Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Forecasting of photovoltaic power using deep belief network

Neo, Y.Q., Teo, T.T.ORCID: 0000-0002-7552-8497, Woo, W.L., Logenthiran, T. and Sharma, A. (2017) Forecasting of photovoltaic power using deep belief network. In: IEEE Region 10 Conference TENCON 2017, 5 - 8 November 2017, Penang, Malaysia

Link to Published Version:
*Subscription may be required


This main focus of this paper aims to forecast photovoltaic power. The accuracy for forecasting Renewable Energy Sources (RES) are important as it is needed for power grids to operate. It can help make necessary adjustments to operate with RES, which can be highly complexed. As penetration level of renewable generation increases overtime, there may result in a shift towards a generation-dominant grid, causing severe power quality concerns. The proposed methodology of this paper is artificial neural network (ANN) and the training algorithm is Deep Belief Network (DBN). The parameters that are used to configure the software are studied in close observation. The objective of this paper is to determine the parameters of the DBN to accurately forecast photovoltaic power. The proposed methodology is validated by cross-validation and comparing it with another training algorithm.

Item Type: Conference Paper
Item Control Page Item Control Page