Search

China's recently announced directive on tackling climate change, namely, to reach carbon peak by 2030 and to achieve carbon neutrality by 2060, has led to an unprecedented nationwide response among the academia and industry. Under such a directive, a rapid increase in the grid penetration rate of solar in the near future can be fully anticipated. Although solar radiation is an atmospheric process, its utilization, as to produce electricity, has hitherto been handled by engineers. In that, it is thought important to bridge the two fields, atmospheric sciences and solar engineering, for the common good of carbon neutrality. In this überreview, all major aspects pertaining to solar resource assessment and forecasting are discussed in brief. Given the size of the topic at hand, instead of presenting technical details, which would be overly lengthy and repetitive, the overarching goal of this review is to comprehensively compile a catalog of some recent, and some not so recent, review papers, so that the interested readers can explore the details on their own. [ABSTRACT FROM AUTHOR]

Site adaptation refers to procedures for correcting systematic errors in an extended period of gridded modeled data using a short period of ground-based measurements used as unbiased reference. Traditionally, site adaptation leverages a single gridded product and issues point predictions. Currently, remote-sensed and reanalysis data are available from different sources providing multiple versions of estimates of a same atmospheric variable, for any location on Earth. These datasets allow what is called an ensemble prediction. In this regard, this contribution proposes a probabilistic site-adaption framework, and describes how one can use parametric and nonparametric techniques within the framework. On top of the stand-alone probabilistic site-adaption methods, heuristics are optionally used to combine quantiles, to further improve the accuracy of site adaptation. To exemplify the framework, global horizontal irradiance data from 26 sites worldwide with different climate characteristics and weather regimes are used to side-adapt the corresponding predictions from up to 5 satellite-derived databases and 2 reanalyses spanning various periods, collectively. It is found that the proposed site-adaptation methods using multiple gridded products are able to attain, on average, a 5 W/m 2 reduction in continuous ranked probability score than that leveraging just a single product. [ABSTRACT FROM AUTHOR]

The National Solar Radiation Database (NSRDB) offers satellite-derived irradiance over most of the Americas. The latest version of the NSRDB hosts broadband irradiance estimates at 5-min–2-km resolution, which have improved from the previous 30-min–4-km resolution. Whereas this new dataset is expected to continue providing invaluable solar resource information to researchers and practitioners worldwide, its accuracy needs to be validated. Validating high-resolution data must go beyond reporting the overall error metrics. To that end, time series features are used here to study the correspondence between the satellite-derived and measured irradiance transients. This paper also provides an in-depth discussion on the data quality issues observed during the validation. [ABSTRACT FROM AUTHOR]

• The most comprehensive validation of satellite-derived and reanalysis irradiance by far. • Hourly gridded irradiance from 8 products are validated at 57 BSRN stations worldwide. • The entire temporal record of BSRN, over 27 years, is used. • Murphy–Winkler distribution-oriented verification framework is emphasized. Gridded solar radiation products, namely satellite-derived irradiance and reanalysis irradiance, are key to the next-generation solar resource assessment and forecasting. Since their accuracies are generally lower than that of the ground-based measurements, providing validation of the gridded solar radiation products is necessary in order to understand their qualities and characteristics. This article delivers a worldwide validation of hourly global horizontal irradiance derived from satellite imagery and reanalysis. The accuracies of 6 latest satellite-derived irradiance products (CAMS-RAD, NSRDB, SARAH-2, SARAH-E, CERES-SYN1deg, and Solcast) and 2 latest global reanalysis irradiance products (ERA5 and MERRA-2) are verified against the complete records from 57 BSRN stations, over 27 years (1992–2018). This scope of validation is unprecedented in the field of solar energy. Moreover, the importance of using distribution-oriented verification approaches is emphasized. Such approaches go beyond the traditional measure-oriented verification approach, and thus can offer additional insights and flexibility to the verification problem. [ABSTRACT FROM AUTHOR]

• Ensemble modeling for 1-min diffuse fraction is discussed. • The raw ensemble is post-processed using ensemble model output statistics (EMOS). • EMOS outperforms the best stand-alone models such as Yang2 and Engerer2. • EMOS is superior to simple model blending in terms of CRPS and ignorance score. Separation models split diffuse and direct components of solar radiation from the global horizontal radiation. At the moment, all separation models only issue predictions that are deterministic (as opposed to probabilistic). Since the best prediction is necessarily probabilistic, a parametric post-processing framework called the ensemble model output statistics (EMOS) is introduced in this paper, to make probabilistic predictions. EMOS takes the diffuse fractions predicted by an ensemble of existing 1-min separation models, and outputs a predictive distribution, with parameters optimized by maximum likelihood estimation. Clearly, the EMOS-based separation modeling goes beyond the current literature, in terms of uncertainty quantification. Eight popular separation models from the literature, with different architectures, are used to demonstrate the predictive power of EMOS. Using 1-min high-quality radiometric data from seven stations in the USA and four stations in Europe, it is found that Y ang2 is the best stand-alone model with an average RMSE of 21.8%, in terms of direct normal irradiance prediction, contrasting the 26.3% of the previously reported best model, namely, E ngerer2. On the other hand, the EMOS post-processed predictions have an average RMSE of 20.8%, which is lower than that of the best stand-alone model. Moreover, EMOS is shown superior to simple model averaging, in terms of continuous ranked probability score and ignorance score. [ABSTRACT FROM AUTHOR]

As a crucial component of the model chain, which facilitates irradiance-to-power conversion during solar resource assessment and forecasting, separation modeling continues to draw attention in both academia and industry. However, when evaluating even the best separation model today, one can quickly recognize its limited accuracy compared to other energy meteorology models such as transposition models. The task of separating global horizontal irradiance into diffuse and beam components does not seem soluble by any derivative effort aimed at tweaking the existing semi-physical models. As a result, an appealing alternative is to consider end-to-end data-driven models, which have demonstrated predictive capability in scenarios where the volume of data is substantial and the interaction among variables is complex. This work discusses the separation of 1-min irradiance from a data-driven perspective. In this preliminary study, a total of 10 representative data-driven separation models are developed and compared to the state-of-the-art semi-physical models, using a comprehensive 1-min irradiance database that spans five years and covers numerous climate types. The average error of the data-driven models is found to be 15.2% to 22.6% lower than that of the semi-physical models for training locations and 7.9% to 17.6% lower for completely unseen locations. Data-driven models also have significantly lower standard deviations (up to 87.2% even for completely unseen locations), highlighting their robustness. In addition, this work provides a guideline for choosing between data-driven and semi-physical models based on data availability, application needs, computational resources, interpretability, and model adaptability. Furthermore, the study underscores the challenges in accurately predicting the diffuse fraction using available input features and indicates that the incorporation of additional weather-related variables and domain knowledge could enhance the performance of data-driven separation models. • Irradiance separation using end-to-end data-driven techniques. • Thorough comparison of data-driven models with semi-physical models. • Data-driven models demonstrate enhanced overall performance. • In-depth examination of key meteorological features and their relationships. [ABSTRACT FROM AUTHOR]

Since directly measuring beam and diffuse irradiance is not feasible on many occasions, one often has to resort to estimating the beam and diffuse irradiance components from the global irradiance, which is known as separation modeling. Separation modeling is essentially a nonlinear regression problem, with the clearness index being the main input and the diffuse fraction being the output. Hundreds of separation models with various complexities have been proposed, among which the Yang4 model was recently validated using worldwide data as the quasi-universal choice for 1-min data. In this work, Yang4 is further improved by regime-dependent fitting, i.e., fitting a separate set of model coefficients for each climatological regime. Different regimes are determined through clustering of cloud cover frequency, aerosol optical depth, and surface albedo climatology maps. The new Yang5 model is able to outperform its predecessor at the 126 stations tested, covering a wide range of climate types. Overall, the normalized root mean square errors for beam normal irradiance (BNI) and diffuse horizontal irradiance (DHI) of Yang5 are 17.55% and 32.92% on average, as compared to 19.13% and 34.94% for the next best model, namely, Yang4. Furthermore, through conducting pairwise Diebold–Mariano tests, Yang5 is shown superior to Yang4 at 110/126 sites for BNI prediction and 93/126 for DHI. [Display omitted] • A concise summary of recent advances in separation modeling is given. • A regime-dependent version of the Yang4 model is proposed. • Regimes are defined by clustering cloud, aerosol, and albedo climatology. • The new Yang5 model outperforms the quasi-universal Yang4. [ABSTRACT FROM AUTHOR]

• Developing an R and Python package to enhance the openness and accessibility of publicly available solar datasets. • Providing access to solar data with multiple types, including sky images, behind-the-meter photovoltaic generation, etc. • Demonstrating the package usage by self-contained scripts. Observational solar data is the foundation of data-driven research in solar power grid integration and power system operations. Compared to other fields in data science, the openness and accessibility of solar data is lacking, which prevents solar data science from catching up with the emerging trends of data science (e.g., deep learning). In this paper, OpenSolar, a package with both R and Python versions, is developed to enhance the openness and accessibility of publicly available solar datasets. The OpenSolar package provides access to multiple types of solar data, primarily from four datasets: (1) the National Renewable Energy Laboratory (NREL) Solar Power Data for Integration Studies dataset, (2) the NREL Solar Radiation Research Laboratory dataset, (3) the Sheffield Solar-Microgen database, and (4) the Dataport database. Unlike other open solar datasets that only contain meteorological data, the four datasets in the OpenSolar package also contain behind-the-meter power data, sky images, and solar power data with satisfactory temporal and spatial resolution and coverage. The overview, quality-control methods, and potential usage of the datasets, in conjunction with sample code implementing the OpenSolar functions, are described in this paper. The package is expected to assist in bridging the gaps between the research fields of solar energy, power systems, and data science. [ABSTRACT FROM AUTHOR]

• An update to the previously published R package "SolarData" is provided. • Functions to read, quality control, and aggregate the BSRN data are included in this update. • As compared to the BSRN-Toolbox, these functions provide additional flexibility in data handling. Previously in "SolarData: An R package for easy access of publicly available solar datasets" [Sol. Energy 171 (2017) A3–A12], the R package SolarData was built for easy access of five publicly available solar datasets. In this update, code for reading data from the Baseline Surface Radiation Network (BSRN), the largest research-grade solar radiation monitoring network, is added to the package. BSRN comprises 66 stations (as of 2019-02) around the globe, which collect 1-min or 3-min radiation data since 1992. These data are stored in the so-called "station-to-archive" files, each containing records from one month and one station, and can be downloaded via ftp. The functions in SolarData v1.1 directly interact with these station-to-archive files, without using the BSRN-Toolbox. It should however be noted that SolarData is not a replacement of BSRN-Toolbox. Instead, it gives R users improved accessibility and freedom in BSRN data manipulation. Like the previous release, all contents herein described are made available on GitHub. [ABSTRACT FROM AUTHOR]

Highlights • A universal benchmarking model called "complete-history persistence ensemble" is proposed. • CH-PeEn is independent of model parameters, forecast horizon, and forecast lead time. • For a given site and temporal resolution of data, the CRPS calculated using CH-PeEn is approximately constant. • Skill scores computed based on the CRPS of CH-PeEn are more amenable for forecast comparison. Abstract Probabilistic solar irradiance forecasting is often benchmarked using the clear-sky persistence ensemble (PeEn). By comparing the continuous ranked probability score (CRPS) of a forecasting model to that of PeEn, the skill score can be obtained. Such skill score can be interpreted as the percentage improvement over the baseline model—PeEn. However, the CRPS of PeEn depends heavily on the model parameters and forecast setup, e.g., the number of ensemble members. The skill score is meant to provide a possibility for universal forecast comparison, but because of the different PeEn implementations, the score can be hard, if not impossible, to interpret. On this point, the complete-history PeEn (CH-PeEn) is herein proposed as a universal benchmarking method for probabilistic solar forecasting. CH-PeEn utilizes the entire history of measurements, and forms empirical distributions of the forecast clear-sky index that only depend on the time of day. The CRPS calculated based on CH-PeEn only depends on the location and temporal resolution of the data, not on forecast horizon nor lead time. Hence, CH-PeEn can lead to a near unique CRPS, and such uniqueness greatly improves the interpretability of the skill scores. [ABSTRACT FROM AUTHOR]

Separation models predict diffuse horizontal irradiance from other meteorological parameters such as the global horizontal irradiance or zenith angle. From a mathematical point of view, the separation modeling problem is a regression, where the regressors are observed or computed variables and the regressand is the unobserved diffuse fraction. The most successful minute-resolution separation model prior to 2016 was proposed by Engerer, which is constructed using a trend component (cloud enhancement) and a main effect (logistic function). Subsequently, the Starke model published in 2018 further improved the Engerer model. It is herein shown that the logistic-function type of model, and many other separation models, can be transformed into a (condition-based) multiple linear regression problem. Under this transformation framework, two new models are proposed, which strictly dominate the performance of the Engerer model and the Starke model, at all 7 test sites across the continental United States, making them probably the most accurate separation models to date. The new models are also tested at 5 European sites with unseen data (i.e., not involved during model parameter fitting); their performance again dominates all benchmarking models. The new separation models leverage half-hourly satellite-derived diffuse fraction. Since satellite data are available globally, the satellite-augmented separation models have universal applicability. However, despite their good performance, empirical separation models suffer from a series of issues. Hence, models driven by atmospheric physics are the "true gems" that one should pursue. [ABSTRACT FROM AUTHOR]

Satellite-derived irradiance data, as an alternative to ground-based measurements, offer a unique opportunity to verify gridded solar forecasts generated by a numerical weather prediction model. Previously, it has been shown that the mean square errors (MSE) evaluated against ground-based measurements and satellite-derived solar irradiance are comparable, which might warrant the use of satellite-based products for regional forecast verification. In this paper, the 24-h-ahead hourly forecasts issued by the North American Mesoscale forecast system are verified against both ground-based (Surface Radiation Budget Network, or SURFRAD) and satellite-based (National Solar Radiation Data Base, or NSRDB) measurements, at all 7 SURFRAD stations over 2015–2016. Three different MSE decomposition methods are used to characterize—e.g., through association, calibration, refinement, resolution, or likelihood—how well the two types of measurements can gauge the forecasts. However, despite their comparable MSEs, NSRDB is found suboptimal in its ability to verify forecasts as compared to SURFRAD. Nonetheless, if a new forecasting model produces significantly better forecasts than the benchmarking model, satellite-derived data are able to detect such improvements and make conclusions. This article comes with supplementary material (data and code) for reproducibility. [ABSTRACT FROM AUTHOR]

Abstract The physical solar model (PSM) data from the National Solar Radiation Data Base offers half-hourly and hourly satellite-derived irradiance data on a regular 0.04 ° × 0.04 ° grid over most of America, from 1998 to 2016. Previously in "The National Solar Radiation Data Base (NSRDB)" [Renewable & Sustainable Energy Reviews 89 (2018) 51–60] and "Evaluation of the National Solar Radiation Database (NSRDB): 1998–2015" [National Renewable Energy Laboratory technical report NREL/TP-5D00-67722], the accuracy of the PSM data is validated against the ground data from the SURFRAD network. Unfortunately, those validations appear to be incorrect, most likely due to a data-aggregation issue. Whereas the PSM data is valuable for solar system design, simulation and evaluation, the high data uncertainty reported limits the confidence in using the dataset. The validation of the PSM data is herein reiterated. It is found that the actual uncertainty is much lower than those reported earlier. Highlights • NSRDB PSM v3 satellite-derived irradiance is validated against SURFRAD ground-based measurements. • The actual uncertainties of the hourly data are much smaller than the previously reported values. • The three commonly used data-aggregation schemes are discussed in a solar engineering context. [ABSTRACT FROM AUTHOR]

Highlights • Kriging ensemble is proposed for spatial prediction problems. • Kriging ensemble is validated using both geostatistical and lattice data. • Similar to time series ensemble, kriging ensemble improves prediction accuracy. • Similar to time series ensemble, using kriging ensemble is less risky than choosing the best component model. Abstract It is surprising how a commonly used concept in temporal prediction—combining forecasts, or rather combining predictions—has not really been brought forward in spatial prediction. Analogous to forecasting, where forecasts made using models such as exponential smoothing or neural networks are combined through regressions, the various prediction combination methods are herein transferred to spatial prediction problems. Through a series of empirical studies, the advantage and potential of kriging ensemble, or more generally, spatial-interpolator ensemble, are demonstrated. Both geostatistical and lattice data (solar irradiance) are considered. Although in theory, the improvement in predictive performance is not guaranteed, just like how we cannot guarantee that ensemble improves forecasts, in practice, a validated ensemble performs at least as good as the best component model, just like how the ensembles in forecasting would behave. [ABSTRACT FROM AUTHOR]

Highlights • NSRDB PSM v3 satellite-derived irradiance is validated against SURFRAD ground-based measurements. • Kriging is used to perform spatial prediction with various dimension-reduced lattices. • The dimension-reduced data (100 times sparser) is as good as the original data in terms of nRMSE and nMBE. Abstract The National Solar Radiation Database (NSRDB) has provided solar resource data for over 25 years. Its most recent update, namely, physical solar model (PSM) version 3, generates half-hourly, gridded, satellite-derived irradiance data with a spatial resolution of 4 km × 4 km, covering most of America. The total volume of the data is over 40 TB. Since the main method to access the data is using API with a daily limit of 2000 requests, it would require 1000 days to obtain one year of PSM data from approximately 2 million pixels. Furthermore, such a big dataset is difficult to store and manipulate. In this regard, this paper empirically investigates the accuracies of various kriging methods, so that a suitable, dimension-reduced (in space) dataset can be opted during various spatio-temporal analyses, such as forecasting or monitoring network design. [ABSTRACT FROM AUTHOR]

Text mining is an emerging topic that advances the review of academic literature. This paper presents a preliminary study on how to review solar irradiance and photovoltaic (PV) power forecasting (both topics combined as “solar forecasting” for short) using text mining, which serves as the first part of a forthcoming series of text mining applications in solar forecasting. This study contains three main contributions: (1) establishing the technological infrastructure (authors, journals & conferences, publications, and organizations) of solar forecasting via the top 1000 papers returned by a Google Scholar search; (2) consolidating the frequently-used abbreviations in solar forecasting by mining the full texts of 249 ScienceDirect publications; and (3) identifying key innovations in recent advances in solar forecasting (e.g., shadow camera, forecast reconciliation). As most of the steps involved in the above analysis are automated via an application programming interface, the presented method can be transferred to other solar engineering topics, or any other scientific domain, by means of changing the search word. The authors acknowledge that text mining, at its present stage, serves as a complement to, but not a replacement of, conventional review papers. [ABSTRACT FROM AUTHOR]

To integrate solar energy into the electricity grid reliably and efficiently, different forecasting techniques are used to produce forecasts at various forecast horizons using data of different levels of aggregation. For example, power generated by distributed photovoltaics (PV) can be disaggregated in a geographical hierarchy into transmission zones, distribution nodes, PV plants, subsystems and inverters; forecasts are required at all of these levels to facilitate different power system operations and power plant management. Due to the different information sets visible to different players in the hierarchy, such as PV plant owners and independent system operators, the forecasts produced at different aggregation levels are often not optimal. Furthermore, these forecasts are aggregate inconsistent, i.e., forecasts made using data collected at lower levels do not add up exactly to the forecasts made using higher level data. In this paper, the state-of-the-art forecast reconciliation techniques are explored. By minimizing the trace of forecast error covariance matrix, the base forecasts obtained across the hierarchy can be optimally reconciled. The reconciled forecasts not only aggregate consistently, and thus lead to collaborative decision making, but also improve the base forecasts at each level significantly. The merit of the reconciliation framework goes to the fact that it does not require any additional information other than those base forecasts. Furthermore, forecast improvements are independent of the base forecasting methods. In other words, once the base forecasts are generated with our favorite methods, the reconciliation will most likely improve those forecasts further. The reconciliation techniques can be applied to a variety of hierarchies with different forecast horizons. The empirical part of this work considers two examples, namely, an NWP-based day-ahead forecast reconciliation over 318 simulated PV plants in the state of California, and a spatio-temporal statistical 1-min-ahead forecast reconciliation over 17 irradiance sensors on the Oahu Island, Hawaii. It is shown that reconciliation benefits both the gird operators and PV system owners by bringing more accurate forecasts, and thus motivates information sharing in an electricity grid. [ABSTRACT FROM AUTHOR]

Predicting solar radiation on inclined surfaces is a critical task for photovoltaic energy systems design, simulation and performance evaluation. Many transposition models have been proposed in the literature; and there are abundant evaluation studies. However, these models are sometimes used incorrectly. Moreover, these errors tend to propagate through the literature. This paper aims to identify and correct some errors in the literature. It also provides a benchmark on transposition model accuracies. Twenty-six models are described using a consistent nomenclature. Model performance is ranked and pairwise accuracies are evaluated with one year of solar irradiance measurements from four locations. Although no universal model is found in this study, some are recommended. The paper comes with computer code and a small portion of experimental data. [ABSTRACT FROM AUTHOR]

In modern smart grids and deregulated electricity markets, accurate forecasting of solar irradiance is critical for determining the total energy generated by PV systems. We propose a mixed wavelet neural network (WNN) in this paper for short-term solar irradiance forecasting, with initial application in tropical Singapore. The key advantage of using wavelet transform (WT) based methods is the high signal compression ability of wavelets, making them suitable for modeling of nonstationary environmental parameters with high information content, such as short timescale solar irradiance. In this WNN, a combination of the commonly known Morlet and Mexican hat wavelets is used as the activation function for hidden-layer neurons of a feed forward artificial neural network (ANN). To demonstrate the effectiveness of the proposed approach, hourly predictions of solar irradiance, which is an aggregate sum of irradiance value observed using 25 sensors across Singapore, are considered. The forecasted results show that WNN delivers better prediction skill when compared with other forecasting techniques. [ABSTRACT FROM AUTHOR]

Several variants of spatio-temporal kriging are used to perform very short-term solar irradiance forecasting by utilizing data from a sensor network. Kriging can produce forecasts not only at the locations of the irradiance monitoring stations, but also at locations where sensors are not installed. Leave-one-out cross-validation is used to test the kriging performance at unobserved locations. Kriging weights are determined either empirically or using a correlation function. Four parametric correlation functions (correlograms) are herein considered, namely, separable, fully symmetric, and two polynomial-adjusted correlation functions. A dense 1 km × 1.2 km network of 17 stations located on Oahu island, Hawaii, is used in this paper. We find that kriging based on a polynomial-adjusted correlation function (the best among the parametric models) is able to obtain forecast skill up to 0.43 and 0.36 for observed and unobserved locations respectively, for a forecast horizon of 50 s. It is also found that empirical kriging performs better than parametric models at small forecast horizons (such as 30 s). However, it loses accuracy for forecast horizons longer than 100 s. [ABSTRACT FROM AUTHOR]

We find an application of the lasso (least absolute shrinkage and selection operator) in sub-5-min solar irradiance forecasting using a monitoring network. Lasso is a variable shrinkage and selection method for linear regression. In addition to the sum of squares error minimization, it considers the sum of ℓ 1 -norms of the regression coefficients as penalty. This bias–variance trade-off very often leads to better predictions. One second irradiance time series data are collected using a dense monitoring network in Oahu, Hawaii. As clouds propagate over the network, highly correlated lagged time series can be observed among station pairs. Lasso is used to automatically shrink and select the most appropriate lagged time series for regression. Since only lagged time series are used as predictors, the regression provides true out-of-sample forecasts. It is found that the proposed model outperforms univariate time series models and ordinary least squares regression significantly, especially when training data are few and predictors are many. Very short-term irradiance forecasting is useful in managing the variability within a central PV power plant. [ABSTRACT FROM AUTHOR]

We forecast hourly solar irradiance time series using a novel hybrid model based on SOM (self-organizing maps), SVR (support vector regression) and PSO (particle swarm optimization). In order to solve the noise and stationarity problems in the statistical time series forecasting modelling process, SOM is applied to partition the whole input space into several disjointed regions with different characteristic information on the correlation between the input and the output. Then SVR is used to model each disjointed regions to identify the characteristic correlation. In order to reduce the performance volatility of SVM (support vector machine) with different parameters, PSO is implemented to automatically perform the parameter selection in SVR modelling. This hybrid model has been used to forecast hourly solar irradiance in Colorado, USA and Singapore. The technique is found to outperform traditional forecasting models. [ABSTRACT FROM AUTHOR]

The Perez irradiance model offers a practical representation of solar irradiance by considering the sky hemisphere as a three-part geometrical framework, namely, the circumsolar disc, the horizon band and the isotropic background. Furthermore, the simplified Perez diffuse irradiance model, commonly known as the Perez transposition model, is one of the most widely adopted models in tilted irradiance modeling. Although the set of model coefficients reported by Perez et al. (1990) is considered to be at an asymptotic level of optimization, later analyses have shown that coefficients which are adjusted to local conditions may perform better than the original set. The model coefficients can be adjusted locally based on multiple datasets of diffuse and global irradiance on tilted and horizontal planes. In this paper, we present a different approach to adjust the coefficients, by using only measurements of global irradiance on tilted and horizontal planes from a tropical climate site, Singapore. A complete set of mathematical solutions to the inverse problem, i.e., irradiance transposition from tilt to horizontal, is also proposed. The data can then be used to generate irradiance maps from in-plane irradiance measurements at photovoltaics (PV) systems. Such maps provide relevant information for PV grid integration. [ABSTRACT FROM AUTHOR]

Highlights: [•] We propose two statistics to quantify the threshold distance within a solar irradiance random field. [•] The determined threshold distance is used as a parameter shrinkage tool for solar irradiance forecasting. [•] The shrunken models improve the predictive performances of the spatio-temporal models. [Copyright &y& Elsevier]

Highlights: [•] Satellite image analysis is performed and cloud cover index is classified using self-organizing maps (SOM). [•] The ESSS model is used to forecast cloud cover index. [•] Solar irradiance is estimated using multi-layer perceptron (MLP). [•] The proposed model shows better accuracy than other investigated models. [Copyright &y& Elsevier]

Abstract: Electricity power grid operations require information about demand and supply on a variety of timescales and areas. The advent of significant generation contributions by time variable renewable energy sources means that forecasting methods are increasingly required. Some of the earliest requirements will be for spatial-temporal estimation of solar irradiance and the resulting photovoltaic-generated electricity. Accurate forecasts represent an important step towards building a smart grid for renewable energy driven cities or regions, and to this end we develop forecasting tools that use data from ground-based irradiance sensors. Spatial-temporal datasets that enjoy the properties of stationarity, full symmetry and separability are in general more amenable to forecasting using time-forward kriging algorithms. Usually, none of these properties obtain in meteorological data such as wind velocity fields and solar irradiance distributions. In this paper, we construct a statistical forecast system to mitigate this problem. We first achieve temporal stationarity by detrending solar irradiance time series at individual monitoring stations. We then approximate spatial stationarity through deformations of the geographic coordinates. Various spatial-temporal variance-covariance structures are formed to explore full symmetry and separability. Finally, time-forward kriging is used to forecast the hourly spatial-temporal solar irradiance data from 10 Singapore weather stations. The aim of the proposed system is to forecast irradiance and PV electricity generation at arbitrary spatial locations within a monitored area. [Copyright &y& Elsevier]

Separation models, which are used to split beam and diffuse irradiance components from the global one, constitute the largest class of radiation models. Over the years, there have been more than 150 models proposed, and public views on the ranking of these models have been divergent, due to the climate-, weather-, and sky-condition dependency in model performance. In a study conducted in 2016, 140 separation models have been validated using high-quality radiometry data from 54 research-grade stations worldwide, which offer an objective and comprehensive assessment of the then available models. It was found that the Engerer2 model had the best overall performance. Since 2016, numerous other models have been proposed, and most of them are able to claim superiority over Engerer2 , once again making the question "what is the best separation model to date" relevant. On this point, this article first reviews these latest advances in separation modeling. Next, as to promote fair comparison, an exceedingly comprehensive benchmarking dataset, which consists of 5 years (2016–2020) of 1-min data, from 126 stations located in all 7 continents and on islands in all 4 oceans, is considered in the empirical part of the article. With this dataset, which has more than 80 million valid 1-min data points, 10 latest separation models with diverse modeling philosophies are compared. It is found that the Yang4 model has the best overall performance, and thus is able to replace Engerer2 as the new quasi-universal separation model. • A review on the latest advancements on separation modeling is presented. • A comprehensive dataset with 126 sites worldwide is used in this research. • An extensive worldwide comparison of 10 latest models is conducted. • Yang4 model is able to replace Engerer2 as the new quasi-universal model. [ABSTRACT FROM AUTHOR]

Highlights: [•] We propose a model to convert global solar irradiance from tilted planes to horizontal. [•] Ten transposition models for conversion from horizontal to tilt are validated using tropical data. [•] Five decomposition models are validated for the tropics. [•] Decrease in zenith angle can improve the conversion accuracy. [Copyright &y& Elsevier]

Abstract: We forecast high-resolution solar irradiance time series using an exponential smoothing state space (ESSS) model. To stationarize the irradiance data before applying linear time series models, we propose a novel Fourier trend model and compare the performance with other popular trend models using residual analysis and the Kwiatkowski-Phillips-Schmidt-Shin (KPSS) stationarity test. Using the optimized Fourier trend, an ESSS model is implemented to forecast the stationary residual series of datasets from Singapore and Colorado, USA. To compare the performance with other time series models, autoregressive integrated moving average (ARIMA), linear exponential smoothing (LES), simple exponential smoothing (SES) and random walk (RW) models are tested using the same data. The simulation results show that the ESSS model has generally better performance than other time series forecasting models. To assess the reliability of the forecasting model in real-time applications, a complementary study of the forecasting 95% confidence interval and forecasting horizon of the ESSS model has been conducted. [Copyright &y& Elsevier]

Abstract: We apply time series analysis to forecast next hour solar irradiance including cloud cover effects. Three forecasting methods are proposed using different types of meteorological data as input parameters, namely, global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI), direct normal irradiance (DNI) and cloud cover. The first method directly uses GHI to forecast next hour GHI through additive seasonal decomposition followed by an Auto-Regressive Integrated Moving Average (ARIMA) model. The second method forecasts DHI and DNI separately using additive seasonal decomposition followed by an ARIMA model and then combines the two forecasts to predict GHI using an atmospheric model. The third method considers cloud cover effects. An ARIMA model is used to predict cloud transients. GHI at different zenith angles and under different cloud cover conditions is constructed using nonlinear regression, i.e., we create a look-up table of GHI regression models for different cloud cover conditions. All three methods are tested using data from two weather stations in the USA: Miami and Orlando. It is found that forecasts using cloud cover information can improve the forecast accuracy. [Copyright &y& Elsevier]

Abstract: We analyse Singapore global horizontal irradiance (GHI) measurements to estimate clear sky GHI in the equator region. The data consists of a one-year period of clear sky GHI at three solar irradiance monitoring stations in Singapore, namely, First Zero Energy House in Singapore (1.3094∘N, 103.9160∘E) located in the east of the country, Solar Energy Research Institute of Singapore (1.3007∘N, 103.7718∘E) in the mid-west, and Nanyang Technology University Intelligence Systems Center (1.3436∘N, 103.6792∘E) in the west. Several empirical clear sky models are considered for a Singapore case study. An regression method is proposed to parameterise the model of choice for Singapore''s measured GHI. The developed model is validated using GHI measurements from difference stations over different time periods. [Copyright &y& Elsevier]

With the objective of producing high-resolution and high-accuracy maps of mean annual irradiance at country scale, this contribution exploits the complementary properties of two distinct sources of solar irradiance data: gridded modeled data derived from satellite observations, and station-specific typical meteorological year (TMY) data. A data fusion procedure based on Gaussian process modeling is used to optimally combine the two sources of data and derive solar resource maps. Gridded physical solar model version 3 (PSM3) satellite-derived data at 4-km resolution and TMY3 data from 67 stations in California are used to produce a map of mean annual global horizontal irradiance at 2-km resolution and exemplify the procedure. It is shown that by integrating the PSM3 data with TMY3 data, the original 5.2% mean error in the PSM3 map is reduced to 1.6%. The demonstrated approach is suitable for a variety of regional-scale applications for which both high-resolution data of low accuracy and low-resolution measurements of high accuracy are available. • High- and low-accuracy measurements are integrated through Gaussian process modeling. • A few high-accuracy measurements are used to correct the many low-accuracy ones. • Gaussian process provides flexible fusion of data with high- and low-accuracy counterparts. • An empirical study on solar radiation mapping is presented. [ABSTRACT FROM AUTHOR]

• A novel method to optimally adapt the Heliosat-2 parameters is developed. • Region-adapted Heliosat-2 is applied to China's FY-4A satellite. • GHI with a spatio-temporal resolution of 0.5 km and 15 min is derived. • Sub-kilometer spatial distribution of solar resource in China is obtained. It is, now, universally accepted that modern solar resource assessment should proceed with acquiring satellite-derived irradiance; the underlying problem is one of retrieval. The Advanced Geostationary Radiation Imager (AGRI), which is one of the main payloads of China's latest-generation geostationary satellite series, namely, Fengyun-4 (FY-4), offers such an opportunity, albeit its full potential remains largely unexplored. On this account, a region-adapted semi-empirical model based on Heliosat-2 is herein developed to estimate global horizontal irradiance (GHI) at the native resolution of FY-4 satellite (0.5 km, 15 min), using AGRI reflectance measurement at 640 nm. The adaptation of Heliosat-2 parameters to China's climatology and geography is based on a research-grade ground-based radiometry network with 38 stations evenly distributed across the country. Model validation is conducted against ground-based observations made at 34 unseen stations, over the course of one year (2018). As compared to the 5-km Himawari-8 radiation product, the developed product not only has a 10-fold refinement in resolution, which can clearly depict those small-scale variations in solar resources that are not observable from the Himawari-8 product, but also attains a substantially lower bias (1.4% versus 7.6% under clear skies and –0.3% versus 9.5% under cloudy skies). This ultra-high-resolution product is of vital importance to solar resource assessment on a scale consistent with the size of typical utility-scale photovoltaic plants. The product is also of great value for agricultural and ecological studies, which often depends upon high-quality solar radiation at a high spatio-temporal resolution. [ABSTRACT FROM AUTHOR]

With an ever-increasing share of intermittent renewable energy in the world's energy mix, there is an increasing need for advanced solar power forecasting models to optimize the operation and control of solar power plants. In order to justify the need for more elaborate forecast modeling, one must compare the performance of advanced models with naïve reference methods. On this point, a rigorous formalism using statistical tools, variational calculation and quantification of noise in the measurement is studied and five naïve reference forecasting methods are considered, among which there is a newly proposed approach called ARTU (a particular autoregressive model of order two). These methods do not require any training phase nor demand any (or almost no) historical data. Additionally, motivated by the well-known benefits of ensemble forecasting, a combination of these models is considered, and then validated using data from multiple sites with diverse climatological characteristics, based on various error metrics, among which some are rarely used in the field of solar energy. The most appropriate benchmarking method depends on the salient features of the variable being forecast (e.g., seasonality, cyclicity, or conditional heteoroscedasity) as well as the forecast horizon. Hence, to ensure a fair benchmarking, forecasters should endeavor to discover the most appropriate naïve reference method for their setup by testing all available options. Among the methods proposed in this paper, the combination and ARTU statistically offer the best results for the proposed study conditions. • Benchmark of six Statistical Reference Methods (SRM). • Direct multi-step forecast strategy without training phase. • Validation of results using data from multiple climates. • Theory mixing statistical tools, variational calculation and measurement error. • Combination of models and ARTU are the best performing models. [ABSTRACT FROM AUTHOR]