Probabilistic solar nowcasting based on all-sky imagers

Nouri, Bijan; Wilbert, Stefan; Blum, Niklas; Fabel, Yann; Lorenz, Elke; Hammer, Annette; Schmidt, Thomas; Zarzalejo, Luis F.; Pitz‐Paal, Robert

doi:10.1016/j.solener.2023.01.060

Cited by 8 publications

(15 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We briefly summarize the calculation of probabilistic nowcasts (Section 3.3) and present the utilized baseline model (Section 3.3.1), however, for a more detailed description, the interested reader is referred to ref. [22]. Afterward, we present the validation procedure for deterministic and probabilistic nowcasting (Section 3.4).…”

Section: Methodsmentioning

confidence: 99%

“…The applied approach is based on the idea that the primary source of uncertainties in solar nowcasts lies in the variability of prevailing irradiance. [ 22 ] As a nonparametric method, no specific distribution function (e.g., Gaussian) is assumed; instead, the probability distribution for each nowcast is derived from historical prediction errors observed under similar conditions. To identify the prevailing conditions, we employ the identical DNI variability classification as introduced in Section 2.2.…”

Section: Methodsmentioning

confidence: 99%

“…Process flow of quantile‐based postprocessing approach to generate probabilistic nowcasts from deterministic ones as described in ref. [22] As input (orange), the current deterministic nowcast, recent reference measurements, and the look‐up table of quantiles of historical errors are required. Interim results are depicted in blue and the output is the final probabilistic nowcast (green).…”

Section: Methodsmentioning

confidence: 99%

“…Alternatively, quantiles from a distribution of historic prediction errors can be combined with the deterministic forecasts in a postprocessing procedure to generate PIs. [ 22 ] In both cases, parametric, i.e., assuming a distribution function, and nonparametric methods are applicable.…”

Section: Introductionmentioning

confidence: 99%

“…[19] As a result, recent studies also focus on probabilistic forecasts. [20][21][22] There are basically two types of probabilistic methods applied in solar forecasting: quantile and ensemble forecasts. [23] The latter is common practice in NWP, where an ensemble forecast is generated by running the model multiple times with slightly modified initial conditions.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Combining Deep Learning and Physical Models: A Benchmark Study on All‐Sky Imager‐Based Solar Nowcasting Systems

Fabel,

Nouri,

Wilbert

et al. 2024

Solar RRL

Self Cite

View full text Add to dashboard Cite

Intermittent solar irradiance due to passing clouds poses challenges for integrating solar energy into existing infrastructure. By making use of intra‐hour nowcasts (very short‐term forecasts), changing conditions of solar irradiance can be anticipated. All‐sky imagers, capturing sky conditions at high spatial and temporal resolution, can be the basis of such nowcasting systems. In this work, a deep learning (DL) model for solar irradiance nowcasts based on the transformer architecture is presented. The model is trained end‐to‐end using sequences of sky images and irradiance measurements as input to generate point‐forecasts up to 20 min ahead. Further, the effect of integrating this model into a hybrid system, consisting of a physics‐based model and smart persistence, is examined. A comparison between the DL and two hybrid models (with and without the DL model) is conducted on a benchmark dataset. Forecast accuracy for deterministic point‐forecasts is analyzed under different conditions using standard error metrics like root‐mean‐square error (RMSE) and forecast skill (FS). Furthermore, spatial and temporal aggregation effects are investigated. In addition, probabilistic nowcasts for each model are computed via a quantile approach. Overall, the DL model outperforms both hybrid models under the majority of conditions and aggregation effects.This article is protected by copyright. All rights reserved.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%