ArticlePublisher preview available

Prediction of grid-connected photovoltaic performance using artificial neural networks and experimental dataset considering environmental variation

Authors:
To read the full-text of this research, you can request a copy directly from the author.

Abstract and Figures

Photovoltaic (PV) investment requires a feasibility study of the PV system in terms of environmental parameters at the location, which is the implementation time and cost. In this study, a 1.4 PV system was installed in Sohar, Oman and the system recorded data, which was modelled using an artificial neural network (ANN). The contribution of this study is to use three proposed ANN models (MLP, SOFM, and SVM) to predict similar systems in twelve other locations throughout the country based on measured solar irradiance and ambient temperature in these locations. The experimental results of Sohar show feasible values of 6.82 A, 150–160 V, 800–1000 W, and 245.8 kWh, peak current, voltage, power, and energy, respectively. Also, the proposed models show an excellent prediction with less error and high accuracy. Furthermore, statistical and sensitivity analyses are presented with a comparison of results found by researchers in the literature for validation. The lowest RMSE was found for SOFM (0.2514) in the training phase compared with (0.2528) for MLP and (0.2167) for SVM. The same sequence but with a higher accuracy was found for SOFM (95.25%), while (92.55%) and (89.19%) for MLP and SVM, respectively. In conclusion, the sensitivity analysis shows that solar irradiance has more effect on the output compared with ambient temperature. Also, a prediction of PV output for Duqm was forecasted till 2050, where it is found insignificant deviation due to climate change compared with 2020.
This content is subject to copyright. Terms and conditions apply.
Vol.:(0123456789)
Environment, Development and Sustainability
https://doi.org/10.1007/s10668-022-02174-0
1 3
Prediction ofgrid‑connected photovoltaic performance
using artificial neural networks andexperimental dataset
considering environmental variation
HusseinA.Kazem1,2
Received: 28 June 2021 / Accepted: 31 January 2022
© The Author(s), under exclusive licence to Springer Nature B.V. 2022
Abstract
Photovoltaic (PV) investment requires a feasibility study of the PV system in terms of envi-
ronmental parameters at the location, which is the implementation time and cost. In this
study, a 1.4 PV system was installed in Sohar, Oman and the system recorded data, which
was modelled using an artificial neural network (ANN). The contribution of this study is
to use three proposed ANN models (MLP, SOFM, and SVM) to predict similar systems
in twelve other locations throughout the country based on measured solar irradiance and
ambient temperature in these locations. The experimental results of Sohar show feasible
values of 6.82 A, 150–160V, 800–1000W, and 245.8 kWh, peak current, voltage, power,
and energy, respectively. Also, the proposed models show an excellent prediction with less
error and high accuracy. Furthermore, statistical and sensitivity analyses are presented with
a comparison of results found by researchers in the literature for validation. The lowest
RMSE was found for SOFM (0.2514) in the training phase compared with (0.2528) for
MLP and (0.2167) for SVM. The same sequence but with a higher accuracy was found for
SOFM (95.25%), while (92.55%) and (89.19%) for MLP and SVM, respectively. In conclu-
sion, the sensitivity analysis shows that solar irradiance has more effect on the output com-
pared with ambient temperature. Also, a prediction of PV output for Duqm was forecasted
till 2050, where it is found insignificant deviation due to climate change compared with
2020.
Keywords Grid-connected PV· Artificial neural network· MLP· SOFM· SVM· PV
performance
List of symbols
PV Photovoltaic
GCPV Grid-connected photovoltaic
ANN Artificial neural networks
VOC Open circuit voltage
* Hussein A. Kazem
h.kazem@su.edu.om
1 Sohar University, PO Box44, SoharPCI311, Oman
2 Solar Energy Research Institute, Universiti Kebangsaan Malaysia, 43600Bangi, Selangor,
Malaysia
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Due to the expected lack of fossil fuels in near future as well as climate change produced by greenhouse effect as consequence of environmental emissions, renewable energy generation, and specifically solar photovoltaic generation, has become relevant in present energy generation challenge. Photovoltaic generators have strong relationship with solar irradiation and outdoor temperature in energy generation process. These meteorological parameters are volatile and uncertain in nature so, unexpected changes on these parameters produce variations on solar photovoltaic generators’ output power. While many researchers have been focused in recent years on the development of novel models for forecasting involved meteorological parameters in photovoltaic generation, they commonly do not consider an analysis step of the data before using it in the developed models. Hence, the aim of this study consists in assembling a wavelet based time-frequency analysis of the used data with deep learning neural networks to forecast solar irradiation, in next 10 min, to compute solar photovoltaic generation. Results of the validation step showed that the deviation of the proposed forecaster was lower than 4% in 90.60% of studied sample days. Final forecaster’s root mean square error was 35.77 W/m², which was an accuracy improvement of 37.52% compared against persistence benchmark model.
Article
Full-text available
The Middle East and North Africa can exploit solar energy resources and export them to Europe and South Asia for a sustainable future of the world. A high voltage direct current (HVDC) multi-terminal transmission grid is employed in this research to export solar energy to South Asia from the Middle East and from North Africa to Europe. The 4 GW HVDC multi-terminal transmission line is proposed in the Middle East and North Africa countries of Oman, UAE, Saudi Arabia, Kuwait, Iraq, Iran, Turkey, Egypt, Libya, Tunisia, Algeria, and Morocco. Photovoltaic energy from the furthest point in Riyadh, Saudi Arabia is proposed to be exported to east and west to Gwadar, Pakistan, and Alexandroupoli, Greece, respectively. In another route, photovoltaic energy export is analyzed from Riyadh to Gwadar via UAE and Oman, while energy export from Tabuk, Saudi Arabia to Sevilla, Spain via North Africa is analyzed. The energy export from all locations is found to be economically feasible. The most viable route is energy export from Riyadh, Saudi Arabia to Alexandroupoli, Greece via Kuwait, Iraq, Iran, and Turkey which achieves an internal rate of return of 19.37% with the net present value of 2561 MUSD for a power purchase agreement of 0.11 USD/kWh. The large-scale PV plants and transmission line strengthen the local grids and provide a photovoltaic energy shifting of 1.5 h between neighboring countries of the Middle East and North Africa and a 7100 km long solar silk road offers reliable photovoltaic energy to customer countries. The project will enhance the Middle East and North Africa countries' technical capability and reduce air pollution.
Article
Full-text available
Renewable energy is essential for planet sustainability. Renewable energy output forecasting has a significant impact on making decisions related to operating and managing power systems. Accurate prediction of renewable energy output is vital to ensure grid reliability and permanency and reduce the risk and cost of the energy market and systems. Deep learning's recent success in many applications has attracted researchers to this field and its promising potential is manifested in the richness of the proposed methods and the increasing number of publications. To facilitate further research and development in this area, this paper provides a review of deep learning-based solar and wind energy forecasting research published during the last five years discussing extensively the data and datasets used in the reviewed works, the data pre-processing methods, deterministic and probabilistic methods, and evaluation and comparison methods. The core characteristics of all the reviewed works are summarised in tabular forms to enable methodological comparisons. The current challenges in the field and future research directions are given. The trends show that hybrid forecasting models are the most used in this field followed by Recurrent Neural Network models including Long Short-Term Memory and Gated Recurrent Unit, and in the third place Convolutional Neural Networks. We also find that probabilistic and multistep ahead forecasting methods are gaining more attention. Moreover, we devise a broad taxonomy of the research using the key insights gained from this extensive review, the taxonomy we believe will be vital in understanding the cutting-edge and accelerating innovation in this field.
Article
Full-text available
Given the ever-increasing people’s need for housing, affordable housing and environmental comfort are important, and the use of inactive design methods is the cheapest solution to achieve low-energy buildings. This study aimed at understanding the natural ventilation potential in Gorgan. This research has been conducted aiming at providing solutions based on airflow to reduce energy consumption in the building for the warm months of the year. Solutions for using natural ventilation in the design of a residential complex include a terrace, void, direct ventilation for the corridors, cross ventilation, and arranging the units to get suitable ventilation. The energy influence the air through the intermediate spaces of the floor units; as the external consumption during the warm period of the year was simulated in the form of an applied project. In this research, a model with mechanical ventilation and 4 patterns of 22 modes were designed using natural ventilation. The results showed that the factors such as using void, stair layout of the plan, using semi-empty spaces in the building, creating internal lateral ventilation, being the interior doors opened, placing the smaller unit from west to south, the connection between the kitchen and the living space in the building causes a reduction of the energy consumption in the warm seasons of the year in moderate and humid climates.
Article
Full-text available
The current work compares the feasibility of using a non-renewable system, a standalone renewable system, and a hybrid renewable system to power up a remote community in Malaysian Borneo. The analysis shows that both the photovoltaic (PV) and hydrokinetic systems can be used to generate renewable electricity, either as a standalone or a hybrid system. Specifically, using a non-renewable system (50 kW standalone diesel system) based on a conventionally high diesel price ($ 1.482/L) resulted in a high net present cost (NPC) of $ 1.788 million. In contrast, a standalone renewable system, i.e., a 280 kW standalone PV system, resulted in a lower NPC ($ 1.615 million) than the standalone diesel system, and it can cater for highly fluctuated solar output with 100% renewables generation. Alternatively, the use of a hybrid PV/hydrokinetic system with a 100 kW PV, 60 kW hydrokinetic turbines, and an 85 kW converter was both economically and technically feasible (with $ 1.714 million NPC), with low water speed. Meanwhile, the use of a hybrid PV/diesel system with 60 kW PV, a 50 kW converter, 120 batteries, and a backup diesel generator could effectively reduce the dependence of electricity generation on diesel supply and resulted in the lowest NPC ($ 1.129 million) among all the analyzed system types. Significantly, the current work demonstrates that the use of PV and hydrokinetic renewable sources of energy in Malaysian Borneo is feasible and deserves more attention to electrify remote communities in Malaysian Borneo.
Article
Full-text available
In many Brazilian dairy farms, the supply of electrical energy does not fully meet the demands of the production systems and/or presents frequent fluctuations and interruptions, making the expansion of production chain unfeasible. These and other barriers have motivating the adoption of renewable and independent energy sources, installed directly in the farms. The present study was conducted to perform a technical analysis of photovoltaic energy generation required for supplying the electricity demand in dairy farms located at Minas Gerais State, Brazil. Daily global radiations on tilted panels were estimated by mathematical models and long-term meteorological database (greater than 33 years). Electrical energy generated by grid-connected photovoltaic systems was predicted considering minimum, average, and maximum daily global radiation conditions. Electrical energy demands were monitored in small (66 cows), medium (106 cows), and large (158 cows) dairy farms. The direct use of electrical energy in operations performed at Brazilian dairy farms were 4, 28, 15, 39, and 15%, on average, for lighting, milking, water heating/pumping, milk cooling/refrigeration, and miscellaneous, respectively. The estimated energy consumptions were 0.96, 0.87, and 0.81 kWh cow−1 day−1 for farms with 66, 106, and 158 cows, respectively. The photovoltaic panel area varied considerably among small, medium, and large dairy farms in order to meet 80% of renewable fraction for the electricity demand.
Article
This study examines the multiscale spillovers and nonlinear causalities between the crude oil futures market and the stock markets of the United States (US), Canada, China, Russia, and Venezuela before and during the COVID-19 pandemic. Using the wavelet coherency method, we find strong co-movement between the oil futures market and these five stock markets, particularly from March 2020 to May 2020 (initial period of the COVID-19 outbreak) at high frequency. Furthermore, we find positive co-movements at low frequency during the overall COVID-19 period. This finding suggests that the bearish trend of stock markets is associated with a downward movement in oil prices. Using the wavelet-based Granger causality approach, we find that the oil and stock indices have less co-movement on a smaller scale but greater movement on a larger scale across all periods. As an exception, the Russian market is significantly influenced by oil prices, even on a small scale, before the COVID-19 period, but not after the beginning of the pandemic. We also find effects in the opposite direction—the Canadian and U.S. markets influence oil prices on a small scale during the COVID-19 period, an effect that is not visible for the U.S. market in the pre-COVID-19 sample. The results also show a significant bidirectional causality from oil to stock markets and vice versa during Russian-Saudi oil price war at high scale. Furthermore, we find that investors should hold more oil futures than stock shares in their portfolios for all periods. This evidence confirms that oil instruments are important for hedging during normal periods and act as safe-haven assets during crisis periods. We observe that the U.S. and Canadian stock markets were more affected by oil price shocks than were other countries.
Article
Nowadays the demand for energy increased rapidly duo to escalating population and global technological development. Consequently, the focus on non-conventional energy sources is increasing rapidly due to change in climates and unhealthy environment caused by fossil fuels. Solar energy is vital to fulfill this growing need of clean energy. The performance of the photovoltaic (PV) technology used for capturing sun’s radiation is directly affected by solar irradiation. The optimum performance of the PV systems would be possible if the solar panel is always orientated towards the direction of maximum radiations of sun. Hence, in order improve the performance of the PV, tracking of maximum radiation of the sun is extremely important. In this experimental study, solar tracking is done using a single axis solar sun tracker which not only provides accurate but also cost efficient solar sun tracking in comparison with an existing stationary system of same capacity. The results show that a single axis solar tracking systems have generated approximately 28.3% more compared to static systems and is also found to be more economic than a solar tracking system which uses microcontroller.
Article
In this study, the aging measurements of a 1.4 kW grid-connected photovoltaic system were analyzed. The system is located at the Solar Energy Laboratory at the College of Engineering, Sohar University, Sohar, Oman. The system variables were monitored and measured for a period of seven years, starting from 1 October 2012 until 30 September 2019, during which the electricity produced to the network was fed. Weather data metrics measurements showed solar irradiation values, ambient temperature, and the PV module temperature were measured for seven years. The performance criteria for the different PV system, which included (the PV module and system efficiency), performance ratio, and capacity factor were measured in addition to the different losses and system productivity. The results showed that the system is exposed to aging, although the amount of its impact is relatively small, as the study was conducted for a seven-year-old, while the effect of aging is more at longer ages. The measured aging caused the system efficiency to decrease by 6.3% and the production rate to 5.88% while the mean daily array capture loss and system loss were 6.95% and 6.13%, respectively, and the capacity factor decreased by 4.91%. Measurements during the seven years showed that the rate of degradation is greater during the summer than the rest of the seasons due to the high radiation intensity which causes high temperatures of the PV units in addition to the high dust density during this season. The results showed that the use of GCPV systems in the Sohar region is very successful in terms of the limited aging of the system and its long-term viability with appropriate efficiency.
Article
A grid-connected photovoltaic system was tested and investigated for the entire year under desertic weather exhibited. The system contains 1.4 kW PV and 1.7 kW inverter —the data was measured every second and used to model and evaluate the system performance. However, dust is one of the essential parameters that affect grid-connected photovoltaic performance, yield, and profitability. In this paper, a proposed model has been takes into consideration the dust impact on grid-connected photovoltaic performance in a novel way. The outcomes demonstrate that the proposed model precisely anticipated the system performance and validated through experimental results. However, it is found that the photovoltaic, inverter, and performance efficiencies are 10.80%, 94.00%, and 73.00%, respectively. Besides, the average yield factor is 141.39 kWh/kWp, and capacity factor is 19.64%. It concluded that the system productivity in Oman is within the prospected rate. The information in this paper is valuable for investors, researchers, and engineers interested in grid-connected photovoltaic and dust impact in Oman and neighboring countries.