Solar Power Forecasting Research Articles

Enhanced Solar Power Prediction Using Attention-Based DiPLS-BiLSTM Model

The data for solar power generation contain a huge amount of data with a large number of features which are difficult to extract effectively. It is important for the grid management and operational efficiency of the solar farm to accurately predict the solar power. The existing prediction models utilize historical data but often fail to capture critical latent features. This limitation leads to overlooked complex dependencies or temporal relationships, reducing prediction accuracy, especially in load and generation forecasting. An attention-based dynamic inner partial least squares (DiPLS) model and a bidirectional long short-term memory (BiLSTM) model were used for solar power prediction. First, DiPLS is used to dynamically extract features, and then, an attention process is used to predict the importance of these features. Finally, the raw sets are input to the BiLSTM model to make predictions of solar power in the future. The proposed method improves prediction accuracy, achieving an R2 value of 0.965 for training and 0.961 for testing, compared to conventional models. Additionally, the method demonstrated lower root mean squared error (RMSE), indicating enhanced stability and accuracy for solar power forecasting.

Establishing Lightweight and Robust Prediction Models for Solar Power Forecasting Using Numerical–Categorical Radial Basis Function Deep Neural Networks

As green energy technology develops, so too grows research interest in topics such as solar power forecasting. The output of solar power generation is uncontrollable, which makes accurate prediction of output an important task in the management of power grids. Despite a plethora of theoretical models, most frameworks encounter problems in practice because they assume that received data is error-free, which is unlikely, as this type of data is gathered by outdoor sensors. We thus designed a robust solar power forecasting model and methodology based on the concept of ensembling, with three key design elements. First, as models established using the ensembling concept typically have high computational costs, we pruned the deep learning model architecture to reduce the size of the model. Second, the mediation model often used for pruning is not suitable for solar power forecasting problems, so we designed a numerical–categorical radial basis function deep neural network (NC-RBF-DNN) to replace the mediation model. Third, existing pruning methods can only establish one model at a time, but the ensembling concept involves the establishment of multiple sub-models simultaneously. We therefore designed a factor combination search algorithm, which can identify the most suitable factor combinations for the sub-models of ensemble models using very few experiments, thereby ensuring that we can establish the target ensemble model with the smallest architecture and minimal error. Experiments using a dataset from real-world solar power plants verified that the proposed method could be used to build an ensemble model of the target within ten attempts. Furthermore, despite considerable error in the model inputs (two inputs contained 10% error), the predicted NRMSE of our model is still over 10 times better than the recent model.

Advancing global solar photovoltaic power forecasting with sub-seasonal climate outlooks

Given the high weather dependency of solar photovoltaic energy, accurate weather information ranging from days to weeks in advance are required for stable plant operation and management. This study emphasizes that sub-seasonal climate outlooks can advance global solar power forecasting. We evaluate the capacity factor (CF) for standard test conditions (CFstc), calculated from atmospheric reanalysis of surface solar irradiance and ambient air temperature, by comparing it with actual CF from two solar plants in Korea. Results confirm CFstc as a reliable estimate of actual CF. Application of this methodology to four-week outlooks from two state-of-the-art sub-seasonal climate forecast systems shows a significant anomaly correlation coefficient (ACC) skill for global solar power forecasting at least one week in advance (5–11 forecast days). Regional ACC skills vary at extended lead times, with South Asia, eastern South America and eastern Australia maintaining ACC > 0.6 for up to two weeks (12–18 forecast days). Notably, useful ACC skill persists for up to four weeks (26–32 forecast days) across 70.9 % of global land areas. These findings suggest that sub-seasonal climate outlooks can provide decision-making information in developing more reliable solar energy strategies, highlighting the importance of transdisciplinary collaboration between renewable energy and meteorological sectors.

Quantum long short-term memory (QLSTM) vs. classical LSTM in time series forecasting: a comparative study in solar power forecasting

Accurate solar power forecasting is pivotal for the global transition towards sustainable energy systems. This study conducts a meticulous comparison between Quantum Long Short-Term Memory (QLSTM) and classical Long Short-Term Memory (LSTM) models for solar power production forecasting. The primary objective is to evaluate the potential advantages of QLSTMs, leveraging their exponential representational capabilities, in capturing the intricate spatiotemporal patterns inherent in renewable energy data. Through controlled experiments on real-world photovoltaic datasets, our findings reveal promising improvements offered by QLSTMs, including accelerated training convergence and substantially reduced test loss within the initial epoch compared to classical LSTMs. These empirical results demonstrate QLSTM’s potential to swiftly assimilate complex time series relationships, enabled by quantum phenomena like superposition. However, realizing QLSTM’s full capabilities necessitates further research into model validation across diverse conditions, systematic hyperparameter optimization, hardware noise resilience, and applications to correlated renewable forecasting problems. With continued progress, quantum machine learning can offer a paradigm shift in renewable energy time series prediction, potentially ushering in an era of unprecedented accuracy and reliability in solar power forecasting worldwide. This pioneering work provides initial evidence substantiating quantum advantages over classical LSTM models while acknowledging present limitations. Through rigorous benchmarking grounded in real-world data, our study illustrates a promising trajectory for quantum learning in renewable forecasting.

State-of-the-Art Probabilistic Solar Power Forecasting: A Structured Review

In recent years, the installed capacity increment with regard to solar power generation has been highlighted as a crucial role played by Photovoltaic (PV) generation forecasting in integrating a growing number of distributed PV sites into power systems. Nevertheless, because of the PV generation’s unpredictable nature, deterministic point forecast methods struggle to accurately assess the uncertainties associated with PV generation. This paper presents a detailed structured review of the state-of-the-art concerning Probabilistic Solar Power Forecasting (PSPF), which covers forecasting methods, model comparison, forecasting horizon and quantification metrics. Our review methodology leverages the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) approach to systematically identify primary data sources, focusing on keywords such as probabilistic forecasting, Deep Learning (DL), and Machine learning (ML). Through an extensive and rigorous search of renowned databases such as SCOPUS and Web of Science (WoS), we identified 36 relevant studies (n=36). Consequently, expert scholars decided to develop three themes: (1) Conventional PSPF, (2) PSPF utilizing ML, and (3) PSPF using DL. Probabilistic forecasting is an invaluable tool concerning power systems, especially regarding the rising proportion of renewable energy sources in the energy mix. We tackle the inherent uncertainty of renewable generation, maintain grid stability, and promote efficient energy management and planning. In the end, this research contributes to the development of a power system that is more resilient, reliable, and sustainable.

Enhancing photovoltaic power prediction using a CNN-LSTM-attention hybrid model with Bayesian hyperparameter optimization

Improving the accuracy of solar power forecasting is crucial to ensure grid stability, optimize solar power plant operations, and enhance grid dispatch efficiency. Although hybrid neural network models can effectively address the complexities of environmental data and power prediction uncertainties, challenges such as labor-intensive parameter adjustments and complex optimization processes persist. Thus, this study proposed a novel approach for solar power prediction using a hybrid model (CNN-LSTM-attention) that combines a convolutional neural network (CNN), long short- term memory (LSTM), and attention mechanisms. The model incorporates Bayesian optimization to refine the parameters and enhance the prediction accuracy. To prepare high-quality training data, the solar power data were first preprocessed, including feature selection, data cleaning, imputation, and smoothing. The processed data were then used to train a hybrid model based on the CNN-LSTM-attention architecture, followed by hyperparameter optimization employing Bayesian methods. The experimental results indicated that within acceptable model training times, the CNN-LSTM-attention model outperformed the LSTM, GRU, CNN-LSTM, CNN-LSTM with autoencoders, and parallel CNN-LSTM attention models. Furthermore, following Bayesian optimization, the optimized model demonstrated significantly reduced prediction errors during periods of data volatility compared to the original model, as evidenced by MRE evaluations. This highlights the clear advantage of the optimized model in forecasting fluctuating data.

Advancing Solar Power Forecasting: Integrating Boosting Cascade Forest and Multi-Class-Grained Scanning for Enhanced Precision

Advancing Solar Power Forecasting: Integrating Boosting Cascade Forest and Multi-Class-Grained Scanning for Enhanced Precision

On soft measurement modeling for predicting photovoltaic power with uncertainty based on the Takagi–Sugeno model

Abstract Accurate solar power forecast is becoming more essential for safe and reliable power grid operation with the increasing number of grid-connected photovoltaic (PV) power production units. However, PV power exhibits significant output fluctuation due to both external inputs and intrinsic stochasticity in system dynamics. Therefore, an efficient and reliable soft measurement model of PV power with uncertainty is demanded in practice applications. The technique described in this paper captures the impacts produced by the fundamental uncertainty observed in the data, instead of relying on unrealistic assumptions about uncertainty. A soft measurement model using the Takagi-Sugeno (TS) Fuzzy Logic System (FLS) based on several input-output time series of PV plants is presented in this study. Chebyshev's inequality from probability theory and statistics is adopted to create the confidence interval-based response envelopes for these time series at each moment. An envelope-based measure of output uncertainty and a center-valued response forecasting model can be obtained by the proposed identification technique. PV datasets are employed to demonstrate the concept, which indicates the proposed soft measurement may outperform existing methods in terms of prediction accuracy. The average values of Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the Correlation Coefficient (R) are only 0.0787, 0.1113, and 0.9979. The average values of the prediction interval coverage probability (PICP) and the prediction interval normalized average width (PINAW) are 0.9806 and 0.1051.

Inclusion of Shading and Soiling With Physical and Data-Driven Algorithms for Solar Power Forecasting

Shading and soiling are the biggest environmental factors that negatively affect the yield of PV systems. In order to integrate PV systems into the grid as easily and on large scale as possible, it is important that energy generation forecasts are as accurate as possible. The scope of this paper is to present a method how shading and soiling can be integrated into machine learning based PV forecasts even if they have already been pre trained by a large dataset. This paper focuses on shading by buildings, trees, obstacles, while shading by clouds can only be considered to a limited extent by weather forecasts. This study uses a dataset of three years of training data to build a base model. Subsequently, the power loss due to shading and soiling is determined using a digital twin and used to correct the forecast values of the baseline model. Finally, an evaluation of the corrected and original predicted values is performed. This study is able to show that the forecast error could be reduced in the same way as the loss due to shading and soiling using various machine learning methods. The results were compared against a Physical Informed Neural Network (PINN), which outperformed classical machine learning methods both with and without shading and soiling.

Bayesian neural networks for solar power forecasts in advanced thermoelectric systems

This study presents a Bayesian neural network for the comprehensive performance forecast of an advanced thermoelectric system featuring segmented thermocouples with variable area shape for optimal solar energy conversion. The training data is obtained from a three-dimensional numerical model developed in ANSYS software to account for the energy, exergy, and entropy flows in the thermoelectric module to maximize the power output, energy, and exergy efficiencies while minimizing the entropy generation and system irreversibilities. The thermodynamic model optimizes key parameters, including geometry (height, cross-section, and percentage skutterudite content) across realistic diverse heat fluxes and cooling coefficients. To facilitate a faster and more efficient optimization pipeline, various neural networks, leveraging algorithms like Levenberg-Marquardt, scaled conjugate gradient, and Bayesian regularization, are utilized for efficient data management and accurate performance prediction. The Bayesian neural network emerges as the most accurate and efficient model providing the lowest training and testing loss of ∼10−9. Furthermore, key findings highlight the importance of skutterudite content and solar concentration ratio. For instance, an optimal skutterudite content of 6.2 % at 25 Suns produces maximum power of 29.7 W, while 37 % skutterudite at 75 Suns minimizes exergy destruction. This study's methodology and findings contribute to designing more efficient clean energy systems and advancing sustainable energy goals.

Machine learning-based short-term solar power forecasting: a comparison between regression and classification approaches using extensive Australian dataset

Solar energy production is an intermittent process that is affected by weather and climate conditions. This can lead to unstable and fluctuating electricity generation, which can cause financial losses and damage to the power grid. To better control power production, it is important to predict solar energy production. Big data and machine learning algorithms have yielded excellent results in this regard. This study compares the performance of two different machine learning approaches to solar energy production prediction: regression and classification. The regression approach predicts the actual power output, while the classification approach predicts whether the power output will be above or below a certain threshold. The study found that the random forest regressor algorithm performed the best in terms of accuracy, with mean absolute errors and root mean square errors of 0.046 and 0.11, respectively. However, it did not predict peak power values effectively, which can lead to higher errors. The long short-term memory algorithm performed better in classifying peak power values. The study concluded that classification models may be better at generalizing than regression models. This proposed approach is valuable for interpreting model performance and improving prediction accuracy.

Machine Learning Models for Solar Power Generation Forecasting in Microgrid Application Implications for Smart Cities

In the context of escalating concerns about environmental sustainability in smart cities, solar power and other renewable energy sources have emerged as pivotal players in the global effort to curtail greenhouse gas emissions and combat climate change. The precise prediction of solar power generation holds a critical role in the seamless integration and effective management of renewable energy systems within microgrids. This research delves into a comparative analysis of two machine learning models, specifically the Light Gradient Boosting Machine (LGBM) and K Nearest Neighbors (KNN), with the objective of forecasting solar power generation in microgrid applications. The study meticulously evaluates these models’ accuracy, reliability, training times, and memory usage, providing detailed experimental insights into optimizing solar energy utilization and driving environmental sustainability forward. The comparison between the LGBM and KNN models reveals significant performance differences. The LGBM model demonstrates superior accuracy with an R-squared of 0.84 compared to KNN’s 0.77, along with lower Root Mean Squared Error (RMSE: 5.77 vs. 6.93) and Mean Absolute Error (MAE: 3.93 vs. 4.34). However, the LGBM model requires longer training times (120 s vs. 90 s) and higher memory usage (500 MB vs. 300 MB). Despite these computational differences, the LGBM model exhibits stability across diverse time frames and seasons, showing robustness in handling outliers. These findings underscore its suitability for microgrid applications, offering enhanced energy management strategies crucial for advancing environmental sustainability. This research provides essential insights into sustainable practices and lays the foundation for a cleaner energy future, emphasizing the importance of accurate solar power forecasting in microgrid planning and operation.

Short-Term Solar Power Prediction using Machine Learning Algorithms : A High Performing approach

The use of solar energy is growing in popularity across the globe as a clean and sustainable energy source. Nevertheless, integrating solar power into the grid and guaranteeing a steady supply of electricity is made more difficult by the weather patterns’ tendency to cause unpredictability in solar power generation. One potential solution to these issues is the use of machine learning (ML) techniques for short term solar photovoltaic (PV) power forecasting. This research looks into how well various machine learning algorithms work for short-term PV power forecasting. It also offers a summary of the variables that affect time-series data performance and suggests high-performance methods. Long Short-term Memory (LSTM), Support Vector Machines (SVM), and Artificial Neural Networks (ANN) were tested on an 88,494-item dataset with 20 features at a 15-minute resolution. As accuracy metrics, root mean square error (RMSE) and mean square error (MSE) show that ANNs and LSTM perform better than SVM. The superiority of the suggested methods is demonstrated by a careful comparison with recently published results. The reason for the effectiveness of ANNs and LSTMs is their capacity to represent the complex and non-linear relationships found in the data. Short-term solar power forecasting is complex, and SVMs are better equipped to handle linear relationships. These findings will be a useful guidance for those who intend to work in the field of PV power forecasting, even though further research is required to generalize this observation. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 9(1), 2022 P 82-95

Solar photovoltaic power forecasting system with online manner based on adaptive mode decomposition and multi-objective optimization

Constructing an accurate and reliable solar photovoltaic (PV) power forecasting system is crucial for smart grid management and dispatch. However, due to the intermittent, non-stationary and random nature of solar energy, current methods cannot effectively capture the dynamic change patterns of PV data, resulting in a forecasting accuracy that cannot satisfy the requirement for stable operation of smart grids. To address this issue, in this paper, a new solar PV power forecasting system is proposed based on a new adaptive mode decomposition method that combines machine learning models. First an adaptive mode decomposition method is proposed for adaptively eliminating high-frequency information of PV data, mitigating the impact of high noise on the forecasting performance. Then the powerful mapping ability of machine learning model is utilized to construct a solar PV power generation combination forecasting system. The system adopts a multi-objective optimization strategy to determine the optimal weights of the combination model, which reduces the volatility of the forecasting results and enhances the stability of the forecasting system. Experimental results demonstrate that the proposed system achieves the best overall performance with online manner on 10 state-of-the-art baselines. Furthermore, the evaluation analysis by Diebold–Mariano test, improvement ratio and sensitivity analysis further show that the stability and accuracy of the proposed system outperforms the comparative baselines.

IoT based solar power forecasting using SSA-ELM technique

Abstract The optimizing of renewable energy use and grid integration relies on accurate solar power predictions. In order to predict the amount of power that solar photovoltaic (PV) systems would produce inside an IoT framework, this study suggests a new method that integrates Singular Spectrum Analysis (SSA) with Extreme Learning Machine technology. The SSA algorithm makes sense of solar power data by separating it into its component parts, such as trend, seasonality, and noise. The ELM model, a quick and effective feedforward neural network with a single hidden layer, takes these broken-down parts as input characteristics. In order to enhance the accuracy of solar power forecasts, the suggested strategy combines the decomposition skills of SSA with the predictive capability of ELM. Data acquired by solar PV sensors is input into the IoT-based forecasting model, which then undergoes preprocessing with SSA, feature extraction, model training with ELM, and performance evaluation. The SSA-ELM methodology has been successfully tested on real solar power data and has shown promising results in terms of accuracy measures such as low mean absolute error and mean absolute percentage error. By implementing the suggested method, accurate projections of solar output can be made, leading to better energy management, lower costs, and the smooth incorporation of renewables into smart grids. A dependable and computationally efficient method for solar forecasting in Internet of Things applications is provided by the combination of SSA and ELM.

Short-term forecasting of rooftop retrofitted photovoltaic power generation using machine learning

This paper explores short-term forecasting of rooftop retrofitted photovoltaic (PV) power generation using a Neural Networks (NN) model, highlighting its importance for energy management and grid integration. The study used data from the Universiti Malaysia Pahang Al-Sultan Abdullah (UMPSA) of the Faculty of Electrical & Electronics Engineering Technology (FTKEE), capturing a 570.4 kWp rooftop PV system. The data, collected at 15-min intervals from January 29 to February 4, 2024, included thirty-three input features such as ambient temperature, horizontal irradiation, AC voltages, AC currents, and Maximum Power Point Trackers (MPPTs). The target was the total active power in kilowatts. The methodology involved partitioning the data into a training set covering the first five days and a testing set for the last two days. The NN model was compared with other machine learning approaches, including Long Short-Term Memory Networks (LSTM), Gated Recurrent Units (GRU), Random Forest (RF), and k-Nearest Neighbors (k-NN). Performance metrics: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Maximum Error, and Standard Deviation were used to evaluate the models. The results showed that the NN model outperformed all other models, achieving an RMSE of 0.7098, an MAE of 0.3629, a Maximum Error of 3.901, and a Standard Deviation of 0.7076. These findings suggest that NN effectively capture complex patterns in rooftop PV system data, contributing to enhanced reliability and efficiency in short-term solar power forecasting. The study's implications extend to improved grid management and energy efficiency, underlining the significance of advanced machine learning techniques in renewable energy forecasting.

Optimal Scheduling of a Cascade Hydropower Energy Storage System for Solar and Wind Energy Accommodation

The massive grid integration of renewable energy necessitates frequent and rapid response of hydropower output, which has brought enormous challenges to the hydropower operation and new opportunities for hydropower development. To investigate feasible solutions for complementary systems to cope with the energy transition in the context of the constantly changing role of the hydropower plant and the rapid evolution of wind and solar power, the short-term coordinated scheduling model is developed for the wind–solar–hydro hybrid pumped storage (WSHPS) system with peak shaving operation. The effects of different reservoir inflow conditions, different wind and solar power forecast output, and installed capacity of pumping station on the performance of WSHPS system are analyzed. The results show that compared with the wind–solar–hydro hybrid (WSH) system, the total power generation of the WSHPS system in the dry, normal, and wet year increased by 10.69%, 11.40%, and 11.27% respectively. The solar curtailment decreased by 68.97%, 61.61%, and 48.43%, respectively, and the wind curtailment decreased by 76.14%, 58.48%, and 50.91%, respectively. The high proportion of wind and solar energy connected to the grid in summer leads to large net load fluctuations and serious energy curtailment. The increase in the installed capacity of the pumping station will promote the consumption of wind and solar energy in the WSHPS system. The model proposed in this paper can improve the operational flexibility of hydropower station and promote the consumption of wind and solar energy, which provides a reference for the research of cascade hydropower energy storage system.

Enhancing Solar Power Forecasting using Grasshopper optimization and Whale Optimization Algorithm

Solar power forecasting is essential for effectively integrating solar energy into the power system. Accurate forecasting allows for more effective power system planning, operation, and management. In this research work, we employed the Grasshopper Optimization Algorithm (GOA) and Whale Optimization Algorithm (WOA) to choose features for solar power forecasting utilizing time series data from the OPSD dataset. The dataset contains measurements taken at 15-minute intervals, giving a wealth of data for training and verifying forecasting algorithms. The WOA is used to adjust the parameters of a forecasting model, hence increasing its accuracy and reliability. The suggested approach's performance is evaluated on a large-scale dataset, with training, validation, and test sets of 100,000, 50,000, and 51,347 data points, respectively. The results show that the WOA is effective at improving solar power forecasting accuracy, which contributes to the efficient use of renewable energy resources.

Solar power forecasting using domain knowledge

Integrating renewable energy into the existing energy market is crucial. Solar power forecasting is essential since it depends on weather parameters and must integrate with the central grid to use the produced solar power effectively. Contemporary studies indicate that machine learning has the potential to predict the future generation of solar energy based on past data. This research demonstrates a broad range of solar power forecasting, combining the one-year time series solar power generation data, solar panel physical features, and weather information with the help of machine learning and deep learning tools with domain knowledge. The dataset is curated and preprocessed. We propose a deep learning ensemble model based on BI-LSTM. The proposed model can forecast well regardless of geographical position and is able to predict both short-term and long-term time horizons. We compared the results of the proposed model with the existing dataset and multiple standard deep learning models and found that our model produced better performance than traditional models. We also validated our model using different solar plants in Durgapur, India. For long-term forecasting, our model also outperformed the base model.

Moving beyond the Aerosol Climatology of WRF-Solar: A Case Study over the North China Plain

Abstract Numerical weather prediction (NWP), when accessible, is a crucial input to short-term solar power forecasting. WRF-Solar, the first NWP model specifically designed for solar energy applications, has shown promising predictive capability. Nevertheless, few attempts have been made to investigate its performance under high aerosol loading, which attenuates incoming radiation significantly. The North China Plain is a polluted region due to industrialization, which constitutes a proper testbed for such investigation. In this paper, aerosol direct radiative effect (DRE) on three surface shortwave radiation components (i.e., global, beam, and diffuse) during five heavy pollution episodes is studied within the WRF-Solar framework. Results show that WRF-Solar overestimates instantaneous beam radiation up to 795.3 W m−2 when the aerosol DRE is not considered. Although such overestimation can be partially offset by an underestimation of the diffuse radiation of about 194.5 W m−2, the overestimation of the global radiation still reaches 160.2 W m−2. This undesirable bias can be reduced when WRF-Solar is powered by Copernicus Atmosphere Monitoring Service (CAMS) aerosol forecasts, which then translates to accuracy improvements in photovoltaic (PV) power forecasts. This work also compares the forecast performance of the CAMS-powered WRF-Solar with that of the European Centre for Medium-Range Weather Forecasts model. Under high aerosol loading conditions, the irradiance forecast accuracy generated by WRF-Solar increased by 53.2% and the PV power forecast accuracy increased by 6.8%. Significance Statement Numerical weather prediction (NWP) is the “go-to” approach for achieving high-performance day-ahead solar power forecasting. Integrating time-varying aerosol forecasts into NWP models effectively captures aerosol direct radiation effects, thereby enhancing the accuracy of solar irradiance forecasts in heavily polluted regions. This work not only quantifies the aerosol effects on global, beam, and diffuse irradiance but also reveals the physical mechanisms of irradiance-to-power conversion by constructing a model chain. Using the North China Plain as a testbed, the performance of WRF-Solar on solar power forecasting on five severe pollution days is analyzed. This version of WRF-Solar can outperform the European Centre for Medium-Range Weather Forecasts model, confirming the need for generating high spatial–temporal NWP.