Research Papers

Direct Power Output Forecasts From Remote Sensing Image Processing

[+] Author and Article Information
David P. Larson

Department of Mechanical and
Aerospace Engineering,
Jacobs School of Engineering,
University of California,
San Diego, La Jolla, CA 92093-0411
e-mail: dplarson@ucsd.edu

Carlos F. M. Coimbra

Department of Mechanical and
Aerospace Engineering,
Jacobs School of Engineering,
University of California,
San Diego, La Jolla, CA 92093-0411
e-mail: ccoimbra@ucsd.edu

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 June 5, 2017; final manuscript received December 14, 2017; published online February 20, 2018. Assoc. Editor: Geoffrey T. Klise.

J. Sol. Energy Eng 140(2), 021011 (Feb 20, 2018) (8 pages) Paper No: SOL-17-1214; doi: 10.1115/1.4038983 History: Received June 05, 2017; Revised December 14, 2017

A direct methodology for intra-day forecasts (1–6 h ahead) of power output (PO) from photovoltaic (PV) solar plants is proposed. The forecasting methodology uses publicly available images from geosynchronous satellites to predict PO directly without resorting to intermediate irradiance (resource) forecasting. Forecasts are evaluated using four years (January 2012–December 2015) of hourly PO data from 2 nontracking, 1 MWp PV plants in California. For both sites, the proposed methodology achieves forecasting skills ranging from 24% to 69% relative to reference persistence model results, with root-mean-square error (RMSE) values ranging from 90 to 136 kW across the studied horizons. Additionally, we consider the performance of the proposed methodology when applied to imagery from the next generation of geosynchronous satellites, e.g., Himawari-8 and geostationary operational environmental satellite (GOES-R).

Copyright © 2018 by ASME
Your Session has timed out. Please sign back in to continue.


Quesada-Ruiz, S. , Chu, Y. , Tovar-Pescador, J. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2014, “ Cloud-Tracking Methodology for Intra-Hour DNI Forecasting,” Sol. Energy, 102, pp. 267–275. [CrossRef]
Marquez, R. , and Coimbra, C. F. M. , 2011, “ Forecasting of Global and Direct Solar Irradiance Using Stochastic Learning Methods, Ground Experiments and the NWS Database,” Sol. Energy, 85(5), pp. 746–756. [CrossRef]
Marquez, R. , Gueorguiev, V. , and Coimbra, C. F. M. , 2012, “ Forecasting of Global Horizontal Irradiance Using Sky Cover Indices,” ASME J. Sol. Energy Eng., 135(1), p. 011017. [CrossRef]
Saade, E. , Clough, D. E. , and Weimer, A. W. , 2013, “ Use of Image-Based Direct Normal Irradiance Forecasts in the Model Predictive Control of a Solar-Thermal Reactor,” ASME J. Sol. Energy Eng., 136(1), p. 010905. [CrossRef]
Perez, R. , Lorenz, E. , Pelland, S. , Beauharnois, M. , Van Knowe, G. , Hemker, K. J. , Heinemann, D. , Remund, J. , Muller, S. C. , Traunmuller, W. , Steinmauer, G. , Pozo, D. , Ruiz-Arias, J. A. , Lara-Fanego, V. , Ramirez-Santigosa, L. , Gaston-Romero, M. , and Pomares, L. M. , 2013, “ Comparison of Numerical Weather Prediction Solar Irradiance Forecasts in the US, Canada and Europe,” Sol. Energy, 94, pp. 305–326. [CrossRef]
Pelland, S. , Galanis, G. , and Kallos, G. , 2013, “ Solar and Photovoltaic Forecasting Through Post-Processing of the Global Environmental Multiscale Numerical Weather Prediction Model,” Prog. Photovolt.: Res. Appl., 21(3), pp. 284–296. [CrossRef]
Larson, D. P. , Nonnenmacher, L. , and Coimbra, C. F. M. , 2016, “ Day-Ahead Forecasting of Solar Power Output From Photovoltaic Plants in the American Southwest,” Renewable Energy, 91, pp. 11–20. [CrossRef]
Pierro, M. , Bucci, F. , De Felice, M. , Maggioni, E. , Perroto, A. , Spada, F. , Moser, D. , and Cornaro, C. , 2016, “ Deterministic and Stochastic Approaches for Day-Ahead Solar Power Forecasting,” ASME J. Sol. Energy Eng., 139(2), p. 021010. [CrossRef]
Aryaputera, A. W. , Yang, D. , and Walsh, W. M. , 2015, “ Day-Ahead Solar Irradiance Forecasting in a Tropical Environment,” ASME J. Sol. Energy Eng., 137(5), p. 051009. [CrossRef]
Cornaro, C. , Bucci, F. , Del Frate, F. , Peronaci, S. , and Taravat, A. , 2015, “ Twenty-Four Hour Solar Irradiance Forecast Based on Neural Networks and Numerical Weather Prediction,” ASME J. Sol. Energy Eng., 137(3), p. 031011. [CrossRef]
Perez, R. , Kivalov, S. , Schlemmer, J. , Hemker , K., Jr. , Renné, D. , and Hoff, T. E. , 2010, “ Validation of Short and Medium Term Operation Solar Radiation Forecasts in the US,” Sol. Energy, 84(12), pp. 2161–2172. [CrossRef]
Marquez, R. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2013, “ Hybrid Solar Forecasting Method Uses Satellite Imaging and Ground Telemetry as Inputs to ANNs,” Sol. Energy, 92, pp. 176–188. [CrossRef]
Inman, R. H. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2013, “ Solar Forecasting Methods for Renewable Energy Integration,” Prog. Energy Combust. Sci., 39(6), pp. 535–576.
Law, E. W. , Prasad, A. A. , Kay, M. , and Taylor, R. A. , 2014, “ Direct Normal Irradiance Forecasting and Its Application to Concentrated Solar Thermal Output Forecasting—A Review,” Sol. Energy, 108, pp. 287–307. [CrossRef]
Antonanzas, J. , Osorio, N. , Urraca, R. , de Pison, F. J. M. , and Antonanzas-Torres, F. , 2016, “ Review of Photovoltaic Power Forecasting,” Sol. Energy, 136, pp. 78–111. [CrossRef]
Nonnenmacher, L. , and Coimbra, C. F. M. , 2014, “ Streamline-Based Method for Intra-Day Solar Forecasting Through Remote Sensing,” Sol. Energy, 108, pp. 447–459. [CrossRef]
Chu, Y. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2013, “ Hybrid Intra-Hour DNI Forecasts With Sky Image Processing Enhanced by Stochastic Learning,” Sol. Energy, 98(Pt. C), pp. 592–603. [CrossRef]
Chu, Y. , Li, M. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2015, “ Real-Time Prediction Intervals for Intra-Hour DNI Forecasts,” Renewable Energy, 83, pp. 234–244. [CrossRef]
Pedro, H. T. C. , and Coimbra, C. F. M. , 2012, “ Assessment of Forecasting Techniques for Solar Power Output With No Exogenous Inputs,” Sol. Energy, 86(7), pp. 2017–2028. [CrossRef]
Zamo, M. , Mestre, O. , Arbogast, P. , and Pannekoucke, O. , 2014, “ A Benchmark of Statistical Regression Methods for Short-Term Forecasting of Photovoltaic Electricity Production, Part I: Deterministic Forecast of Hourly Production,” Sol. Energy, 105, pp. 792–803. [CrossRef]
Zamo, M. , Mestre, O. , Arbogast, P. , and Pannekoucke, O. , 2014, “ A Benchmark of Statistical Regression Methods for Short-Term Forecasting of Photovoltaic Electricity Production—Part II: Probabilistic Forecast of Daily Production,” Sol. Energy, 105, pp. 804–816. [CrossRef]
Perez, R. , Schlemmer, J. , Hemker , Kivalov, S. , Kankiewicz, A. , and Gueymard, C. , 2015, “ Satellite-to-Irradiance Modeling—A New Version of the SUNY Model,” 42nd IEEE Photovoltaic Specialist Conference (PVSC), New Orleans, LA, June 14–19, pp. 1–7.
Perez, R. , Ineichen, P. , Moore, K. , Kmiecik, M. , Chain, C. , George, R. , and Vignola, F. , 2002, “ A New Operational Model for Satellite-Derived Irradiances: Description and Validation,” Sol. Energy, 73, pp. 307–317. [CrossRef]
Perez, R. , Ineichen, P. , Kmiecik, M. , Moore, K. , Renne, D. , and George, R. , 2004, “ Producing Satellite-Derived Irradiances in Complex Arid Terrain,” Sol. Energy, 77(4), pp. 367–371. [CrossRef]
Vignola, V. , Harlan, P. , Perez, R. , and Kmiecik, M. , 2007, “ Analysis of Satellite Derived Beam and Global Solar Radiation Data,” Sol. Energy, 81(6), pp. 768–772. [CrossRef]
Zagouras, A. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2015, “ On the Role of Lagged Exogenous Variables and Spatial-Temporal Correlations in Improving the Accuracy of Solar Forecasting Methods,” Renewable Energy, 78, pp. 203–218. [CrossRef]
Mozorra Aguiar, L. , Pereira, B. , David, M. , Diaz, F. , and Lauret, P. , 2015, “ Use of Satellite Data to Improve Solar Radiation Forecasting With Bayesian Artificial Neural Networks,” Sol. Energy, 122, pp. 1309–1324. [CrossRef]
Chu, Y. , Urquhart, B. , Gohari, S. M. I. , Pedro, H. T. C. , Kleissl, J. , and Coimbra, C. F. M. , 2015, “ Short-Term Reforecasting of Power Output From a 48 MWe Solar PV Plant,” Sol. Energy, 112, pp. 68–77. [CrossRef]
Lipperheide, M. , Bosch, J. L. , and Kleissl, J. , 2015, “ Embedded Nowcasting Method Using Cloud Speed Persistence for a Photovoltaic Power Plant,” Sol. Energy, 112, pp. 232–238. [CrossRef]
Lonij, V. P. , Brooks, A. E. , Cronin, A. D. , Leuthold, M. , and Koch, K. , 2013, “ Intra-Hour Forecasts of Solar Power Production Using Measurements From a Network of Irradiance Sensors,” Sol. Energy, 97, pp. 58–66. [CrossRef]
Kaur, A. , Nonnenmacher, L. , Pedro, H. T. C. , and Coimbra, C. F. M. , 2016, “ Benefits of Solar Forecasting for Energy Imbalance Markets,” Renewable Energy, 86, pp. 819–830. [CrossRef]
Fonseca, J. G. , Oozeki, T. , Ohtake, H. , Shimose, K. , Takashima, T. , and Ogimoto, K. , 2014, “ Regional Forecasts and Smoothing Effect of Photovoltaic Power Generation in Japan: An Approach With Principal Component Analysis,” Renewable Energy, 68, pp. 403–413. [CrossRef]
Wolff, B. , Kuhnert, J. , Lorenz, E. , Krame, O. , and Heinemann, D. , 2016, “ Comparing Support Vector Regression for PV Power Forecasting to a Physical Modeling Approach Using Measurement, Numerical Weather Prediction, and Cloud Motion Data,” Sol. Energy, 135, pp. 197–208. [CrossRef]
Vapnik, V. N. , 1998, Statistical Learning Theory, Wiley, New York.
Chang, C. C. , and Lin, C. J. , 2002, “ Training ν-Support Vector Regression: Theory and Algorithms,” Neural Comput., 14(8), pp. 1959–1977. [CrossRef] [PubMed]
Chang, C. C. , and Lin, C. J. , 2011, “ LIBSVM: A Library for Support Vector Machines,” ACM Trans. Intell. Syst. Technol., 2(3), pp. 1–27. [CrossRef]
Ineichen, P. , and Perez, R. , 2002, “ A New Airmass Independent Formulation for the Linke Turbidity Coefficient,” Sol. Energy, 73(3), pp. 151–157. [CrossRef]
Ineichen, P. , 2006, “ Comparison of Eight Clear Sky Broadband Models against 16 Independent Data Banks,” Sol. Energy, 80(4), pp. 468–478. [CrossRef]
Gueymard, C. A. , 2012, “ Clear-Sky Irradiance Predictions for Solar Resource Mapping and Large-Scale Applications: Improved Validation Methodology and Detailed Performance Analysis of 18 Broadband Radiative Models,” Sol. Energy, 86(8), pp. 2145–2169. [CrossRef]
Marquez, R. , and Coimbra, C. F. M. , 2012, “ Proposed Metric for Evaluation of Solar Forecasting Models,” ASME J. Sol. Energy Eng., 135(1), p. 011016. [CrossRef]
Hintze, J. L. , and Nelson, R. D. , 1998, “ Violin Plots: A Box Plot-Density Trace Synergism,” Am. Stat., 52(2), pp. 181–184.
Zagouras, A. , Inman, R. H. , and Coimbra, C. F. M. , 2014, “ On the Determination of Coherent Solar Microcolimates for Utility Planning and Operators,” Sol. Energy, 102, pp. 173–188. [CrossRef]
Hammer, A. , Kuhnert, J. , Weinreich, K. , and Lorenz, E. , 2015, “ Short-Term Forecasting of Surface Solar Irradiance Based on Meteosat-SEVIRI Data Using a Nighttime Cloud Index,” Remote Sens., 7(7), pp. 9070–9090. [CrossRef]


Grahic Jump Location
Fig. 1

A diagram of the overall forecast methodology. First, a 480 × 680 pixel satellite image at time t is cropped down to a w × w region of interest, centered around the target site. The w × w cropped image is then transformed into a n-length feature vector x(t) and fed into a forecast model, e.g., SVR, to produce a prediction of the target output ŷ(t+Δt).

Grahic Jump Location
Fig. 2

The merged PO and satellite image data sets for the two power plants: (a) CCA and (b) LCC. Two years of data (2012–2013) are designated for training the forecast models (shown in gray), with two additional years (2014–2015) used for testing (shown in black).

Grahic Jump Location
Fig. 3

Effect of image size on forecast accuracy of the CCA training set (2012–2013), measured by RMSE (kW). The mean RMSE for the LS (solid) and SVR (dashed) models across the six horizons (1–6 h ahead) are shown bounded by the minimum and maximum RMSE values. A similar trend is observed for the LCC.

Grahic Jump Location
Fig. 4

Forecast errors on the testing set (2014–2015) for (a) CCA and (b) LCC, grouped by horizon (1–6 h ahead). The distributions for LS and SVR are shown side by side to enable direct comparisons of the two models for the same horizon. The horizontal dashed and dotted lines within each distribution represent the quartiles, with the width approximated using kernel density estimate techniques. Negative MBE values correspond to the model overpredicting the PO.

Grahic Jump Location
Fig. 5

Kernel density estimates of the AE of the 1 h ahead SVR forecasts on the CCA testing set. The AE values are grouped into columns by quartile: (1) first quartile, (2) second quartile, (3) third quartile, and (4) fourth quartile. Darker areas represent higher densities. The top row (Reference) shows the AE densities as functions of the solar azimuthal (ϕ) and zenith (θz) angles when the forecast was generated, while the bottom row (Valid) shows the solar angles of the targeted time for the forecasts. Sunrise and sunset are at θz=90deg while ϕ<180deg corresponds to the morning and ϕ>180deg to the evening. The majority of the largest errors, i.e., the fourth quartile of errors, are produced by forecasts generated near sunrise. A similar error distribution is seen for the LCC site (figure omitted due to length constraints).




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