Solar Irradiation Prediction using back Propagation and Artificial Neural Network
Journal Title: International Journal of Trend in Scientific Research and Development - Year 2021, Vol 5, Issue 4
Abstract
Solar Energy is one of most promising potential renewable sources of energy. But among all the conventional sources of renewable energy, its nature is quite unpredictable owing to the fact that the solar irradiation keeps on changing and fluctuating. This leads to uncertainty in ascertaining the exact measure of solar power that can be harnessed. Hence forth, a solar irradiation model based on solar irradiation using ANN shall aid in preliminary estimation of the measure of solar power energy that is available for usage incorporating the load dispatch and grid. This proposed work of study puts forward an Artificial Intelligence based model of a Solar Irradiation Forecasting implemented by Artificial Neural Networks ANN . ANN is utilized accredited to its adaptive, data driven and nonlinear nature that exhibits high efficacy in forecasting domains. An endeavor has been also made for solar irradiation forecast using the ANN along with Levenberg Marquardt LM algorithm. The primary reason for using the Levenberg Marquardt algorithm is the fact that is an extremely effective algorithm employing back propagation. The attributes of the algorithm are high stability and speed which yields lesser number of iterations and relatively less magnitude of errors. The evaluation of the proposed system needs to be done on performance specific parameters which in this case are chosen as Mean Absolute Percentage Error MAPE and Regression. Mean absolute percentage error indicates the amount of errors present in the prediction which is to say the deviation between the actual targets and predicted output. Another parameter which validates the mean absolute percentage error is the regression which is a graphical representation of discrete values in the target set and the predicted output. Additional analysis mechanisms such as training states has been also presented which depicts how the mean square error plummets as the number of iterations increase. The variation of mean square error can be seen in training, testing and validation phases. The neural network topology used is 1 20 1 indicating one neuron in the input layer, 20 neurons in the hidden layer and 1 neuron in the output layer respectively. It has been shown that the proposed methodology attains a very good accuracy of approximately 97.74 with the error rate amounting to a meager 2.76 . This model serves to be a robust mechanism and shows good performance. The low error and high accuracy can be attributed to the efficacy of back propagation in Artificial Neural Networks. A comparative analysis is also presented with contemporary work that attains an error of 30 , proving the fact that the proposed system outperforms the contemporary techniques. Harendra Kumar Verma | Ashish Bhargava "Solar Irradiation Prediction using back Propagation and Artificial Neural Network" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-4 , June 2021, URL: https://www.ijtsrd.compapers/ijtsrd43673.pdf Paper URL: https://www.ijtsrd.comengineering/electrical-engineering/43673/solar-irradiation-prediction-using-back-propagation-and-artificial-neural-network/harendra-kumar-verma
Authors and Affiliations
Harendra Kumar Verma | Ashish Bhargava
Keywords
Related Articles
- EP ID EP695915
- DOI -
- Views 85
- Downloads 0
Investment Behavior Among Generational Group
Objective: Share Market investment is a lively platform to invest in the various industries. The objective of the study is to identify investment behavior of the different age groups and to find out the difference in per...
Invisible Web
In this research paper our topic is to analyze and dig deeper into deep web and what can we find when we go dive deep into the internet world. Web means internet as we all know, today almost everyone usesthe internet. We...
A Study of Derivative Market in India
The Indian derivative market has become multi trillion dollar markets over the years. Financial markets are very inventive in augmenting the popularity of derivatives instruments which exemplifies how resourcefully marke...
Grid Connected Solar Photovoltaic Array with MPPT Matlab Simulation
Now a days, due to the problem of generation of electricity and consumption required is not sufficient so, solar generation system is essential. Nowadays, with generates electricity to match the required demand. So this...
ESGP PA Scholars in the University of Eastern Philippines Their Academic Performance, Challenges, and Perceived Solutions
The ESGP PA scholars in the University of Eastern Philippines, in spite of the financial grant, have their fair share of challenges at home, in school, and with the scholarship management. This descriptive qualitative st...