Machine learning based models for solar energy

It is well-known that solar power depends on many complex parameters such as humidity, radiation, temperature, dust and wind speed. In order to cope with these complex structures and to provide an accurate forecast, the development of reliable and effective forecasting methodology is very significant. In this study, an Artificial Neural Network (ANN)-based system has been developed to forecast very short-term (2 to 4-h) power output of a grid tied Photovoltaic Power Plant (PVPP). An algorithm called Extreme Learning Machine (ELM) has gained increasing interest through its extremely fast learning and good generalization capability. The Kernel Extreme Learning Machine (KELM) which is the improved version of the ELM is proposed to develop a very short-term PVPP power forecast system. The most important feature of KELM is that it has less adjustable parameters and better generalization ability when compared with classical ELM. Experimental studies have been carried out on a grid tied PVPP that has 1 (MW) installed power capacity. The inputs of the KELM-based forecast system are selected as solar power, humidity, radiation and temperature. All data are divided into four parts to analyze the effect of seasons on the performance of the proposed forecast system. The comparison studies are carried out to clearly observe the forecast ability and performance of the KELM. From the experimental studies of the winter season which has not symmetric diurnal mean power curve, Correlation Coefficients (R) values of KELM are calculated as 0.896 for 2-h ahead, 0.865 for 3-h ahead, 0.833 for 4-h ahead. Those of ELM and (Levenberg Marquardt) LM are 0.866–0.812 for 2-h ahead, 0.831–0.669 for 3-h ahead, 0.799–0.502 for 4-h ahead, respectively. Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values of KELM for winter season are 12.08%-6.14% for 2-h ahead, 14.06%-7.43% for 3-h ahead, 15.33%-7.74% for 4-h ahead, respectively. While these values of ELM are obtained as 13.82%-7.05% for 2-h ahead, 15.52%-8.77% for 3-h ahead, 17.63%-9.33% for 4-h ahead. The RMSE and MAE values of LM are 17.90%-9.14% for 2-h ahead, 20.04%-9.85% for 3-h ahead, 21.22%-10.68% for 4-h ahead. According to the obtained results, it is clearly seen that KELM provides a more powerful and reliable forecasting performance.

Short-term reforecasting of power output from a 48 MWe solar PV plant

Photovoltaic (PV) technology, as an efficient solution for mitigating impacts of climate change, has been increasingly used across the world to replace fossil-fuel power to minimize greenhouse gas emissions. With the world's highest cumulative and fastest built PV capacity, China needs to assess the environmental and social impacts of these established photovoltaic (PV) power plants. However, a comprehensive map regarding the locations and extent of the PV power plants remains to be scarce at the country scale. This study developed a workflow combining machine learning and visual interpretation methods with big satellite data to map the PV power plants in China. We applied a pixel-based Random Forest (RF) model to classify the PV power plants from composite images in 2020 with 30-meter spatial resolution on Google Earth Engine (GEE). The result classification map was further improved by a visual interpretation approach. Eventually, we established a map of PV power plants in China by 2020, covering a total area of 2917 km2. Based on the derived national PV map, we found that most PV power plants were sited on cropland, followed by barren land and grassland. In addition, the installation of PV power plants has generally decreased the vegetation cover. This new dataset is expected to be conducive to policy management, environmental assessment, and further classification of PV power plants.

In this paper, the Artificial Neural Networks (ANN) model have been exploited for estimation of the energy production of the 20-MW solar photovoltaic (PV) plant installed in Adrar in terms of energy yield, using a minimal knowledge of meteorological parameter include global solar radiation, relative humidity, wind speed, and ambient or air temperature. A comparative analysis has been carried out between the proposed artificial neural networks model and multiple linear regression (MLR) models in PV output estimation and performance is measured based on evaluation indexes, namely mean absolute percentage error, normalized mean absolute error, and normalized root mean square error respectively. Further, the comparison of the ANN model with MLR based models reveals the advantages and supremacy of the ANN model.

Western China has good conditions for constructing large-scale photovoltaic (PV) power stations; however, such power plants with large fluctuations and strong randomness suffer from the long-distance power transmission problem, which needs to be solved. For large-scale PV power stations that do not have the conditions for simultaneous hydropower and PV power, this study examined long-distance delivery mode and energy storage optimization. The objective was to realize the long-distance transmission of electrical energy and maximize the economic value of the energy storage and PV power storage. For a large-scale PV power station, the energy storage optimization was modelled under a given long-distance delivery mode, and the economic evaluation system quantified using the net present value (NPV) of the battery was based on the energy dispatch optimization model. By contrast, a lithium battery performance model was developed. Therefore, further analysis of the economics of the energy storage and obtaining the best capacity of the energy storage battery and corresponding replacement cycle considered battery degradation. The case study of Qinghai Gonghe 100 MWp demonstration base PV power station showed that the optimal energy storage capacity was 5 MWh, and the optimal replacement period was 2 years. Therefore, the annual abandoned electricity was reduced by 3.051 × 10 4 MWh compared with no energy storage. The utilization rate of both the PV power station and quality of the delivered electricity were modelled to realize a long-distance transmission to the grid net. This will have an important guiding significance to develop and construct large-scale single PV power stations.

This study developed a rational approach to the consideration of solar photovoltaic (PV) power plants applied to electric utility systems, then performed a requirements assessment and preliminary impact and penetration analyses by studying PV generation in three actual utility systems. Conventional utility loss-of-load probability and production simulation methods were used, along with a PV plant performance model developed for the study.Evaluations were based on comparison of total utility generation system costs with and without PV plants. The study is reported under the following sections: 1) geographic areas for photovoltaic power plants; 2) siting of photovoltaic power plants; 3) electric utility characteristics; 4) photovoltaic power plant characteristics; 5) methods of study; 6) study inputs; 7) results; 8) cost and performance goals for viability; 9) reference PV selection; 10) reference PV plant definitions; 11) comparison with conventional generation; 12) requirements for PV applications; 13) preliminary impact analysis; and 14) preliminary penetration analysis. (WHK)

Performance study on a grid connected 20 kWp solar photovoltaic installation in an industry in Tiruchirappalli (India)

In this work, we evaluate the performance of photovoltaic (PV) power generation forecast using various hybrid deep-learning algorithms, including long and short term memory (LSTM), LSTM with an Autoencoder ( LSTM-Autoencoder ) and LSTM with an attention mechanism (LSTM-Attention). We show that the LSTM-Attention model is significantly more accurate in predicting the hourly power generation of a PV plant with 162kW capacity than the other two reference counterparts. After 100 epochs training, the model achieves a superior Root Mean Square Error (RMSE) below 0.01, Mean Absolute Error (MAE) below 0.005, the Absolute Deviation (AD) below 0.02, and the Mean Absolute Percantage Error (MAPE) is around 35%. Moreover, since the correlation coefficient is up to 92%, this hybrid model not only can be used in solar power generation prediction in PV plants, but potentially can also be extended to other renewable energy sources such as predicting wind power or tide power generation.

Nowadays , nonlinear time series and artificial neural networks (ANN) models are used for forecasting in the field of business, agriculture and soon. Recent studies have shown, ANN have been successfully used for forecasting of financial and agriculture data series The classical methods used for time series prediction like Box-Jenkins or ARIMA assumes that there is a linear relationship between inputs and outputs. ANN have more advantages that can approximate to model both linear and nonlinear structures in time series, they are not able to handling both structures equally well. The autoregressive integrated moving average (ARIMA) model and two ANN models namely, Radial basis function neural networks (RBFNN), and Elman recurrent neural networks (ERNN) methods were applied to Hyderabad airport traffic data. The data obtained for 15 years from 2002–2003 to 2016–2017 about domestic and international passenger of International Airport of Hyderabad, India. In this research paper, we compared the performances of ARIMA, RBFNN and ERNN were based on three measures: mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE). The results showed that RBFNN obtained the smallest MAE, MAPE and RMSE in both the modeling and forecasting processes. The performances of the three models ranked in ascending order were: ARIMA, ERNN and the RBFNN model. Keywords: T ime series, forecasting, artificial neural networks, ARIMA models, radial basis function neural networks, and Elman recurrent neural networks Cite this Article R. Ramakrishna, Berhe Aregay, Tewodros Gebregergs. The Comparison in Time Series Forecasting of Air Traffic Data by Autoregressive Integrated Moving Average Model, Radial Basis Function and Elman Recurrent Neural Networks. Research & Reviews: Journal of Statistics . 2018; 7(3): 75–90p.

Considering the recent drop (up to 86%) in photovoltaic (PV) module prices from 2010 to 2017, many countries have shown interest in investing in PV plants to meet their energy demand. In this study, a detailed design methodology is presented to achieve high benefits with low installation, maintenance and operation costs of PV plants. This procedure includes in detail the semi-hourly average time meteorological data from the location to maximise the accuracy and detailed characteristics of different PV modules and inverters. The minimum levelised cost of energy (LCOE) and maximum annual energy are the objective functions in this proposed procedure, whereas the design variables are the number of series and parallel PV modules, the number of PV module lines per row, tilt angle and orientation, inter-row space, PV module type, and inverter structure. The design problem was solved using a recent hybrid algorithm, namely, the grey wolf optimiser-sine cosine algorithm. The high performance for LCOE-based design optimisation in economic terms with lower installation, maintenance and operation costs than that resulting from the use of maximum annual energy objective function by 12%. Moreover, sensitivity analysis showed that the PV plant performance can be improved by decreasing the PV module annual reduction coefficient.

One of the primary sources of renewable energy in the coming years is thought to be solar energy. Solar energy and other renewable energy sources do, moreover, have a disadvantage in that it is hard to forecast when they will be available. The best use of solar energy is impacted by this issue, particularly when it is combined with other sources. As a result, the organization and economy of solar energy depend on accurate solar energy forecasting techniques. Predicting solar energy shortly is the study’s major goal. This paper describes the study of Neutrosophic fuzzy logic with artificial neural networks (NFL-ANN) to anticipate solar photovoltaic (PV) plant output power with the use of specified input factors known as meteorological information, such as sunshine length, humidity levels, temperature, air pressure, and others, artificial neural networks are used to forecast the outcome. NFL represents a generalised logic, which can manage stochasticity learning mistakes and unpredictability that fuzzy logic lacks. It offers the results of the calculation section. Excellent performance computer processors and NFL provide reasonable accuracy estimates of solar plant outputs as well as system reliability to consider environmental factors. The investigation was carried out with the use of MATLAB programming. With the assistance of statistical markers like mean absolute percentage error (MAPE), mean absolute error (MAE), root means square error (RMSE), and determinant coefficient, the suggested NFL-ANN approach is evaluated and compared to other approaches that are already in use. In comparison to existing techniques, the suggested NFL-ANN provides superior accuracy and lesser prediction error, according to the study’s findings. This research will be enhanced to forecast power without any loss.

With the declining costs of photovoltaics (PV), and the excellent solar energy resource availability in the country, the Brazilian government and the electricity sector have started to evaluate and consider PV as a serious potential contributor to the National electricity mix. Since the late 1990s, Brazilian electrical utilities are required by the National Electrical Energy Regulatory Agency ANEEL to invest 1% of their operational income on R&D. In 2011 ANEEL issued an R&D call dedicated to utility-scale PV. The solar energy research group at Universidade Federal de Santa Catarina (www.fotovoltaica.ufsc.br) has been actively investigating and promoting PV in Brazil, operates since 1997 the first grid-connected, thin-film PV generator in the country. Under the ANEEL R&D call, a 4-year, US$ 20 million project was started in 2012. The project aims at assessing the performance of seven different PV technologies at eight different Evaluation Sites (ES) in Brazil, and also to design, procure, install and monitor the performance of a utility-scale 3 MWp R&D PV power plant, which is located at one of these eight ES. The 3 MWp PV power plant and all the eight ES are fully monitored, with all electrical and environmental parameters measured at 1-s intervals. PV technologies include thin-film amorphous silicon (a-Si), microcrystalline silicon (μ c-Si), cadmium telluride (CdTe), copper indium gallium diselenide, mono- and multi-crystalline silicon (c-Si and m-Si), all at fixed tilt, as well as double-axis tracking, concentrated PV using triple-junction InGap/GaAs/Ge at 820 suns concentration. All ES are identical, except for the fixed PV arrays tilt angle, which is equal to the latitude at each site. The 3 MWp R&D PV power plant is co-located at one of the ES sites. Thin-film PV technologies with a low temperature coefficient of power presented superior output performance, and cloud-edge and cloud-enhancement effects of solar irradiance resulted in operational issues that were not previously described in the literature. Inverter Loading Ratios commonly described in the literature (for less sunny sites) led to considerable annual energy losses.

Photovoltaic (PV) power generation is highly stochastic and volatile, a trait that presents a notable challenge to the prediction accuracy of distributed PV systems. To address this challenge, this study proposes a short-term photovoltaic power prediction strategy that integrates variational modal decomposition (VMD) for feature extraction with an improved RIME (IRIME) optimization algorithm for parameter optimization. Firstly, the raw PV power data are split into several intrinsic mode functions (IMFs) using VMD. The decomposed IMFs are reconstructed by using the sample entropy (SE) method, and a new subsequence with enhanced features is obtained. Secondly, a bidirectional gated recurrent unit (BIGRU) prediction model is established, and its structural parameters are optimized by the IRIME algorithm. Finally, the prediction results of each subsequence are summarized to obtain the final prediction value. Information from a centralized PV power station located in southern China is employed to verify the suggested prediction model. Experimental findings indicate that in comparison with other models, the proposed model achieves the smallest PV power prediction error; the mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) of the proposed model are reduced at least by 21.95%, 3.03%, and 12.33%, respectively. The coefficient of determination (R2) is increased at least by 10.56‰. The method presented in this research is capable of improving prediction accuracy efficiently and holds specific engineering practicality.

Short-term PV power prediction based on meteorological similarity days and SSA-BiLSTM

Regions with limited space for constructing renewable power generation systems need to maximize electricity generation by optimizing the operational efficiency of existing plants and selecting an optimal location for the new construction of PV power plants with favorable weather conditions and surrounding environment. Utilizing monthly input and output data, including four inputs (solar irradiation, temperature, number of modules, and photovoltaic (PV) array rated capacity) and one output (electricity generation) from utility-scale PV power plants, meta-frontier data envelopment analysis was employed in this study to identify factors contributing to power generation inefficiency. Additionally, a Monte Carlo experiment analyzed the impact of solar irradiation uncertainty on power generation efficiency. The findings revealed that the average power generation inefficiency during the study period was 0.445, primarily attributable to seasonal and technical factors. Since these factors are difficult to control once a power plant is in operation, it is important to select an optimal site for power plants by considering meteorological and geographical data. Moreover, the results indicated a significant gap between observed and simulated values of power generation efficiency, arising from variations in weather conditions, power plant site area, and the data used in the research. Employing PV modules with higher electricity output levels can boost the DC/AC ratio, thereby increasing power generation, enhancing efficiency, and contributing to a stable power supply, thus reducing daily and seasonal fluctuations in power generation.