0
Journal of Solar Energy Engineering | Volume 137 | Issue 3 | research-article
Research Papers

Twenty-Four Hour Solar Irradiance Forecast Based on Neural Networks and Numerical Weather Prediction

[+] Author and Article Information
C. Cornaro

Department of Enterprise Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, ItalyCHOSE,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: cornaro@uniroma2.it

F. Bucci

Department of Enterprise Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: frabucci@gmail.com

M. Pierro

Department of Enterprise Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: marco.pierro@gmail.com

F. Del Frate

Department of Civil Engineering and
Computer Science Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: fabio.delfrate@disp.uniroma2.it

S. Peronaci

Department of Civil Engineering and
Computer Science Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: simone.peronaci@hotmail.it

A. Taravat

Department of Civil Engineering and
Computer Science Engineering,
University of Rome “Tor Vergata”,
Via del Politecnico 1, Rome 00133, Italy
e-mail: art23130@gmail.com

1Corresponding author.

Contributed by the Solar Energy Division of ASME for publication in the JOURNAL OF SOLAR ENERGY ENGINEERING: INCLUDING WIND ENERGY AND BUILDING ENERGY CONSERVATION. Manuscript received March 28, 2014; final manuscript received December 18, 2014; published online January 8, 2015. Assoc. Editor: Philippe Blanc.

J. Sol. Energy Eng 137(3), 031011 (Jun 01, 2015) (9 pages) Paper No: SOL-14-1104; doi: 10.1115/1.4029452 History: Received March 28, 2014; Revised December 18, 2014; Online January 08, 2015
Article
References
Figures
Tables
Citing Articles

In this paper, several models to forecast the hourly solar irradiance with a day in advance using artificial neural network techniques have been developed and analyzed. The forecast irradiance is the one incident on the plane of the modules array of a photovoltaic plant. Pure statistical (ST) models that use only local measured data and model output statistics (MOS) approaches to refine numerical weather prediction data are tested for the University of Rome “Tor Vergata” site. The performance of ST and MOS, together with the persistence model (PM), is compared. The ST models improve the performance of the PM of around 20%. The combination of ST and NWP in the MOS approach gives the best performance, improving the forecast of approximately 39% with respect to the PM.
FIGURES IN THIS ARTICLE
<>
Copyright © 2015 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

HV of irradiance for two different days. Variable day: January 1, 2009 and not variable day: January 11, 2009.

+HV of irradiance for two different days. Variable day: January 1, 2009 and not variable day: January 11, 2009.
View Large  |  Save Figure  |  Download Slide (.ppt)
HV of irradiance for two different days. Variable day: January 1, 2009 and not variable day: January 11, 2009.
Grahic Jump Location
Fig. 2

PDF of normalized daily variation

+PDF of normalized daily variation
View Large  |  Save Figure  |  Download Slide (.ppt)
PDF of normalized daily variation
Grahic Jump Location
Fig. 3

Sketch of the MLPNN architecture. P1: input vector with R rows. For i-layer, IWi: input weights, LWi: layer weights, bi: bias vector, Si: number of neurons, ai: output vector, and fi: transfer function.

+Sketch of the MLPNN architecture. P1: input vector with R rows. For i-layer, IWi: input weights, LWi: layer weights, bi: bias vector, Si: number of neurons, ai: output vector, and fi: transfer function.
View Large  |  Save Figure  |  Download Slide (.ppt)
Sketch of the MLPNN architecture. P1: input vector with R rows. For i-layer, IWi: input weights, LWi: layer weights, bi: bias vector, Si: number of neurons, ai: output vector, and fi: transfer function.
Grahic Jump Location
Fig. 4

(a) correlations between measured data and MPL4.1 and MPL4.2 model forecast data (hourly shape forecasting), (b) correlations between measured data and 1MLP and 5NWPMLP

+(a) correlations between measured data and MPL4.1 and MPL4.2 model forecast data (hourly shape forecasting), (b) correlations between measured data and 1MLP and 5NWPMLP
View Large  |  Save Figure  |  Download Slide (.ppt)
(a) correlations between measured data and MPL4.1 and MPL4.2 model forecast data (hourly shape forecasting), (b) correlations between measured data and 1MLP and 5NWPMLP
Grahic Jump Location
Fig. 5

Improvement 1MLP model (a) and 5NWPMLP model (b) with respect to the PM

+Improvement 1MLP model (a) and 5NWPMLP model (b) with respect to the PM
View Large  |  Save Figure  |  Download Slide (.ppt)
Improvement 1MLP model (a) and 5NWPMLP model (b) with respect to the PM
Grahic Jump Location
Fig. 6

Example of sequence of measured and forecast data for the 1MLP and the 5NWPMLP models

+Example of sequence of measured and forecast data for the 1MLP and the 5NWPMLP models
View Large  |  Save Figure  |  Download Slide (.ppt)
Example of sequence of measured and forecast data for the 1MLP and the 5NWPMLP models
Grahic Jump Location
Fig. 7

On the left: MBE values versus forecast Kcs for the ECMWF; on the right: the same for the 5NWPMLP

+On the left: MBE values versus forecast Kcs for the ECMWF; on the right: the same for the 5NWPMLP
View Large  |  Save Figure  |  Download Slide (.ppt)
On the left: MBE values versus forecast Kcs for the ECMWF; on the right: the same for the 5NWPMLP
Grahic Jump Location
Fig. 8

5NWPMLP model KS test

+5NWPMLP model KS test
View Large  |  Save Figure  |  Download Slide (.ppt)
5NWPMLP model KS test
Grahic Jump Location
Fig. 9

Monthly performance of different forecast models

+Monthly performance of different forecast models
View Large  |  Save Figure  |  Download Slide (.ppt)
Monthly performance of different forecast models

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Comparisons of Gaussian ARTMAP and Distributed Gaussian ARTMAP Classifiers: The Category Proliferation Problem
Intelligent Engineering Systems through Artificial Neural Networks, Volume 16
A Novel Approach for LFC and AVR of an Autonomous Power Generating System
International Conference on Mechanical Engineering and Technology (ICMET-London 2011)
Fitting a Function and Its Derivative
Intelligent Engineering Systems Through Artificial Neural Networks, Volume 17
A Semi-Adaptive Fractional Order PID Control Strategy for a Certain Gun Control Equipment
International Conference on Instrumentation, Measurement, Circuits and Systems (ICIMCS 2011)
Using Neural Networks for Dynamic Vehicle Navigation Using Integrated GNSS / INS Your Romania
International Conference on Instrumentation, Measurement, Circuits and Systems (ICIMCS 2011)
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In