A Comparative Analysis of Solar PV Forecast using SVM under CO2
Concentration
BHABANI PATNAIK1, SARAT CHANDRA SWAIN2, ULLASH KUMAR ROUT3,
RITESH KUMAR DASH4
1,2Department of Electrical Engineering, KIIT University, Bhubaneswar, INDIA
3Department of Electrical Engineering, OUTR, Bhubaneswar, INDIA
4Department of Electrical and Electronics Engineering, REVA University, Bengaluru. INDIA
Abstract: - Installation of a solar PV plant requires an understanding of the performance of solar PV panels in
various weather conditions by which solar output power has to be predicted well in advance. Solar PV
technology is the most reliable and cost-effective technology compared to other renewable energy technologies.
To minimize the cost of production and pollution, it is very essential to improve the techno-economic
parameters of the technology and have a better understanding of the development of solar PV technology but
the efficiency of solar PV technology depends on various environmental factors. Irradiation and temperature
are the main inputs in solar PV technology. Again both the terms depend on greenhouse content and its
concentration in the atmosphere. Due to the influence of many factors, the forecasting of the solar PV
performance in terms of output is a difficult task. So various comparative analysis has been used for forecasting
solar PV power. This paper has analysed support vector machines through the MATLAB simulation model for
forecasting the performance parameters of the solar PV system. An experiment has been conducted at NRRI,
Cuttack in collecting real-time data for analysis.
Keywords: - Irradiation, temperature, ANN, carbon dioxide, OTC, support vector machine (SVM), etc.
Received: April 18, 2021. Revised: March 14, 2022. Accepted: April 16, 2022. Published: May 6, 2022.
1 Introduction
To balance the generation and demand of electricity,
it is important to better plan to produce more energy
in the dispatch schedule, which is called forecasting
[1]. Nowadays, solar PV power and performance
parameters forecasting are very important for future
electricity demand. Error-free forecasting is a
difficult task for researchers because of the
involvement of various parameters such as
irradiance, temperature, cloud, dust particles, black
carbon, aerosol, etc. and presently many more
factors have been included in greenhouse gases [2].
These gases increase gradually due to heavy
pollution all over the world. With the increase in
technological advancement, renewable energy is
having a significant impact. Average 1.73105
terawatts of solar radiation continuously strike the
earth, while global electrics demand averages 2.6
TW. The installed solar PV on-grid capacity of
India was 8.84 GW, 22.45 GW, 34.26 GW and
50.30 GW by the years 2016, 2018, 2020 and 2022
respectively [3]. In an installation of the solar PV
like plants having virtual inertia is always having
importance for VAr generation and consumption.
The reactive power is coming into the picture due to
the converting elements like silicon controlled
rectifier (SCR) and practical diodes. Underbalanced
VAr condition the system is having unity pu of the
voltage profile otherwise voltage sag and Ferranti
effect are realised majorly [4]. However, solar PV
power generation depends on various environmental
factors and due to the large demand for conventional
energy; greenhouse gases emission are increasing
[5]. And nowadays increase in greenhouse gas is a
very common factor included with factors affecting
renewable energy sources, especially for solar
photovoltaic systems [6]. Greenhouse gas is the
combination of different gases in which the
concentration of CO2 is more than other gases.
With the increase of greenhouse gases, the
atmospheric temperature also increases gradually
and affects the performance of the Solar PV system.
Accordingly, the forecasting of the solar system was
calculated by considering the varying temperature
and humidity with a change in GHG concentrations
[7].
There are more gases than the gases included in the
Kyoto Protocol which are counted for the
atmospheric temperature rise due to higher
concentrations of Green House Gases (GHGs) in the
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
84
Volume 17, 2022
atmosphere. Beyond the Kyoto defined gases, other
gases and factors also contribute to environmental
temperature rise. The local moisture context has a
higher effect on the local temperature rise that is
generally realised on summer days in hot climate
regions in the form of sunstrokes or heat strokes [8].
The performance of the solar PV in a practical field
with an open environment is having many variations
with the uncertainty in local temperature, global
temperature, layers of GHGs concentration, level of
dust particles and its deposition on solar panels, etc.
[9]. As many factors are affecting solar PV
performance, its accurate prediction and forecasting
is a big task always. In this work, we considered
short term forecasting (data in short term is taken by
field test or experiment) of solar PV performance by
applying an innovative methodology [10].
There are different types of forecasting methods are
used like statistical method, machine learning
approach, computational intelligence and physical
approach. The statistical approach requires a large
number of data that is required to solve statistical
method analysis which is collected from state
meteorological departments or different research
institutes [11]. These types of approaches are also
known as data-driven approaches but these methods
have some disadvantages in that weather condition
is not tracked by this method, which has a great
impact on the environment as well as the output
power of the solar PV system. Similarly,
computational intelligence includes artificial
intelligence like genetic algorithms, metaheuristic
algorithms, fuzzy logic algorithms and adaptive
neuro-fuzzy logic systems performance or output
[12]. But proper forecasting of solar PV panels
requires a hybrid algorithm. To reduce uncertainty
and errors by proper forecasting methodology plays
an important role in which support vector machine
is a key one [13]. Forecasting is classified in three
different ways such as short term, long term and
medium term. Again each forecasting has some
error limitation like an error of short term
forecasting is 10% to 20%. Similarly, the other two
types of forecasting have also some errors which
change with different parameters [14]. Solar PV
output mainly depends on solar irradiance and
temperature. So it requires a non-linear forecasting
method which requires a lot of parameters that finds
the best solution. This paper analyses, how an
increase in greenhouse gas affects the performance
of the solar photovoltaic cell. We may predict the
solar output power for the future in different
environmental conditions. As per the literature
survey, a support vector machine is used for non-
linear forecasting of irradiance for solar PV, where
the error is in-between 10% to 20% [15].
Support vector machine is divided into two parts
such as regression analysis and classification.
Researchers are preferred to apply regression
analysis for forecasting as it gives better
performance as compared to other machine learning
programs [16]. Irradiance is the main input of solar
energy and it is non-linear in nature. So to predict
irradiation levels accurately for the future by
support vector machines by using historical and
practical measurement data is having hidden
benefits to the future solar PV system to be adopted.
Those data may be input parameters or
environmental parameters of the system and these
parameters are also non-linear in nature. Due to the
non-linearity nature, it is difficult to predict the
efficiency of the system in a certain geographical
area. The following contribution has done in this
paper which is listed below.
This paper presents the forecasting of solar
output power with real data which is driven by
experimental analysis with a higher
concentration of carbon dioxide. These
experimental setups were analysed at National
Rice Research Institute (NRRI) Cuttack [17].
For input, real-time data was collected from
different meteorological departments [18].
The results are compared with different types
of support vector machines as given in the
result section [19].
2 Methodology
As per the above explanation, the error of solar PV
forecasting can be minimized so that the margin will
be maximized as per the hyperplane formulation and
due to this error can be minimized within the
tolerance band. Here two data sets are considered
for the input data set (ϰ1, ϰ2, ϰ3) and the
labelled output is (1, 2, 3) R. Similarly,
assuming a training data set of T = [(ϰ1, 1), (ϰ2,
2), (ϰ3, 3)…(ϰn, n)]. On assuming the data set as
a quadratic loss function as shown in Fig.-1.
Fig. 1: Quadratic loss function.
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
85
Volume 17, 2022
󰇛󰏇󰇛󰇜󰇜
󰇛󰏇󰇛󰇜󰇜
󰇛󰏇󰇛󰇜󰇜 
(1)
The above equation represents the quadratic loss
function and after minimization of Equation (1) will
be,
Min (
) 2 +
  (2)
In Equation (2), ‘C’ represents a regularization
parameter and it is subjected to,
S.T.󰇫
 (3)
Now after mapping Equation (2), the optimization
equation is given below form.
󰇛󰅇󰇜󰇛󰅓󰇜
󰇛󰇜


(4)
Temperature, humidity and solar irradiance are
taken as input with  respectively and
that of the solar panel output as . So in a single
variable
regression model, the output
O
y
can be
related to the input

and S
x
as shown in Eq.
(5).
(
)
,
) (5)
Eq. (5) can be realized in a more basic form as,
󰇱

󰇛

󰇜
󰇛


󰇜

󰇛

󰇜
󰇛󰇜
Eq. (6) shows the one dimensional linear regression.
Here Eq. (1) represents the quadratic loss function.
Therefore, the minimization function for Eq. (1)
becomes N contradiction to the linear regression, as
in non-linear regression the output depends upon
many factors as shown in Eq. (5).
In Eq.-4
, represent the complexity of the
model and F represents the kernel Hilbert space.
Here radial basic function (RBF) is used as the
kernel. In the above equation, we can conclude that
the minimize function consists of two tunable
parameters and according to RBF, sometime
tuneable parameters may increase up to three. For
the support vector machine, the training data is
prepared in five different ways such as data
geometry analysis, neighbourhood analysis,
evolutionary algorithm method, active learning-
based method and random sampling method.
Stratified random sampling technique is used here
because it is a probability sampling technique and it
is more accurate as compared to other sampling
techniques. The stratified random techniques divide
the population into small subgroups so as to form
the strata. Here each strata is defined by the
members present inside the sampling with attributes.
Total population size such as Solar irradiance and
temperature will be divided into smaller sizes of
clusters and their behaviour has been studied in
evaluating the overall prediction of the system. The
main characteristic of this technique is to classify
the data based on a group resembling certain
characteristics. Here data are divided into two sets
of equal length which are called subsets. Based on
population and characteristics the random samples
were collected from the sets. The process of
applying the stratified sampling technique is as
follows.
Identify the population properly after collection
of the experimental data
The sample size will be determined from the
data which was selected previously.
Create subsets accordingly characteristics of
the optimizer variable.
Level data will be an asset and classify the
members of that subset.
Random data will be selected, applying random
theory.
A flowchart is given below for better
understanding.
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
86
Volume 17, 2022
Fig. 2: Flowchart of the used algorithm.
3 Experimental Analysis
National rice research institute had permitted to
experiment in their laboratory field. A special
chamber was built by the institute where the
concentration of carbon dioxide was kept higher
than the atmospheric concentration. Three open-top
chambers (OTC) were built with 575ppm, 550ppm
and 400ppm (atmospheric) concentrations of carbon
dioxide. So using these three OTC the CO2
enrichment experiment has been done. The design
and shape of the OTC are rectangular (6
) and a polycarbonate transparent sheet had
covered by the OTC. And the upper part of the OTC
was opened so that the other factors mainly sunlight,
rain, humidity, etc. remains the same as the normal
atmosphere. The temperature for all the chambers
was recorded with a digital thermometer. Pure CO2
gas was blown (2.5 kg/cm2) inside the OTC through
a perforated polyvinyl tube regulated by solenoid
valves. The concentration of CO2, sampling and data
logging of each OTC was recorded by a system
through an automatic digital meter and a
microcontroller. Here the sensor is connected to a
microcontroller which is based on a closed-loop
control system so that automatically the carbon
dioxide was blown to the chambers (OTC) to
maintain the respective ppm levels of respective
chambers.
As per IPCC, after several years the concentration
of CO2 may increase in upcoming decades. So, one
solar PV set up is kept under (400±10)ppm and the
other two setups are kept under (550±10) and
(575±10) ppm chambers. All three experimental set-
ups with respective laband personal computer (PC)
where the concentration of CO2, temperature and
humidity records are given below in Figures 3&4
respectively.
Fig. 3: Laboratory field and recording personal
computer (PC).
The experimental setup has been arranged in such a
way that the researchers can analyse the
performance of solar PV under three different
carbon dioxide concentrations with same irradiation.
Fig. 4: Experimental setup at 400ppm, 550ppm and
575ppm CO2 concentrations.
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
87
Volume 17, 2022
4 Result Analysis
For this experimental set-up, six types of support
vector regression analysis have been presented. And
calculated different types of error, such as root mean
square error (RMSE), R-Squared (coefficient of
determination), mean squared error (MSE) and
mean absolute error (MAE). In Table-1, the
regression analysis of maximum current of 400ppm
CO2 has shown where the root mean square error of
all the SVM techniques was found between 1.0405
to 1.1835. We can observe that among all SVM
technique courses Gaussian is better than all, with
an accuracy of 92%. In Figure-5, we can observe the
regression and forecasting for Imax at 400ppm carbon
dioxide or GHG concentration, where the maximum
current is very from 4.2 amperes to 4.6 amperes.
Table 1. Different errors for Imax at 400ppm CO2
concentration.
Table 2. Different errors for Vmax at 400ppm CO2
concentration.
Table-2 and Figure-6 represent regression analysis
for maximum voltage 400ppm CO2 and it is found
that the RMSE error is in-between 1.037 to 1.1817.
Again from Table-3 and Figure-7, we can find the
regression analysis of maximum power with
400ppm CO2 concentration.
Table 3. Different errors for Pmax at 400ppm CO2
concentration.
Fig. 5: Support vector regression for maximum
current at 400ppm CO2 concentration.
Types of
SVM
RMSE
R-
Squared
MAE
Linear
1.033
-0.03
0.8079
Quadratic
1.0429
-0.03
0.017
Cubic
1.1348
-0.02
0.7135
Fine
Gaussian
1.0515
-0.05
0.624
Medium
Gaussian
1.0292
-0.102
0.028
Coarse
Gaussian
1.0313
-0.21
0.6281
SVM
(PCA)
1.038
-0.02
0.8112
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
88
Volume 17, 2022
Fig. 6: Support vector regression for maximum
voltage at 400ppm CO2 concentration.
Fig. 7: Support vector regression for maximum
power at 400ppm CO2 concentration.
Table 4. Different errors for Imax at 550ppm CO2
concentration.
Table-4 shows the regression analysis of maximum
current under 550ppm concentration where the root
mean square error varies from 1.0419 to 1.1907.
Table 5. Different errors for Vmax at 550ppm CO2
concentration.
Table-5 shows the regression analysis where the
root mean square error varies from 1.0374 to 1.3155
for maximum voltage under 550ppm carbon dioxide
concentration.
Table 6. Different errors for Pmax at 550ppm CO2
concentration.
Table-6 shows the regression analysis where
different error are their using different type of
support vector machine.where root mean square
error varies from 1.036 to 1.07. Similarly all error
values are calculated.
Types of
SVM
RMSE
R-
Squared
MSE
MAE
Linear
1.0652
-0.07
1.1345
0.9154
Quadratic
1.078
-0.08
1.1624
0.9261
Cubic
1.0361
-0.00
1.0738
0.8972
Fine
Gaussian
1.0375
-0.00
1.0765
0.8973
Medium
Gaussian
1.0362
-0.00
1.0739
0.8973
Coarse
Gaussian
1.0360
-0.00
1.0764
0.8953
SVM
(PCA)
1.0432
-0.01
1.0714
0.9855
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
89
Volume 17, 2022
The outcome of regression analysis on maximum
current, maximum voltage and maximum power for
550 ppm has shown on below figure 8, figure 9 and
figure 10 respectively.
Fig. 8: Support vector regression for maximum
current at 550ppm CO2 concentration.
Fig. 9: Support vector regression for maximum
voltage at 550ppm CO2 concentration.
Fig. 10: Support vector regression for maximum
power at 550ppm CO2 concentration.
Table 7. Different errors for at Imax 575ppm CO2
concentration.
Table 8. Different errors for Vmax with 575ppm at
CO2 concentration.
Table 9. Different errors for Pmax 575ppm at CO2
concentration.
Types of
SVM
RMSE
R-
Squared
MSE
MAE
Linear
0.989
-0.01
0.978
0.855
Quadratic
0.993
-0.03
0.986
0.854
Cubic
1.0135
-0.06
1.0272
0.8675
Fine
Gaussian
1.115
-0.29
1.2436
0.9341
Medium
Gaussian
1.0236
-0.09
1.0477
0.8799
Coarse
Gaussian
0.9877
-0.02
0.9795
0.8531
SVM(PCA)
1.0085
-0.06
0.0176
0.8641
Types of
SVM
RMS
E
R-
Squared
MSE
MAE
Linear
0.992
-0.03
0.9832
0.8578
Quadratic
0.996
-0.04
0.995
0.8562
Cubic
1.002
-0.04
1.002
0.8544
Fine
Gaussian
1.084
-0.22
1.175
0.912
Medium
Gaussian
1.006
1
-0.06
1.0127
0.8563
Coarse
Gaussian
0.99
-0.02
0.98
0.855
SVM(PCA)
0.97
-0.03
0.9801
0.8548
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
90
Volume 17, 2022
Similarly Table-7,Table 8 and table 9 shows the
regression analysis for maximum current
,maximum voltage and maximum power of 575 ppm
respectivily. where different error are also
calculated using different type of support vector
machine.
Fig. 11: Support vector regression for maximum
current at 575ppm CO2 concentration.
Fig. 12: Support vector regression for maximum
voltage at 575ppm CO2 concentration.
Fig. 13: Support vector regression for maximum
power at 575ppm CO2 concentration.
Similarly, the regression analysis of maximum
current ,maximum voltage and maximum power for
575ppm of CO2 with all recorded responses are
carried out which has shown in figure 11,figure12
and figure 13 respectively.
5 Conclusion
Prediction of solar PV output with effect of different
environment parameter plays an important role on
renewable energy production. This paper gives an
experimental analysis of solar PV performance
under higher carbon dioxide concentrations and
forecasting has been done by taking the
experimental data. After discussion on result part, it
is found that increase of carbon dioxide affects the
output power and also it increases the error of
forecasting. With the increase in carbon dioxide, the
ambient temperature of the atmosphere increases
and humidity of the atmosphere decreases which
affects the whole solar PV system. At high
temperature the power production of solar PV
system decreases due to excessive heating of solar
PV panel. From the experimental data, it shows that
even with the increase of 25ppm of carbon dioxide,
the temperature increases up to 0.750C to 1.50C.
Again, it has shown that with increase of 1 to 2
degree Celsius of temperature, the output power of
solar PV decreases up to 7% to 10% compare to the
performance of solar PV under normal atmospheric
pressure and ambient temperature condition where
carbon dioxide concentration is 400ppm. Till now
very rare research has done on this topic. Now-a-
days various technologies are there like water
cooling, air cooling and water spray methods to
maintain standard temperature. But in future to
maximize the power production even at higher
ambient temperature, more advanced technologies
should be developed which can be applied to the
outer material of solar PV panel so that it can trap
less amount of heat to lower the temperature even at
higher irradiation.
References:
[1] B. Patnaik, S. C. Swain, U. K. Rout. An Experimental
Study of Greenhouse Gas Concentration on the
Maximum Power Point of Solar PV Panels.
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
91
Volume 17, 2022
Engineering, Technology and Applied Science
Research, Volume 10, Issue 5, pp. 6200-6203, 2020.
< https://doi.org/10.48084/etasr.3682>.
[2] V. J. Shi, W. J. Lee, Y. Liu, “Forecasting power
output of photovoltaic systems based on weather
classification and support vector machines”, IEEE
Transactions on Industry Applications, Vol. 48, pp.
1064-1069, 2012.
[3] Performance of Solar Power Plants in India,
Submitted to Central Electricity Regulatory
Commission, New Delhi, July 2020; Report on
Optimal Generation Capacity Mix for 2029-30,
Govnt. of India, Ministry of Power, Central
Electricity Authority, Jan. 2020.
[4] D. Almeida, J. Pasupuleti, J. Ekanayake, Comparison
of Reactive Power Control Techniques for Solar PV
Inverters to Mitigate Voltage Rise in Low-Voltage
Grids. Electronics, 10(13), 1569, 2021.
<https://doi.org/10.3390/electronics10131569>.
[5] E. M. Crispim, Pedro M. Ferreira and A. E. Ruano,
“Solar radiation prediction using RBF neural
networks and cloudiness indices”, 2006
International Joint Conference on Neural Networks,
pp. 2611-2618, 2006.
[6] B. Patnaik., S. C. Swain, U. K. Rout, “Effect of
Colour Spectrum and Plastic on the Performance of
PV Solar System”. IJRTE, Vol. 8, pp.10843-10846,
2019.
[7] J. You, W.Ampomah,Q.Sun,R. Balch. A
Comprehensive Techno-eco-assessment of CO2
Enhanced Oil Recovery Projects Using a Machine-
learning Assisted Workflow.2021.
SSRNElectronicJournal.<10.2139/ssrn.3816701>.
[8] M. Tao, J. Gao, W. Zhang, Y. Li, Y. He, Y. Shi. A
Novel Phase-Changing Nonaqueous Solution for
CO2 Capture with High Capacity, Thermostability,
and Regeneration Efficiency. Industrial &
Engineering Chemistry Research, 57,28, 9305–
9312, 2018.
<https://doi.org/10.1021/acs.iecr.8b01775>
[9] R. J. Mustafa, M. R. Gomaa, M. Al-Dhaifallah and H.
Rezk. Environmental impacts on the performance of
solar photovoltaic systems. Sustainability, 12(2), pp.
608, 2020.
[10] M. K. Behera, N. Nayak. A comparative study on
short-term PV power forecasting using
decomposition based optimized extreme learning
machine algorithm. Engineering Science and
Technology, an International Journal, Vol. 23, Issue
1, Pages 156-167, 2020. ISSN 2215-0986.
<https://doi.org/10.1016/j.jestch.2019.03.006>.
[11] R. Paul, R. Dash and S. C. Swain, “Design Analysis
of Current Controller for SPV Grid Connected
System through Hysteresis CCT”, 2018
International Conference on Applied
Electromagnetics, Signal Processing and
Communication(AESPC), pp. 16, 2018.
<DOI:10.1109/AESPC44649. 2018.9033188>.
[12] J. Antonanzas, N. Osorio, R. Escobar, R. Urraca, F.
J. Martinez-de-Pison, F. Antonanzas-Torres. Review
of photovoltaic power forecasting. Solar energy,
136, pp. 78-111, 2016.
[13] F. Harrou, F. Kadri, and Y. Sun, “Forecasting of
Photovoltaic Solar Power Production Using LSTM
Approach”, in Advanced Statistical Modeling,
Forecasting, and Fault Detection in Renewable
Energy Systems. London, United Kingdom: Intech
Open, 2020
<https://www.intechopen.com/chapters/71197 doi:
10.5772/intechopen.91248>.
[14] Jr. S. Fonseca, J. Gari, “Use of support vector
regression and numerically predicted cloudiness to
forecast power output of a photovoltaic power plant
in Kitakyushu, Japan”, Progress in photovoltaics:
Research and applications, Vol. 20, pp. 874-882,
2012.
[15] B. Patnaik, S. C. Swain, U. K. Rout “Modelling and
Performance of Solar PV Panel with Different
Parameters”, Applications of Robotics in Industry
Using Advanced Mechanisms. ARIAM 2019.
Learning and Analytics in Intelligent Systems, Vol.
5. Springer, Cham., 2019.
[16] B. Patnaik, S. C. Swain and U. K. Rout, “Effect of
Greenhouse Gas Concentration on Solar
Photovoltaic Performance”, 2019 International
Conference on Intelligent Computing and Remote
Sensing (ICICRS), pp. 1-4, 2019. <doi:
10.1109/ICICRS46726.2019.9555898>.
[17] A. Kumar, A. K. Nayak, R. P. Sah, P. Sanghamitra,
and B. S. Das, “Effects of elevated CO2
concentration on water productivity and antioxidant
enzyme activities of rice (Oryza sativa L.) Under
water deficit stress”, Field Crops Research, Vol.
212, pp. 61-72, 2017.
[18] B. Patnaik, S. C. Swain, U. K. Rout. Performance of
Solar PV Panel Under Higher Concentration of
Carbon Dioxide. Advances in Electronics,
Communication and Computing. Lecture Notes in
Electrical Engineering, Vol. 709, 2021. Springer,
Singapore.
[19] Bjorn Wolff, Jan Kuhnert, Elke Lorenz, Oliver
Kramer, Detlev Heinemann. Comparing support
vector regression for PV power forecasting to a
physical modelling approach using measurement,
numerical weather prediction, and cloud motion
data. Solar Energy, Volume 135, Pages 197-208,
2016.
<ISSN0038092X,doi.org/10.1016/j.solener.2016.05.
051>.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on POWER SYSTEMS
10.37394/232016.2022.17.10
Bhabani Patnaik, Sarat Chandra Swain,
Ullash Kumar Rout, Ritesh Kumar Dash
E-ISSN: 2224-350X
92
Volume 17, 2022