0
Research Papers

Deterministic and Stochastic Approaches for Day-Ahead Solar Power Forecasting

[+] Author and Article Information
Marco Pierro

Institute for Renewable Energy,
EURAC Research,
Bolzano 39100, Italy;
Department of Enterprise Engineering,
University of Rome Tor Vergata,
Rome 00133, Italy
e-mail: marco.pierro@gmail.com

Francesco Bucci

Department of Enterprise Engineering,
University of Rome Tor Vergata,
Rome 00133, Italy

Matteo De Felice

ENEA,
Casaccia R.C.,
Climate Impacts and Modelling Laboratory,
Rome 00123, Italy

Enrico Maggioni, Alessandro Perotto

IDEAM S.r.l.,
Cinisello Balsamo 20092, Italy

Francesco Spada

IDEAM S.r.l.,
Cinisello Balsamo 20092,Italy

David Moser

Institute for Renewable Energy,
EURAC Research,
Bolzano 39100, Italy

Cristina Cornaro

CHOSE,
Department of Enterprise Engineering,
University of Rome Tor Vergata,
Rome 00133, Italy

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 February 5, 2016; final manuscript received September 6, 2016; published online November 30, 2016. Assoc. Editor: Carlos F. M. Coimbra.

J. Sol. Energy Eng 139(2), 021010 (Nov 30, 2016) (12 pages) Paper No: SOL-16-1066; doi: 10.1115/1.4034823 History: Received February 05, 2016; Revised September 06, 2016

Photovoltaic (PV) power forecasting has the potential to mitigate some of effects of resource variability caused by high solar power penetration into the electricity grid. Two main methods are currently used for PV power generation forecast: (i) a deterministic approach that uses physics-based models requiring detailed PV plant information and (ii) a data-driven approach based on statistical or stochastic machine learning techniques needing historical power measurements. The main goal of this work is to analyze the accuracy of these different approaches. Deterministic and stochastic models for day-ahead PV generation forecast were developed, and a detailed error analysis was performed. Four years of site measurements were used to train and test the models. Numerical weather prediction (NWP) data generated by the weather research and forecasting (WRF) model were used as input. Additionally, a new parameter, the clear sky performance index, is defined. This index is equivalent to the clear sky index for PV power generation forecast, and it is here used in conjunction to the stochastic and persistence models. The stochastic model not only was able to correct NWP bias errors but it also provided a better irradiance transposition on the PV plane. The deterministic and stochastic models yield day-ahead forecast skills with respect to persistence of 35% and 39%, respectively.

FIGURES IN THIS ARTICLE
<>
Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Schematics of the different approaches for day-ahead PV power forecast

Grahic Jump Location
Fig. 2

(a) Daily reference yield (Yr) and final yield (Yf) from Jan. 1, 2011 to Dec. 31, 2014 and (b) monthly average of daily power yield for all the considered years

Grahic Jump Location
Fig. 3

Daily trend of clear sky index (Kcs), performance ratio (PR), and clear sky performance index (PKcs) in a clear sky day and in an overcast day

Grahic Jump Location
Fig. 4

Inputs (X) and output (Y) of the RHNN model

Grahic Jump Location
Fig. 5

Minimum, maximum, and average values of RMSE obtained by different irradiance forecast approaches reported in Ref. [44] and the RMSE achieved by the MOSRH(WRF) solar irradiance prediction for Bolzano site versus the RMSE of the clear sky persistence model

Grahic Jump Location
Fig. 6

Liu–Jordan isotropic transposition errors in different typologies of days identified by the daily clear sky index

Grahic Jump Location
Fig. 7

Hourly trend of power and performance ratio, measured and estimated by the SAPM

Grahic Jump Location
Fig. 8

Accuracy comparison of PV power generation forecast (yearly trends)

Grahic Jump Location
Fig. 9

Accuracy comparison of PV power generation forecast

Grahic Jump Location
Fig. 10

Behavior of residuals of the deterministic and stochastic models: (1) power distribution of hourly errors, (2) bias errors versus rated power (Pm/Pn), and (3) Kolmogorov–Smirnov test

Grahic Jump Location
Fig. 11

(a) Error propagation in the deterministic model, in box (b) error propagation not considering the nominal power degradation of the PV modules

Grahic Jump Location
Fig. 12

(a) RMSE of the isotropic transposition model on actual GHI data (RMSEG0), RMSE of GHI and GPOAI prediction (RMSEG1 and RMSEG2), and the forecast error due to the transposition model (RMSEd = RMSEG2 − RMSEG1) ((b) and (c)) RMSE and MBE comparison of the deterministic and stochastic power forecast in different typologies of days (RMSEP0–MBEP0 of the model Pm(MOSRH) considering only the GHI forecast; RMSEP1–MBEP1 of the full deterministic model Pm(MOSRH + IM + SAPM), and RMSEP2–MBEP2 of the stochastic model Pm(RHNN))

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
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