Figure - available from: Environment Development and Sustainability
This content is subject to copyright. Terms and conditions apply.
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 st...
... However, in terms of performance, this technology still does not provide that much of great deal, and that is the focus of science and engineering nowadays, as it is known that the efficiency of this conversion process depends on several climatic and ambient conditions that affects the power produced by a PV panel, predicting that power output time ahead is a key factor for the reliability and stability of such a system as it is also important from an economical point of view. However, complexities and nonlinearity arise for that prediction, due to the fact of the relation between the power production of PVs and unpredictable and mostly correlated ambient conditions such as the wind speed, thus and for an accurate prediction, researchers employed artificial neural networks for this task , mimicking the human brain and taking advantage of the ANN's capability to learn and predict nonlinear problems. ...
... The data collected on the 06/08/2022 are presented in Fig. 3 below where it presents the ambient temperature, solar irradiation, wind speed and module temperature respectively, values of environmental conditions have similar trend, where ambient temperature and solar irradiation tends to increment from the morning to noon time, and then after takes decrease towards the evening where the module temperature follows similar trend due to being a function of the solar irradiation and ambient temperature, while we notice a kind of randomness in the values of the wind speed. (2). ...
Energy is currently a critical and important sector, that's why science and engineering are interested in it, specifically renewable energies. Photovoltaic modules are one of the most promising technologies for directly generating electricity from a renewable source. Nevertheless, several environmental and unpredictable parameters may affect their performance, which introduces difficulties forecasting their output. In this paper, the attention is to forecast the power output of a PV module using ANNs, which have the ability to tackle nonlinear problems. The study accounts for different environmental factors, namely, ambient temperature, solar irradiation, wind speed, and module temperature, while the output is power generation of the PV module. The used data was collected on 3 different days. First the ANN model is trained, validated, and tested with the first dataset; then a new dataset is introduced to the model to predict the output; and finally, based on the model prediction, a correlation for the power output is produced and evaluated with different datasets. Furthermore, the performance of different training algorithm is evaluated in this work, the results shows that a high level of accuracy is achieved using ANNs in the prediction process, where the Scaled Conjugate Gradient was the best in terms of performance with a correlation coefficient of 0.99 and low error values, 3.46 (RMSE) and 2.75 (MAE). A great level of precision is obtained with developed correlation as well (0.95 (R^2) and 5.20 (RMSE)).
The energy industry is always looking at smart ways of conserving and managing the growing demand of consumable and renewable energy. One of the main challenges with modern electric grids systems is to develop specialized models to predict various factors, such as optimal power flow, load generation, fault detection, condition-based maintenance on assets, performance characteristics of assets, and anomalies. When it comes to energy prediction, the data that is associated with it is diverse. Accurate analysis and predictions of such data cannot be done manually and hence needs a powerful computing tool that is capable of reading such diverse data, converting them to a form that can be used for analysis and knowledge discovery. In this research effort, data collected from a local utility company over a period of three years from 2019 to 2021 were used to perform the analysis. Machine learning and neural network algorithms were used to build the predictive model. Different levels of data abstraction along with optimization of model parameters were done for better results. A comparative study was done at the end on the various algorithms used to determine the model performance.