade. Nucl
Fusion 2001;41(11):1715.
[17] ISO. ISO 9459-2, Solar heating e domestic water heating systems e Part 2:
Outdoor test methods for system performance characterization and yearly
performance prediction of solar-only systems. Geneva: ISO; 1995.
[18] CEN European Standard EN12976-2:2006. Thermal solar systems and components e factory made systems e part 2: test methods. Brussels: CEN; 2006.
[19] Haykin S. Neural networks: a comprehensive foundation. New York: Macmillan; 1994.
[20] Rumelhart DE, Hinton GE, Williams RJ. Learning representations by backpropagating errors. Nature 1986;323(6088):533e6.
[21] Werbos PJ. The roots of backpropagation. From ordered derivatives to neural
networks and political forecasting. New York, NY: John Wiley & Sons, Inc.;
1994.