Search

The parameter optimisation of designed controllers for power systems is always a big concern and needs a lot of effort of researchers especially when the electricity grid becomes larger and more complex. This study proposes a control parameterisation using genetic algorithms (GAs) for power oscillation damping (POD) incorporating synchronous condensers (SCs) via software-in-the-loop simulation to enhance the damping and frequency stability for low-inertia systems. A closed-loop interfaced setup among real-time digital simulator, MATLAB, and OLE for process communication running in real time is analysed and implemented to optimise the POD parameters of a SC. Furthermore, a Prony technique based on the system measurement is applied to find out the frequency and damping ratio of the dominant oscillation mode. The POD optimal parameters are determined by the GA objective function that maximises the damping ratio of the dominant oscillation mode. The effectiveness of the proposed method in damping power oscillations and frequency stability improvement is verified through simulation results of the future western Danish power system. Simulation results demonstrate that the proposed approach offers good performance for parameter optimisation of the POD.

The field of energy forecasting has attracted many researchers from different fields (e.g., meteorology, data sciences, mechanical or electrical engineering) over the last decade. Solar forecasting is a fast-growing subdomain of energy forecasting. Despite several previous attempts, the methods and measures used for verification of deterministic (also known as single-valued or point) solar forecasts are still far from being standardized, making forecast analysis and comparison difficult. To analyze and compare solar forecasts, the well-established Murphy–Winkler framework for distribution-oriented forecast verification is recommended as a standard practice. This framework examines aspects of forecast quality, such as reliability, resolution, association, or discrimination, and analyzes the joint distribution of forecasts and observations, which contains all time-independent information relevant to verification. To verify forecasts, one can use any graphical display or mathematical/statistical measure to provide insights and summarize the aspects of forecast quality. The majority of graphical methods and accuracy measures known to solar forecasters are specific methods under this general framework. Additionally, measuring the overall skillfulness of forecasters is also of general interest. The use of the root mean square error (RMSE) skill score based on the optimal convex combination of climatology and persistence methods is highly recommended. By standardizing the accuracy measure and reference forecasting method, the RMSE skill score allows—with appropriate caveats—comparison of forecasts made using different models, across different locations and time periods. Refereed/Peer-reviewed

This paper presents a simple forecasting database infrastructure implemented using the open-source database management system MySQL. This proposal aims at advancing the myriad of solar and wind forecast models present in the literature into a production stage. The paper gives all relevant details necessary to implement a MySQL infra-structure that collects the raw data, filters unrealistic values, classifies the data, and produces forecasts automatically and without the assistance of any other computational tools. The performance of this methodology is demonstrated by creating intra-hour power output forecasts for a 1 MW photovoltaic installation in Southern California and a 10 MW wind power plant in Central California. Several machine learning forecast models are implemented (persistence, auto-regressive and nearest neighbors) and tested. Both point forecasts and prediction intervals are generated with this methodology. Quantitative and qualitative analyses of solar and wind power forecasts were performed for an extended testing period (4 years and 6 years, respectively). Results show an acceptable and robust performance for the proposed forecasts.

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.

This work proposes forecast models for solar-integrated, utility-scale feeders in the San Diego Gas & Electric operating region. The models predict the net load for horizons ranging from 10 to 30 min. The forecasting methods implemented include hybrid methods based on Artificial Neural Network (ANN) and Support Vector Regression (SVR), which are both coupled with image processing methods for sky images. These methods are compared against reference persistence methods. Three enhancement methods are implemented to further decrease forecasting error: (1) decomposing the time series of the net load to remove low-frequency load variation due to daily human activities; (2) segregating the model training between daytime and nighttime; and (3) incorporating sky image features as exogenous inputs in the daytime forecasts. The ANN and SVR models are trained and validated using six-month measurements of the net load and assessed using common statistic metrics: MBE, MAPE, rRMSE, and forecast skill, which is defined as the reduction of RMSE over the RMSE of reference persistence model. Results for the independent testing set show that data-driven models, with the enhancement methods, significantly outperform the reference persistence model, achieving forecasting skills (improvement over reference persistence model) as large as 43% depending on location, solar penetration and forecast horizons.

We introduce a simple and novel technique to extract dynamic features from sky images in order to increase the accuracy of intrahour forecasts for both Global Horizontal Irradiance (GHI) and Direct Normal Irradiance (DNI) values. The proposed methodology is based on a block-matching algorithm that correctly identifies the bulk motion of clouds relative to the position of the Sun in the sky. Adaptive rectangular- and wedge-shaped Regions Of Interest are used to select the image pixels for the new features. The results show an average increase of 6.8% (6.7%) in forecast skill for GHI (DNI) across all horizons tested as measured against a model with global (nonadaptive) image features. Relative to clear-sky persistence, the new model achieves skills ranging from 20% to 30% (22%–35%) for GHI (DNI), among the highest ever reported for these time horizons. An analysis based on Mutual Information and Pearson correlation coefficients between the image features and the training data reveals overall improvements in all metrics. The proposed adaptive method also improves the predictability of the ramp magnitude and direction.We introduce a simple and novel technique to extract dynamic features from sky images in order to increase the accuracy of intrahour forecasts for both Global Horizontal Irradiance (GHI) and Direct Normal Irradiance (DNI) values. The proposed methodology is based on a block-matching algorithm that correctly identifies the bulk motion of clouds relative to the position of the Sun in the sky. Adaptive rectangular- and wedge-shaped Regions Of Interest are used to select the image pixels for the new features. The results show an average increase of 6.8% (6.7%) in forecast skill for GHI (DNI) across all horizons tested as measured against a model with global (nonadaptive) image features. Relative to clear-sky persistence, the new model achieves skills ranging from 20% to 30% (22%–35%) for GHI (DNI), among the highest ever reported for these time horizons. An analysis based on Mutual Information and Pearson corre...

A k-nearest neighbor (kNN) ensemble model has been developed to generate Probability Density Function (PDF) forecasts for intra-hour Direct Normal Irradiance (DNI). This probabilistic forecasting model, which uses diffuse irradiance measurements and cloud cover information as exogenous feature inputs, adaptively provides arbitrary PDF forecasts for different weather conditions. The proposed models have been quantitatively evaluated using data from different locations characterized by different climates (continental, coastal, and island). The performance of the forecasts is quantified using metrics such as Prediction Interval Coverage Probability (PICP), Prediction Interval Normalized Averaged Width (PINAW), Brier Skill Score (BSS), and the Continuous Ranked Probability Score (CRPS), and other standard error metrics. A persistence ensemble probabilistic forecasting model and a Gaussian probabilistic forecasting model are employed to benchmark the performance of the proposed kNN ensemble model. The results show that the proposed model significantly outperform both reference models in terms of all evaluation metrics for all locations when the forecast horizon is greater than 5-min. In addition, the proposed model shows superior performance in predicting DNI ramps.

We discuss methods for net load forecasting and their significance for operation and management of power grids with high renewable energy penetration. Net load forecasting is an enabling technology for the integration of microgrid fleets with the macrogrid. Net load represents the load that is traded between the grids (microgrid and utility grid). It is important for resource allocation and electricity market participation at the point of common coupling between the interconnected grids. We compare two inherently different approaches: additive and integrated net load forecast models. The proposed methodologies are validated on a microgrid with 33% annual renewable energy (solar) penetration. A heuristics based solar forecasting technique is proposed, achieving skill of 24.20%. The integrated solar and load forecasting model outperforms the additive model by 10.69% and the uncertainty range for the additive model is larger than the integrated model by 2.2%. Thus, for grid applications an integrated forecast model is recommended. We find that the net load forecast errors and the solar forecasting errors are cointegrated with a common stochastic drift. This is useful for future planning and modeling because the solar energy time-series allows to infer important features of the net load time-series, such as expected variability and uncertainty.

A Sun-tracking imaging system is implemented for minimizing circumsolar image distortion for improved short-term solar irradiance forecasts. This sky-imaging system consists of a fisheye digital camera mounted on an automatic solar tracker that follows the diurnal pattern of the Sun. The Sun is located at the geometric center of the sky images where the fisheye distortion is minimized. Images from this new system provide more information about the circumsolar sky cover, which provides critical information for intra-hour solar forecasts, particularly for direct normal irradiance. An automatic masking algorithm has been developed to separate the sky area from ground obstacles and the image edges for each image that is collected. Then numerical image features are extracted from the segmented sky area and are used as exogenous inputs to MultiLayer Perceptron (MLP) models for direct normal irradiance forecasts. Sixty-seven days of irradiance and image measurements are used to train, optimize, and assess the MLP-based forecast models for solar irradiance. The results show that the MLP forecasts based on the newly proposed sky-imaging system significantly outperform the reference models in terms of statistical metrics and forecast skill, particularly for shorter horizons, achieving forecast skills 18%–50% higher than the skills of a reference MLP-based model that is based on a zenith-pointed, stationary sky-imaging system.

Clouds significantly attenuate ground-level solar irradiance causing substantial reduction in photovoltaic power output capacity. However, partly cloudy skies may lead to temporary enhancement of local Global Horizontal Irradiance (GHI) above the clear-sky ceiling and, at times, the extraterrestrial irradiance. Such enhancements are referred to here as Cloud Enhancement Events (CEEs). In this work we study these CEEs and assess quantitatively the occurrence of resulting coherent Ramp Rates (RRs). We analyze a full year of ground irradiance data recorded at the University of California, Merced, as well as nearly five months of irradiance data recorded at the University of California, San Diego, and Ewa Beach, Hawaii. Our analysis shows that approximately 4% of all the data points qualify as potential CEEs, which corresponds to nearly 3.5 full-days of such events per year if considered sequentially. The surplus irradiance enhancements range from 18 W m −2 day −1 to 73 W m −2 day −1 . The maximum recorded GHI of ∼1400 W m −2 occurred in San Diego on May 25, 2012, which was nearly 43% higher than the modeled clear-sky ceiling. Wavelet decomposition coupled with fluctuation power index analysis shed light on the time-scales on which cloud induced variability and CEEs operate. Results suggest that while cloud-fields tend to induce variability most strongly at the 30 min time-scale, they have the potential to cause CEEs that induce variability on time-scales of several minutes. This analysis clearly demonstrates that CEEs are an indicator for periods of high variability and therefore provide useful information for solar forecasting and integration.

A forecasting method for hourly-averaged, day-ahead power output (PO) from photovoltaic (PV) power plants based on least-squares optimization of Numerical Weather Prediction (NWP) is presented. Three variations of the forecasting method are evaluated against PO data from two non-tracking, 1 MWp PV plants in California for 2011–2014. The method performance, including the inter-annual performance variability and the spatial smoothing of pairing the two plants, is evaluated in terms of standard error metrics, as well as in terms of the occurrence of severe forecasting error events. Results validate the performance of the proposed methodology as compared with previous studies. We also show that the bias errors in the irradiance inputs only have a limited impact on the PO forecast performance, since the method corrects for systematic errors in the irradiance forecast. The relative root mean square error (RMSE) for PO is in the range of 10.3%–14.0% of the nameplate capacity, and the forecast skill ranges from 13% to 23% over a persistence model. Over three years, an over-prediction of the daily PO exceeding 40% only occurs twice at one of the two plants under study, while the spatially averaged PO of the paired plants never exceeds this threshold.

The output of ground-based, solar power generation systems is strongly dependent on cloud cover, which is the main contributor to solar power variability and uncertainty. Cloud optical properties are typically over-simplified in forecasting applications due to the lack of real-time, accurate estimates. In this work, we introduce a method, the Spectral Cloud Optical Property Estimation (SCOPE), for estimating cloud optical properties directly from high-resolution (5-min, 2 km) imagery from Geostationary Operational Environmental Satellite (GOES)-R, which is the newest generation of the GOES system. The SCOPE method couples a two-stream, spectrally resolved radiative model with the longwave GOES-R sensor output to simultaneously estimate the cloud optical depth, cloud top height, and cloud thickness during both day and night at 5-min intervals. The accuracy of SCOPE is evaluated using one year (2018) of downwelling longwave (DLW) radiation measurements from the Surface Radiation Budget Network, which consists of seven sites spread across climatically diverse regions of the contiguous United States. During daytime clear-sky periods, SCOPE predicts DLW within instrument uncertainty (10 W m−2) for four of the seven locations, with the remaining locations yielding errors of the order of 11.2, 17.7, and 20.2 W m−2. For daytime cloudy-sky, daytime all-sky (clear or cloudy), and nighttime all-sky periods, SCOPE achieves root mean square error values of 23.0–34.5 W m−2 for all seven locations. These results, together with the low-latency of the method (∼1 s per sample), show that SCOPE provides a viable solution to real-time, accurate estimation of cloud optical properties for both day and night.

This work compares the performance of machine learning methods (k-nearest-neighbors (kNN) and gradient boosting (GB)) in intra-hour forecasting of global (GHI) and direct normal (DNI) irradiances. The models predict the GHI and DNI and the corresponding prediction intervals. The data used in this work include pyranometer measurements of GHI and DNI and sky images. Point forecasts are evaluated using bulk error metrics while the performance of the probabilistic forecasts are quantified using metrics such as Prediction Interval Coverage Probability (PICP), Prediction Interval Normalized Averaged Width (PINAW) and the Continuous Ranked Probability Score (CRPS). Graphical verification displays like reliability diagram and rank histogram are used to assess the probabilistic forecasts. Results show that the machine learning models achieve significant forecast improvements over the reference model. The reduction in the RMSE translates into forecasting skills ranging between 8% and 24%, and 10% and 30% for the GHI and DNI testing set, respectively. CRPS skill scores of 42% and 62% are obtained respectively for GHI and DNI probabilistic forecasts. Regarding the point forecasts, the GB method performs better than the kNN method when sky image features are included in the model. Conversely, for probabilistic forecasts the kNN exhibits rather good performance.

Short term electricity trading to balance generation and demand provides an economic opportunity to integrate larger shares of variable renewable energy sources in the power grid. Recently, many regulatory market environments are reorganized to allow short term electricity trading. This study seeks to quantify the benefits of solar forecasting for energy imbalance markets (EIM). State-of-the-art solar forecasts, covering forecast horizons ranging from 24 h to 5 min are proposed and compared against the currently used benchmark models, persistence (P) and smart persistence (SP). The implemented reforecast of numerical weather prediction time series achieves a skill of 14.5% over the smart persistence model. Using the proposed forecasts for a forecast horizon of up to 75 min for a single 1 MW power plant reduces required flexibility reserves by 21% and 16.14%, depending on the allowed trading intervals (5 and 15 min). The probability of an imbalance, caused through wrong market bids from PV solar plants, can be reduced by 19.65% and 15.12% (for 5 and 15 min trading intervals). All EIM stakeholders benefit from accurate forecasting. Previous estimates on the benefits of EIMs, based on persistence model are conservative. It is shown that the design variables regulating the market time lines, the bidding and the binding schedules, drive the benefits of forecasting.

Ground based sky imaging and irradiance sensors are used to quantitatively evaluate the impact of cloud transmittance and cloud velocity on the accuracy of short-term direct normal irradiance (DNI) forecasts. Eight representative partly-cloudy days are used as an evaluation dataset. Results show that incorporating real-time sky and cloud transmittances as inputs reduces the root mean square error (RMSE) of forecasts of both the Deterministic model (Det) (16.3%∼ 17.8% reduction) and the multi-layer perceptron network model (MLP) (0.8% ∼ 6.2% reduction). Four computer vision methods: the particle image velocimetry method, the optical flow method, the x-correlation method and the scale-invariant feature transform method have accuracies of 83.9%, 83.5%, 79.2% and 60.9% in deriving cloud velocity, with respect to manual detection. Analysis also shows that the cloud velocity has significant impact on the accuracy of DNI forecasts: underestimating the cloud velocity magnitude by 50% results in 30.2% (Det) and 24.2% (MLP) increase of forecast RMSE; a 50% overestimate results in 7.0% (Det) and 8.4% (MLP) increase of RMSE; a ±30∘ deviation of cloud velocity direction increases the forecast RMSE by 6.2% (Det) and 6.6% (MLP).

In this work, we validate and enhance previously proposed singe-input direct normal irradiance (DNI) models based on numerical weather prediction (NWP) for intra-week forecasts with over 200,000 hours of ground measurements for 8 locations. Short latency re-forecasting methods to enhance the deterministic forecast accuracies are presented and discussed. The basic forecast is applied to 15 additional locations in North America with satellite-derived DNI data. The basic model outperforms the persistence model at all 23 locations with a skill between 12.4% and 38.2%. The RMSE of the basic forecast is in the range of 204.9 W m−2 to 309.9 W m−2. The implementation of stochastic learning re-forecasting methods yields further reduction in error from 204.9 W m−2 to 176.5 W m−2. To a great extent, the errors are caused by inaccuracies in the NWP cloud prediction. Improved assessment of atmospheric turbidity has limited impact on reducing forecast errors. Our results suggest that NWP-based DNI forecasts are very capable of reducing power and net-load uncertainty introduced by concentrated solar power plants at all locations in North America. Operating reserves to balance uncertainty in day-ahead schedules can be reduced on average by an estimated 28.6% through the application of the basic forecast.

The purpose of this work is to present a simple global solar irradiance forecasting framework based on the optimization of the k -nearest-neighbors (kNN) and artificial neural networks algorithms (ANN) for time horizons ranging from 15 min to 2 h. We apply the proposed forecasting models to irradiance from five locations and assessed the impact of different micro-climates on forecasting performance. We also propose two metrics, the density of large irradiance ramps and the time series determinism, to characterize the irradiance forecastability. Both measures are computed from the irradiance time series and provide a good indication for the forecasting performance before any predictions are produced. Results show that the proposed kNN and ANN models achieve substantial improvements relative to simpler forecasting models. The results also show that the optimal parameters for the kNN and ANN models are highly dependent on the different micro-climates. Finally, we show that the density of large irradiance ramps and time series determinism can successfully explain the forecasting performance for the different locations and time horizons.

We develop a standalone, real-time solar forecasting computational platform to predict one minute averaged solar irradiance ramps ten minutes in advance. This platform integrates cloud tracking techniques using a low-cost fisheye network camera and artificial neural network (ANN) algorithms, where the former is used to introduce exogenous inputs and the latter is used to predict solar irradiance ramps. We train and validate the forecasting methodology with measured irradiance and sky imaging data collected for a six-month period, and apply it operationally to forecast both global horizontal irradiance and direct normal irradiance at two separate locations characterized by different micro-climates (coastal and continental) in California. The performance of the operational forecasts is assessed in terms of common statistical metrics, and also in terms of three proposed ramp metrics, used to assess the quality of ramp predictions. Results show that the forecasting platform proposed in this work outperforms the reference persistence model for both locations.

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).

Compared with recirculation and injection modes, once-through direct steam generation (DSG) parabolic troughs are simpler to construct and require the lowest investment. However, the heat transfer fluid (HTF) in once-through DSG parabolic trough systems has the most complicated dynamic behavior, particularly during periods of moving shadows caused by small clouds and jet contrails. In this paper, a nonlinear distributed parameter dynamic model (NDPDM) is proposed to model the dynamic behavior of once-through DSG parabolic trough solar collector row under moving shadow conditions. Compared with state-of-the-art models, the proposed NDPDM possesses three characteristics: (a) adopting real-time local values of the heat transfer and friction resistance coefficients, (b) simulating the whole collector row, including the boiler and the superheated sections, and (c) modeling the disturbance of direct normal irradiance (DNI) level on DSG parabolic trough solar collector row under moving shadow conditions. Validated using experimental data, the NDPDM accurately predicts the dynamic characteristics of HTF during periods of partial and moving DNI disturbance. The fundamental and specific dynamic process of fluid parameters for a DSG parabolic trough solar collector row is provided in this paper. The results show the following: (a) Moving shadows have a significant impact on the outlet temperature and mass flow rate, and the impact lasts up to 1000 s even after the shadows completely leave the collector row. (b) The time for outlet steam temperature to reach a steady-state value for the first time is independent of the shadow width, speed, and moving direction. (c) High-frequency chattering of the outlet mass flow rate can be observed under moving DNI disturbance and will have a longer duration if the shadow width is larger or the shadow speed is slower. Compared with cases in which the whole system is shaded, partially shading cases have shown a longer duration of high-frequency chattering. (d) Both wider widths and slower speeds of shadow will cause a larger amplitude of responses in the outlet temperature and mass flow rate. When the shadow speed is low, there is a longer delay time of response in the mass flow rate of the outlet fluid. (e) The amplitude of response in the outlet temperature does not depend on the direction of clouds movement. However, if the DNI disturbance starts at the inlet of the collector row, there will be significant delay times in both outlet temperature and mass flow rate, and a larger amplitude of response in outlet mass flow rate.

Passive cooling devices take advantage of the partially transparent properties of the atmosphere in the longwave spectral band from 8 to 13 μm (the so-called “atmospheric window”) to reject radiation to outer space. Spectrally designed thermophotonic devices have raised substantial attention recently for their potential to provide passive and carbon-free alternatives to air conditioning. However, the level of transparency of the atmospheric window depends on the local content of water vapor in the atmosphere and on the optical depth of clouds in the local sky. Thus, the radiative cooling capacity of solar reflectors not only depends on the optical properties of their surfaces but also on local meteorological conditions. In this work, detailed radiative cooling resource maps for the contiguous United States are presented with the goal of determining the best climates for large-scale deployment of passive radiative cooling technologies. The passive cooling potential is estimated based on ideal optical properties, i.e., zero shortwave absorptance (maximum reflectance) and blackbody longwave emittance. Both annual and season-averaged maps are presented. Daytime and nighttime cooling potential are also computed and compared. The annual average cooling potential over the contiguous United States is 50.5 m−2. The southwestern United States has the highest annual averaged cooling potential, over 70 W m−2, due to its dry and mostly clear sky meteorological conditions. The southeastern United States has the lowest potential, around 30 W m−2, due to frequent humid and/or overcast weather conditions. In the spring and fall months, the Arizona and New Mexico climates provide the highest passive cooling potential, while in the summer months, Nevada and Utah exhibit higher potentials. Passive radiative cooling is primarily effective in the western United States, while it is mostly ineffective in humid and overcast climates elsewhere.Passive cooling devices take advantage of the partially transparent properties of the atmosphere in the longwave spectral band from 8 to 13 μm (the so-called “atmospheric window”) to reject radiation to outer space. Spectrally designed thermophotonic devices have raised substantial attention recently for their potential to provide passive and carbon-free alternatives to air conditioning. However, the level of transparency of the atmospheric window depends on the local content of water vapor in the atmosphere and on the optical depth of clouds in the local sky. Thus, the radiative cooling capacity of solar reflectors not only depends on the optical properties of their surfaces but also on local meteorological conditions. In this work, detailed radiative cooling resource maps for the contiguous United States are presented with the goal of determining the best climates for large-scale deployment of passive rad...

We describe and release a comprehensive solar irradiance, imaging, and forecasting dataset. Our goal with this release is to provide standardized solar and meteorological datasets to the research community for the accelerated development and benchmarking of forecasting methods. The data consist of three years (2014–2016) of quality-controlled, 1-min resolution global horizontal irradiance and direct normal irradiance ground measurements in California. In addition, we provide overlapping data from commonly used exogenous variables, including sky images, satellite imagery, and Numerical Weather Prediction forecasts. We also include sample codes of baseline models for benchmarking of more elaborated models.We describe and release a comprehensive solar irradiance, imaging, and forecasting dataset. Our goal with this release is to provide standardized solar and meteorological datasets to the research community for the accelerated development and benchmarking of forecasting methods. The data consist of three years (2014–2016) of quality-controlled, 1-min resolution global horizontal irradiance and direct normal irradiance ground measurements in California. In addition, we provide overlapping data from commonly used exogenous variables, including sky images, satellite imagery, and Numerical Weather Prediction forecasts. We also include sample codes of baseline models for benchmarking of more elaborated models.

The higher penetration of renewable resources in the energy portfolios accentuates the need for accurate forecasting of variable resources (solar, wind, tidal) at several different temporal scales to achieve power grid balance. Solar generation technologies have experienced strong energy market growth in the past few years, with corresponding increase in local grid penetration rates. As is the case with wind, the solar resource at the ground level is highly variable mostly due to cloud cover variability, atmospheric aerosol levels, and indirectly, and to a lesser extent, participating gases in the atmosphere. The inherent variability of solar generation at higher grid penetration levels poses problems associated with the cost of reserves, dispatchable and ancillary generation, and grid reliability in general. As a result, high accuracy forecast systems are required for multiple time horizons that are associated with regulation, dispatching, scheduling, and unit commitment. Here we review the theory behind mathematical methods for optimized solar forecasting and a number of successful applications of solar forecasting methods for both the solar resource and the power output of solar plants at the utility scale level.