Short-Term Solar Forecasting Model using Artificial Neural Networks
Naveen Kumar Dubey1, M.P.S. Chawla2
1Naveen Kumar Dubey*, Electrical Engineering Department, SGSITS, Indore, (M.P.), India.
2M.P.S. Chawla, Electrical Engineering Department, SGSITS, Indore, (M.P.), India.
Manuscript received on September 01, 2020. | Revised Manuscript received on September 09, 2020. | Manuscript published on September 15, 2020. | PP: 1-6 | Volume-6, Issue-10, September 2020. | Retrieval Number: 100.1/ijisme.J12590961020 | DOI: 10.35940/ijisme.J1259.0961020
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In present context, Electrical Energy generation in India is mostly based on the conventional sources, but the time has come to be not depend on these conventional sources and to make renewable Energy sources capable of producing total energy demand by its own. So the focus has been shifted towards Wind, Hydro and photovoltaic (PV) power generation. Accurate forecasting of solar irradiance is required for effective and efficient power scheduling & dispatching. And this weather data is needed by the control engineers for planning their control strategies. In this paper a simple approach for weather prediction is proposed which relies on hourly weather data such as Temperature, Relative humidity, surface pressure, wind speed & direction and solar irradiance. The solar forecasting model to predict short-term solar irradiance & other weather parameters, is done by using Leven-berg Marquardt and Bayesian regularization back-propagation algorithms with standard nonlinear autoregressive with external input NARXfeedforward Network. This approachis simple to implement fast in execution and provides good results for short-term time horizon predictions.
Keywords: Solar power, Short-term forecasting, Artificial neural networks.