Global Horizontal Irradiance Forecasts Research Articles

Effects of clouds and aerosols on downwelling surface solar irradiance nowcasting and short-term forecasting

Abstract. Solar irradiance nowcasting and short-term forecasting are important tools for the integration of solar plants into the electricity grid. Understanding the role of clouds and aerosols in those techniques is essential for improving their accuracy. In this study, we introduce improvements in the existing nowcasting and short-term forecasting operational systems SENSE (Solar Energy Nowcasting System) and NextSENSE achieved by using a new configuration and by upgrading cloud and aerosol inputs, and we also investigate the limitations of evaluating such models using surface-based sensors due to cloud effects. We assess the real-time estimates of surface global horizontal irradiance (GHI) produced by the improved SENSE2 operational system at high spatial and temporal resolution (∼ 5 km, 15 min) for a domain including Europe and the Middle East–North Africa (MENA) region and the short-term forecasts of GHI (up to 3 h ahead) produced by the NextSENSE2 system against ground-based measurements from 10 stations across the models' domain for a whole year (2017). Results for instantaneous (every 15 min) comparisons show that the GHI estimates are within ±50 W m−2 (or ±10 %) of the measured GHI for 61 % of the cases after the implementation of the new model configuration and a proposed bias correction. The bias ranges from −12 to 23 W m−2 (or from −2 % to 6.1 %) with a mean value of 11.3 W m−2 (2.3 %). The correlation coefficient is between 0.83 and 0.96 and has a mean value of 0.93. Statistics are significantly improved when integrating on daily and monthly scales (the mean bias is 3.3 and 2.7 W m−2, respectively). We demonstrate that the main overestimation of the SENSE2 GHI is linked with the uncertainties of the cloud-related information within the satellite pixel, while relatively low underestimation, linked with aerosol optical depth (AOD) forecasts (derived from the Copernicus Atmospheric Monitoring Service – CAMS), is reported for cloudless-sky GHI. The highest deviations for instantaneous comparisons are associated with cloudy atmospheric conditions, when clouds obscure the sun over the ground-based station. Thus, they are much more closely linked with satellite vs. ground-based comparison limitations than the actual model performance. The NextSENSE2 GHI forecasts based on the cloud motion vector (CMV) model outperform the persistence forecasting method, which assumes the same cloud conditions for future time steps. The forecasting skill (FS) of the CMV-based model compared to the persistence approach increases with cloudiness (FS is up to ∼ 20 %), which is linked mostly to periods with changes in cloudiness (which persistence, by definition, fails to predict). Our results could be useful for further studies on satellite-based solar model evaluations and, in general, for the operational implementation of solar energy nowcasting and short-term forecasting, supporting solar energy production and management.

NYSolarCast: A solar power forecasting system for New York State

Solar energy generation capacity will need to be greatly increased to meet aggressive clean energy targets by New York State (NYS), which are 70% renewable energy (RE) generation by 2030, and 100% by 2040. Because solar energy is a variable, weather-driven resource, accurate forecasts of solar energy generation both in nowcast (intra-day) and day-ahead time horizons are necessary for electric utilities and independent system operators, as they maintain grid stability and maximize RE use.Meeting this need for NYS, a gridded, open-source solar power forecasting system called NYSolarCast was developed (https://github.com/ncar/NYSolarCast_delivery). NYSolarCast makes 15-min resolution predictions of global horizontal irradiance (GHI) on a 3-km grid covering all of NYS, which are then used to estimate solar power generation both for select utility-scale PV plants (15-min resolution) and for zone-aggregated distributed PV (1-h resolution). The statewide GHI forecasts are made by applying the StatCast statistical forecasting model, which is trained on over two years of GHI observations from pyranometers at all 126 NYS Mesonet stations and gridded numerical weather prediction forecasts from both the Weather Research and Forecasting model tuned for solar applications (WRF-Solar®) and NOAA’s operational High-Resolution Rapid Refresh (HRRR) model. Forecasts are made at each NYS Mesonet site and then blended outward into the rest of the grid. This paper gives an overview of NYSolarCast performance for intra-day (0–6-h) GHI and power forecasts during a one-year period.

Evaluation of the WRF-solar model for 72-hour ahead forecasts of global horizontal irradiance in West Africa: A case study for Ghana

Accurate global horizontal irradiance (GHI) forecasting is critical for integrating solar energy into the power grid and operating solar power plants. The Weather Research and Forecasting model with its solar radiation extension (WRF-Solar) has been used to forecast solar irradiance in different regions around the world. However, the application of the WRF-Solar model to the prediction of GHI in West Africa, particularly Ghana, has not yet been investigated. The aim of this study is to evaluate the performance of the WRF-Solar model for predicting GHI in Ghana, focusing on three automatic weather stations (Akwatia, Kumasi and Kologo) for the year 2021. We used two one-way nested domains (D1 = 15 km and D2 = 3 km) to investigate the ability of the fully coupled WRF-Solar model to forecast GHI up to 72-hour ahead under different atmospheric conditions. The initial and lateral boundary conditions were taken from the ECMWF high-resolution operational forecasts. Our findings reveal that the WRF-Solar model performs better under clear skies than cloudy skies. Under clear skies, Kologo performed best in predicting 72-hour GHI, with a first day nRMSE of 9.62 %. However, forecasting GHI under cloudy skies at all three sites had significant uncertainties. Additionally, WRF-Solar model is able to reproduce the observed GHI diurnal cycle under high AOD conditions in most of the selected days. This study enhances the understanding of the WRF-Solar model’s capabilities and limitations for GHI forecasting in West Africa, particularly in Ghana. The findings provide valuable information for stakeholders involved in solar energy generation and grid integration towards optimized management in the region.

Solar radiation forecasting using gradient boosting based ensemble learning model for various climatic zones

India is seeing a massive boost in solar power installations in the recent years. It is expected that solar energy utilization will be increasing exponentially in near future. Solar radiation forecasting is an essential tool to obtain the optimal performance out of the solar systems. An ensemble model using gradient boosting has been developed for hourly global horizontal irradiance forecasting is proposed for the various climatic zones of India. The gradient boost-based model is benchmarked with Auto Regressive Integrated Moving Average (ARIMA),2-layer feed forward neural network, and Long Short-Term Memory (LSTM). Gradient boosting obtained with marginally better performance metrics, mean absolute error (20.97 W/m2) and mean square error (1357 W2/m4) as compared to that of the neural network (21.66 W/m2, 1479.53 W2/m4) and performed much better than ARIMA and LSTM in terms of mean absolute error (251.66 W/m2) & mean square error (82,290 W2/m4) and mean absolute error (67.60 W/m2) & mean square error (8123 W2/m4). The diverse climatic zones in India and feature importance analysis have been considered during the model development and testing phase. The proposed model is also intended to have practical implications in obtaining solar irradiance data in locations with minimal availability of solar radiation data and solar power systems due to the model ability to provide good performance metrics in a smaller timeframe making it an ideal candidate for real-time solar power predicting model.

A two‐step machine‐learning approach to statistical post‐processing of weather forecasts for power generation

By the end of 2021, the renewable energy share of the global electricity capacity reached 38.3% and the new installations were dominated by wind and solar energy, showing global increases of 12.7% and 18.5% respectively. However, both wind and photovoltaic energy sources are highly volatile, making planning difficult for grid operators; thuss, accurate forecasts of the corresponding weather variables are essential for reliable electricity predictions. The most advanced approach in weather prediction is the ensemble method, which opens the door for probabilistic forecasting. However, ensemble forecasts are often underdispersive and subject to systematic bias. Hence, they require some form of statistical post‐processing, where parametric models provide full predictive distributions of the weather variables at hand. We propose a general two‐step machine‐learning‐based approach to calibrating ensemble weather forecasts, where, in the first step, improved point forecasts are generated, which then together with various ensemble statistics serve as input features of the neural network estimating the parameters of the predictive distribution. In two case studies based on 100 m wind speed and global horizontal irradiance forecasts of the operational ensemble prediction system of the Hungarian Meteorological Service, the predictive performance of this novel method is compared with the forecast skill of the raw ensemble and the state‐of‐the‐art parametric approaches. Both case studies confirm that, at least up to 48 hr, statistical post‐processing substantially improves the predictive performance of the raw ensemble for all forecast horizons considered. The variants of the proposed two‐step method investigated outperform in skill their competitors, and the suggested new approach is well applicable for different weather quantities and for a fair range of predictive distributions.

Ultra-short-term forecasting of global horizontal irradiance (GHI) integrating all-sky images and historical sequences

Accurate minute solar forecasts play an increasingly crucial role in achieving optimal intra-day power grid dispatch. However, continuous changes in cloud distribution and coverage pose a challenge to solar forecasting. This study presents a convolutional neural network-long short-term memory (CNN-LSTM) model to predict the future 10-min global horizontal irradiance (GHI) integrating all-sky image (ASI) and GHI sequences as input. The CNN is used to extract the sky features from ASI and a fully connected layer is used to extract historical GHI information. The resulting temporary information outputs are then merged and forwarded to the LSTM for forecasting the GHI values for the next 10 min. Compared to CNN solar radiation forecasting models, incorporating GHI into the forecasting process leads to an improvement of 18% in the accuracy of forecasting GHI values for the next 10 min. This improvement can be attributed to the inclusion of historical GHI sequences and regression via LSTM. The historical GHI contains valuable meteorological information such as aerosol optical thickness. In addition, the sensitivity analysis shows that the 1-lagged input length of the GHI and ASI sequence yields the most accurate forecasts. The advantages of CNN-LSTM facilitate power system stability and economic operation. Codes of the CNN-LSTM model in the public domain are available online on the GitHub repository https://github.com/zoey0919/CNN-LSTM-for-GHI-forecasting.

Explainable forecasting of global horizontal irradiance over multiple time steps using temporal fusion transformer

Accurate prediction of solar irradiance is essential for the successful integration of solar power plants into electrical systems. Despite recent advancements in deep learning technology yielding impressive results in solar forecasting, their lack of interpretability has hindered their widespread adoption. In this paper, we propose a novel approach that integrates a Temporal Fusion Transformer (TFT) with a McClear model to achieve accurate and interpretable forecasting performance. The TFT is a deep learning model that provides transparency in its predictions through the use of interpretable self-attention layers for long-term dependencies, recurrent layers for local processing, specialized components for feature selection, and gating layers to suppress extraneous components. The model is capable of learning temporal associations between continuous time-series variables, namely, historical global horizontal irradiance (GHI) and clear sky GHI, accounting for cloud cover variability and clear sky conditions that are often ignored by most machine learning solar forecasters. Additionally, it minimizes a quantile loss during training to produce accurate probabilistic forecasts. In this study, we evaluate the performance of hourly GHI forecasts on eight diverse datasets with varying climates: temperate, cold, arid, and equatorial, for multiple temporal horizons of 2, 3, 6, 12, and 24 h. The model is benchmarked against both climatological persistence for deterministic forecasting and Complete History Persistence Ensemble for probabilistic forecasting. To prove that our model is not location locked, it has been blind tested on four completely different datasets. The results demonstrate that the proposed model outperforms its counterparts across all forecast horizons.

A Transformer-based multimodal-learning framework using sky images for ultra-short-term solar irradiance forecasting

The development of solar energy is crucial to combat the global climate change and fossil energy crisis. However, the inherent uncertainty of solar power prevents its large-scale integration into power grids. Although various sky-image-derived modeling methods exist to forecast the variations of solar irradiance, few focus on fully utilizing the coupling correlations between sky images and historical data to improve the forecasting performance. Therefore, a novel multimodal-learning framework is proposed for forecasting global horizontal irradiance (GHI) in the ultra-short-term. First, the historical and empirically estimated clear-sky GHI are encoded by Informer. Then, the ground-based sky images are transformed into optical flow maps, which can be handled by Vision Transformer. Subsequently, a cross-modality attention method is proposed to explore the coupling correlations between the two modalities. Last, a generative decoder is used to implement multi-step forecasting. The experimental results show that the proposed method achieves a normalized root mean square error (NRMSE) of 4.28% in 10-min-ahead forecasting. Several state-of-the-art methods are also used for comparisons. The experimental results show that the proposed method outperforms the benchmark methods and exhibits higher accuracy and robustness in ultra-short-term GHI forecasting.

Development of a day-ahead solar power forecasting model chain for a 250 MW PV park in India

Due to the steep rise in grid-connected solar Photovoltaic (PV) capacity and the intermittent nature of solar generation, accurate forecasts are becoming ever more essential for the secure and economic day-ahead scheduling of PV systems. The inherent uncertainty in Numerical Weather Prediction (NWP) forecasts and the limited availability of measured datasets for PV system modeling impacts the achievable day-ahead solar PV power forecast accuracy in regions like India. In this study, an operational day-ahead PV power forecast model chain is developed for a 250 MWp solar PV park located in Southern India using NWP-predicted Global Horizontal Irradiance (GHI) from the European Centre of Medium Range Weather Forecasts (ECMWF) and National Centre for Medium Range Weather Forecasting (NCMRWF) models. The performance of the Lorenz polynomial and a Neural Network (NN)-based bias correction method are benchmarked on a sliding window basis against ground-measured GHI for ten months. The usefulness of GHI transposition, even with uncertain monthly tilt values, is analyzed by comparing the Global Tilted Irradiance (GTI) and GHI forecasts with measured GTI for four months. A simple technique for back-calculating the virtual DC power is developed using the available aggregated AC power measurements and the inverter efficiency curve from a nearby plant with a similar rated inverter capacity. The AC power forecasts are validated against aggregated AC power measurements for six months. The ECMWF derived forecast outperforms the reference convex combination of climatology and persistence. The linear combination of ECMWF and NCMRWF derived AC forecasts showed the best result.

Combining a deep learning model with multivariate empirical mode decomposition for hourly global horizontal irradiance forecasting

Accurate and reliable global horizontal irradiance forecasting is one of the solutions for the associated problems with grid-integrated PV plants. This study proposes a novel hybrid MEMD-PCA-GRU model for an hour ahead of GHI forecasting. The multivariate empirical mode decomposition (MEMD) breaks the multidimensional data into multivariate subseries termed intrinsic mode functions (IMFs). MEMD helps to remove the naturally produced non-stationary and nonlinear deficiencies within the target series and meteorological predictors. A large number of obtained IMFs necessitates the application of a dimensionality reduction technique. Principal component analysis (PCA) is used here to identify the most informative features from a large set of IMFs. Finally, the gated recurrent unit (GRU) is utilized to predict GHI at four places in India. The performance of the proposed model is tested against some hybrid and standalone models. Double decomposition techniques enhanced the GRU performance by a minimum % RMSE (% MAE) improvement of 48.38 (24.97). The proposed model reported an average nRMSE (RMSE) of 7.82% (36.85 W/m2) across four locations. The lowest error metrics of the proposed model reflect the relatively stable and good performance compared to studied single-stage and hybrid benchmark models under different climatic conditions.

Intraday forecast of global horizontal irradiance using optical flow method and long short-term memory model

Short-term global horizontal irradiance (GHI) forecast methodologies are routinely employed to mitigate the instability of photovoltaic power and, more importantly, to secure early participation in the energy auction market. The intra-day forecast model substantially consists of the combination of a cloud motion vector (CMV) derivation and a GHI extraction model. This study utilized optical flow (OF) method to calculate CMV from two subsequent satellite images and predict cloud displacement up to 180 min ahead. Meanwhile, the conventional GHI extraction models (Hammer and the fast all-sky radiation model for solar applications (FARMS) models) were employed to extract future GHI. Moreover, to the best of our knowledge, only a few studies have been dedicated to combining OF method with deep learning model. Hence, this study proposed a novel combination based on OF method and long short-term memory model (LSTM) to enhance temporal horizon and accuracy. After all combinations of OF method and GHI models were simulated and assessed under all-sky conditions, the OF method that combined with LSTM outperformed the OF method combined with conventional models across all time horizons. For root mean square error (RMSE) and forecast skill (FS), the accuracy of OF and LSTM models for all time horizon ranges from 59.64 W/m2 to 120.63 W/m2 and 16.40 % to 54.42 %, respectively. The study findings are expected to encourage the inclusion of satellite data in GHI forecast as well as the optimization of energy management systems that integrate with a photovoltaic system.

Long-Term Solar Irradiance Forecasting Using Multilinear Predictors

As the demand for crude oil is increasing every day, prices and pollution are both increasing in return, which has harmful effects on the environment. Thus, more attempts are being made to develop clean energy to rescue the planet and provide humanity with a cleaner energy source. As the renewable energy sector grows, new issues and challenges have emerged. Instability in electricity production from wind turbines, solar power plants, and dams creates challenges for energy transmission and storage systems. In order to achieve a more reliable and effective energy system, machine learning techniques have been used to forecast energy changes. Predicting the sun's Global Horizontal Irradiance (GHI) is one of the machine learning applications used in that sector. In this paper, many machine learning methods have been utilized, such as linear regression and long-short term memory (LSTM) methods to have long term GHI forecasting. Moreover, the significance of this paper is located in the way of prediction of the GHI irradiance prediction by using different levels of the linear regressors to find the best regressor level that provides the minimum error for the testing set based on cross-validation. Results showed that the regressor method provides a lower error compared with a single vanilla LSTM system for a shorter time computationally.

Comparison between Time- and Observation-Based Gaussian Process Regression Models for Global Horizontal Irradiance Forecasting

With the development of predictive management strategies for power distribution grids, reliable information on the expected photovoltaic power generation, which can be derived from forecasts of global horizontal irradiance (GHI), is needed. In recent years, machine learning techniques for GHI forecasting have proved to be superior to classical approaches. This work addresses the topic of multi-horizon forecasting of GHI using Gaussian process regression (GPR) and proposes an in-depth study on some open questions: should time or past GHI observations be chosen as input? What are the appropriate kernels in each case? Should the model be multi-horizon or horizon-specific? A comparison between time-based GPR models and observation-based GPR models is first made, along with a discussion on the best kernel to be chosen; a comparison between horizon-specific GPR models and multi-horizon GPR models is then conducted. The forecasting results obtained are also compared to those of the scaled persistence model. Four performance criteria and five forecast horizons (10 min, 1 h, 3 h, 5 h, and 24 h) are considered to thoroughly assess the forecasting results. It is observed that, when seeking multi-horizon models, using a quasiperiodic kernel and time as input is favored, while the best horizon-specific model uses an automatic relevance determination rational quadratic kernel and past GHI observations as input. Ultimately, the choice depends on the complexity and computational constraints of the application at hand.

Deep Learning Approaches for Long-Term Global Horizontal Irradiance Forecasting for Microgrids Planning

Providing sustainable energy to rural communities is considered in Sustainable Development Goal 7. Off-grid renewable energy systems arise as an affordable solution due to their portability and the availability of renewable sources for rural communities. In this work, to deal with the uncertainties of solar resources, we employ two deep learning models (feed forward and recurrent neural networks) to predict renewable sources in a long-term horizon. To this aim, the approach presented takes into account the necessity of a high enough resolution in the forecasting output. As a case study, we employ open source data for a location in Michoacan, Mexico as well as open source programming frameworks to ensure the replicability of the numerical experiments. The results show that our prediction model performs excellently with respect to the baseline methods (ARIMA, exponential smoothing, and seasonal naive) in terms of the evaluation metrics MASE (18.5% of reduction with respect to seasonal naive), RMSE (24.7%), WAPE (13.1%), MAE (12.9%), and APB (8.9%).

A novel sky image-based solar irradiance nowcasting model with convolutional block attention mechanism

Global horizontal irradiance (GHI) is a crucial factor impacting photovoltaic (PV) production, and is required for accurate real-time photovoltaic power forecasting. And it is a new effective solution to obtain the real-time GHI by the sky images because the GHI is mainly affected by the cloud cover and cloud motion. Therefore, the research proposes a unique artificial intelligence approach for real-time GHI forecasting (‘nowcasting’) based on sky images, which can significantly enhance GHI forecasting accuracy on cloudy sky days. First, a new nowcasting model with convolutional block attention module (CBAM) is proposed, which is based on Visual Geometry Group (VGG) networks. Then, taking the local cloud cover (LCC) as a numerical feature, we coupled it with the cloud feature in the sky image to improve the forecasting performance of the GHI model. Finally, to verify the effectiveness and advantages of the proposed method, when compared to state-of-the-art methods, such as Sun’s model, Jiang’s model, and others, the proposed method outperforms them in GHI forecasting as demonstrated by the 11.67% nRMSE, 7.97% nMAE, 27.69% MAPE, and 0.91 CORR results on the ASI-16 dataset.

A Time Series Forecasting of Global Horizontal Irradiance on Geographical Data of Najran Saudi Arabia

Environment-friendly and renewable energy resources are the need of each developed and undeveloped country. Solar energy is one of them, thus accurate forecasting of it can be useful for electricity supply companies. This research focuses on analyzing the daily global solar radiation (GSR) data of Najran province located in Saudi Arabia and proposed a model for the prediction of global horizontal irradiance (GHI). The weather data is collected from Najran University. After inspecting the data, I we found the dependent and independent variables for calculating the GHI. A dataset model has been trained by creating tensor of variables belonging to air, wind, peak wind, relative humidity, and barometric pressure. Furthermore, six machine learning algorithms convolutional neural networks (CNN), K-nearest neighbors (KNN), support vector machines (SVM), logistic regression (LR), random forest classifier (RFC), and support vector classifier (SVC) techniques are used on dataset model to predict the GHI. The evaluation metrics determination coefficients (R2), root mean square error (RMSE), relative root mean square error (rRMSE), mean bias error (MBE), mean absolute bias error (MABE), mean absolute percentage error (MAPE), and T-statistic (t-stat) are used for the result verification of proposed models. Finally, the current work reports that all methods examined in this work may be utilized to accurately predict GHI; however, the SVC technique is the most suitable method amongst all techniques by claiming the precise results using the evaluation metrics.

Fuzzy Time Series Methods Applied to (In)Direct Short-Term Photovoltaic Power Forecasting

Solar photovoltaic energy has experienced significant growth in the last decade, as well as the challenges related to the intermittency of power generation inherent to this process. In this paper we propose to perform short-term forecasting of solar PV generation using fuzzy time series (FTS). Two FTS methods are proposed and evaluated to obtain a global horizontal irradiance (GHI) value. The first is the weighted method and the second is the fuzzy information granular method. Using the direct proportionality of the power with the GHI, the spatial smoothing process was applied, obtaining spatial irradiance on which a first-order low pass filter was applied to simulated power photovoltaic system generation. Thus, this study proposed indirect and direct forecasting of solar photovoltaic generation which was statistically evaluated and the results showed that the indirect prediction showed better performance with GHI than the power simulation. Error statistics, such as RMSE and MBE, show that the fuzzy information granular method performs better than the weighted method in GHI forecasting.

Intra-day global horizontal irradiance forecast using FY-4A clear sky index

Accurate forecasting of Global horizontal irradiance (GHI) plays an irreplaceable role in power system expansion, power generation production scheduling, maintenance scheduling and ensuring continuous power supply. A method for GHI forecasting based on FY-4A satellite images is proposed, called FY-4A-Heliosat. This method is derived from the clear sky model (MCclear) and the future clear sky index (CSI), which is obtained by comparing the continuous cloud albedo (CAL) maps derived from AGRI satellite images according to Heliosat-2 method. The forecast horizon is 30 min to 3 h with a 30-min horizon resolution. The combination of climatology and persistence model (CP) is chosen as reference forecasts. Under all sky conditions, the results show that compared with CP model, FY-4A-Heliosat model is better in most cases and seasons. The FY-4A-Heliosat method underestimates the GHI in four seasons for all time horizons. Annual studies indicate that the FY-4A-Heliosat method performs better than the CP method for all forecast horizons; the nRMSE is 17% to 25% for the annual performance. Forecasts with change in cloud conditions show that: the FY-4A-Heliosat method performs better than the CP method in clear, partially cloudy, overcast and clear-to-cloudy conditions. For the cloudy-to-clear set, the FY-4A-Heliosat method needs to be improved future. The results demonstrate that the proposed method is suitable for very short term GHI forecasting.

An ensemble method to forecast 24-h ahead solar irradiance using wavelet decomposition and BiLSTM deep learning network.

In recent years, the penetration of solar power at residential and utility levels has progressed exponentially. However, due to its stochastic nature, the prediction of solar global horizontal irradiance (GHI) with higher accuracy is a challenging task; but, vital for grid management: planning, scheduling & balancing. Therefore, this paper proposes an ensemble model using the extended scope of wavelet transform (WT) and bidirectional long short term memory (BiLSTM) deep learning network to forecast 24-h ahead solar GHI. The WT decomposes the input time series data into different finite intrinsic model functions (IMF) to extract the statistical features of input time series. Further, the study reduces the number of IMF series by combining the wavelet decomposed components (D1-D6) series on the basis of comprehensive experimental analysis with an aim to improve the forecasting accuracy. Next, the trained standalone BiLSTM networks are allocated to each IMF sub-series to execute the forecasting. Finally, the forecasted values of each sub-series from BiLSTM networks are reconstructed to deliver the final solar GHI forecast. The study performed monthly solar GHI forecasting for one year dataset using one month moving window mechanism for the location of Ahmedabad, Gujarat, India. For the performance comparison, the naïve predictor as a benchmark model, standalone long short term memory (LSTM), gated recurrent unit (GRU), BiLSTM and two other wavelet-based BiLSTM models are also simulated. From the results, it is observed that the proposed model outperforms other models in terms of root mean square error (RMSE) & mean absolute percentage error (MAPE), coefficient of determination (R2) and forecast skill (FS). The proposed model reduces the monthly average RMSE by range from 26.04–58.89%, 5.17–31.35%, 23.26–56.06% & 21.08–57% in comparison with benchmark, standalone BiLSTM, GRU & LSTM networks respectively. On the other hand, the monthly average MAPE is reduced by range from 9 to 51.18%, 12.59–28.14%, 30.43–59.19% & 26.54–58.92% in comparison to benchmark, standalone BiLSTM, GRU & LSTM respectively. Further, the proposed model obtained the value of R2 equal to 0.94 and forecast skill (%) of 47% with reference to the benchmark model.

Resilient Predictive Control Coupled with a Worst-Case Scenario Approach for a Distributed-Generation-Rich Power Distribution Grid

In a context of accelerating deployment of distributed generation in power distribution grid, this work proposes an answer to an important and urgent need for better management tools in order to ‘intelligently’ operate these grids and maintain quality of service. To this aim, a model-based predictive control (MPC) strategy is proposed, allowing efficient re-routing of power flows using flexible assets, while respecting operational constraints as well as the voltage constraints prescribed by ENEDIS, the French distribution grid operator. The flexible assets used in the case study—a low-voltage power distribution grid in southern France—are a biogas plant and a water tower. Non-parametric machine-learning-based models, i.e., Gaussian process regression (GPR) models, are developed for intraday forecasting of global horizontal irradiance (GHI), grid load, and water demand, to better anticipate emerging constraints. The forecasts’ quality decreases as the forecast horizon grows longer, but quickly stabilizes around a constant error value. Then, the impact of forecasting errors on the performance of the control strategy is evaluated, revealing a resilient behaviour where little degradation is observed in terms of performance and computation cost. To enhance the strategy’s resilience and minimise voltage overflow, a worst-case scenario approach is proposed for the next time step and its contribution is examined. This is the main contribution of the paper. The purpose of the min–max problem added upstream of the main optimisation problem is to both anticipate and minimise the voltage overshooting resulting from forecasting errors. In this min–max problem, the feasible space defined by the confidence intervals of the forecasts is searched, in order to determine the worst-case scenario in terms of constraint violation, over the next time step. Then, such information is incorporated into the decision-making process of the main optimisation problem. Results show that these incidents are indeed reduced thanks to the min–max problem, both in terms of frequency of their occurrence and the total surface area of overshooting.