Prediction Of Photovoltaic Power Output Research Articles

Short term hydropower scheduling considering cumulative forecasting deviation of wind and photovoltaic power

The cumulative forecasting errors of wind and photovoltaic (PV) power pose serious challenges to the short-term scheduling of hydropower stations that connected to the same regional power grid. The variable and intermittent nature of wind and PV power necessitate a flexible power source to achieve power balance. Consequently, hydropower, with its flexible adjustment capability, plays a crucial role in supporting the consumption of wind and PV power stations in short-term scheduling. Meanwhile, to address the grid's peak shaving requirement, the scheduling of cascaded hydropower system becomes increasingly complex. Given the limitations of current wind and PV power output prediction technologies, achieving precise forecast remains challenging. In practice, the actual outputs of wind and PV power stations often diverge considerably from predicted values over multiple consecutive periods. This variability may result in the underutilization of wind and PV power if the reserve capacity of hydropower stations is insufficient. Therefore, it is essential to account for the cumulative deviation in electricity from wind and PV sources. By computing probability distribution and establishing boundary condition of the cumulative deviation, the operation of the cascaded hydropower system can be controlled more effectively, thus enhancing the stability of a multi-energy complementary system. This study employs Gaussian Mixture Model (GMM) to characterize the probability distributions of cumulative electricity deviation across multiple periods of wind and PV power and takes it as restriction to regulate the operation of hydropower stations. In order to balance the deviation electricity, the downward electricity reserve, upward electricity reserve and the storage energy of the cascaded reservoirs are introduced as restrictive conditions. The proposed Hydro-wind-PV peak shaving model is then applied to the short-term optimal scheduling of the Lancang River cascaded hydropower system to validate the feasibility and efficacy.

Deep learning-based evaluation of photovoltaic power generation

Photovoltaic (PV) power generation has emerged as a rapidly growing renewable energy source. However, the PV system output’s intermittent and weather-dependent nature poses challenges when integrating with the power grid. These challenges manifest as critical issues, including voltage fluctuations, harmonic distortion, and current deviation, making it difficult to accurately predict grid conditions. Moreover, the spatial variability caused by PV system intermittency further complicates the situation. To address these challenges and ensure efficient grid integration, this paper proposes a comprehensive approach encompassing deep learning-based state prediction of PV power output. The paper introduces the utilization of a long short-term memory (LSTM) model, a type of deep learning architecture, for learning patterns from historical PV power generation data and weather forecasts. The LSTM model enables accurate predictions for effective grid management by capturing long-term dependencies in PV power generation data. Real-world PV power generation data was employed to evaluate the proposed approach. The results demonstrated the significant improvement in PV power generation prediction accuracy achieved by the LSTM model compared to traditional methods. The proposed approach offers a promising solution for addressing the challenges associated with PV system integration into the power grid. It enables enhanced grid planning, resource allocation, and protection measures to accommodate the increasing penetration of solar energy harnessed through PV systems and related power electronics interfaces.

Efficient shrinkage temporal convolutional network model for photovoltaic power prediction

Power stations operating on PhotoVoltaic (PV) power generation are challenged by demand forecasting as PV power generation is random and intermittent. This paper presents an Efficient Shrinkage Temporal Convolutional Network (ESTCN) model, which combines the Temporal Convolutional Network (TCN) and an improved Deep Residual Shrinkage Network (DRSN) to forecast PV power output. First, the attention sub-network of the DRSN is improved, the fully connected layer of the sub-network is canceled, and feature extraction is done directly on the model features following global average pooling using a one-dimensional convolution to obtain cross-channel interaction and increased model efficiency. Next, an improved attention mechanism and adaptive soft thresholding are introduced into TCN to automatically determine the noise threshold to address the issue of information weight dispersion caused by redundant information in the input samples. By incorporating dilated causal convolution, attention module, soft thresholding, and residual connection, the ESTCN model is formed and is shown to enhance PV power output prediction with only a minimal increase in the number of parameters. The power station in Ningxia, China is adopted as a validation example with data taken from the year 2020. The fitting and prediction results of the ESTCN model are compared against the Convolutional Neural Network, Long Short-Term Memory, and TCN models. The ESTCN model yielded the following values of the evaluation metrics: Root Mean Square Error (RMSE) of 1.47 kW, Mean Absolute Error (MAE) of 0.79 kW, and coefficient of determination of 0.999. Compared to the TCN model, the prediction errors based on RMSE and MAE improved by 0.39 kW and 0.18 kW, respectively. Applying the ESTCN model to predict PV power generation of power stations in two regions in China, Ningxia and Xinjiang, shows the ESTCN model to be superior, scalable, and universal over other deep learning models.

Distributed photovoltaic power output prediction based on satellite cloud map video frames

To address the challenge of predicting distributed photovoltaic (PV) power output for improved system integration and stability, this study proposes a novel method. Given the expanding scale of distributed PV systems and their economic constraints, accurate power output prediction becomes pivotal. Conventional prediction methods are hindered by the lack of meteorological stations at most distributed PV stations. In response, we present a dynamic convolutional generative adversarial network (DC-GAN) approach coupled with satellite cloud map video frames. By extracting shading features from satellite cloud images and utilizing DC-GAN, our method forecasts short-term cloud shading effects on future radiation. We further integrate radiation data from centralized PV stations, spatial correlations of distributed PV stations, and cloud shading characteristics. This information is used to construct a predictive model combining Convolutional Neural Networks (CNN) and Long short-term memory (LSTM), enhancing prediction accuracy. Comparative experiments confirm the superiority of our proposed method over traditional approaches, substantiating its effectiveness and practicality. Our method achieves notable accuracy improvements, establishing its value in predicting distributed PV power output. This research’s findings offer a valuable contribution to the field of renewable energy integration. In numerical assessments, our method demonstrates a significant increase in prediction accuracy, outperforming conventional techniques by 3.3%. This underscores the practical relevance and efficiency of our proposed approach in enhancing distributed PV power output prediction.

A hybrid machine learning method with explicit time encoding for improved Malaysian photovoltaic power prediction

Nowadays, with the growing interest in green energy, further improvements in photovoltaic (PV) power systems are needed. In this regard, the main aim is to find an optimal method to predict the output power of PV systems to maintain a sustainable operation. Hence, this work proposes a hybrid Machine Learning (ML) method LASSO-RFR for an hourly PV power output prediction. The model consists of Least Absolute Shrinkage and Selection Operator (LASSO) and Random Forest Regressor (RFR), where the former model makes a prediction and the latter fine tune it by the addition or subtraction of a relatively small value. The proposed model outperformed other models when tested on real data recorded from 2016 to 2019 for three Malaysian PV systems, namely Thin-Film (TF), Monocrystalline (MC), and Polycrystalline (PC). LASSO-RFR attained the lowest root mean square error (RMSE) of 23.7, 18.2, and 20.8 Wh/m2 for the TF, MC, and PC, respectively. This work also highlights the importance of explicit time encoding in improving PV power prediction. Although it is used to be ignored in the literature when developing ML models, the time feature is the second most influencing factor of PV power prediction after solar irradiance, as shown by the SHAP analysis (shapely additive explanation). For the study implications, the developed prediction model can assist the industry in predicting 1 h ahead of PV power output, demand-side management, and building operations and maintenance.

Prediction of photovoltaic power output based on similar day analysis using RBF neural network with adaptive black widow optimization algorithm and K-means clustering

Solar photovoltaic power generation has become the focus of the world energy market. However, weak continuity and variability of solar power data severely increase grid operating pressure. Therefore, it is necessary to propose a new refined and targeted forecasting method to broaden the forecasting channels. In this paper, a hybrid model (KM-SDA-ABWO-RBF) based on radial basis function neural networks (RBFNNs), adaptive black widow optimization algorithm (ABWO), similar day analysis (SDA) and K-means clustering (KM) has been developed. The ABWO algorithm develops adaptive factors to optimize the parameters of RBFNNs and avoid getting trapped in local optima. SDA and K-means clustering determine the similarity days and the optimal similarity day through meteorological factors and historical datasets. Nine models compared forecast accuracy and stability over four seasons. Experiments show that compared with other well-known models on the four indicators, the proposed KM-SDA-ABWO-RBF model has the highest prediction accuracy and is more stable.

Prediction of photovoltaic power output based on different non-linear autoregressive artificial neural network algorithms

Prediction of power output plays a vital role in the installation and operation of photovoltaic modules. In this paper, two photovoltaic module technologies, amorphous silicon and copper indium gallium selenide installed outdoors on the rooftop of the University of Dodoma, located at 6.5738° S and 36.2631° E in Tanzania, were used to record the power output during the winter season. The average data of ambient temperature, module temperature, solar irradiance, relative humidity, and wind speed recorded is used to predict the power output using a non-linear autoregressive artificial neural network. We consider the Levenberg-Marquardt optimization, Bayesian regularization, resilient propagation, and scaled conjugate gradient algorithms to understand their abilities in training, testing and validating the data. A comparison with reference to the performance indices: coefficient of determination, root mean square error, mean absolute percentage error, and mean absolute bias error is drawn for both modules. According to the findings of our investigation, the predicted results are in good agreement with the experimental results. All the algorithms performed better, and the predicted power out of both modules using the Bayesian regularization algorithm is observed to exhibit good processing capabilities compared to the other three algorithms that are evident from the measured performance indices.

Comparative study of machine learning approaches for predicting short-term photovoltaic power output based on weather type classification

To combat the worsening global energy shortage, global photovoltaic (PV) installation capacity has been increasing rapidly every year. Since the instability and intermittence of PV power output have great impacts on utility grids, accurate PV power output prediction is crucial. This paper proposes the use of machine learning approaches, combined with a weather type classification method, to predict short-term PV power output. The datasets are collected from a commercial PV power station located in Yangjiang, Guangdong province of China (latitude 21.56 °N, longitude 112.09 °E). Firstly, daytime meteorological data from 07:30 to 18:00 are divided into six 2-h intervals, and then the meteorological conditions of each interval are divided into four categories using an Extremely randomized Trees Classification model according to the PV power generation in each period. Secondly, nine machine learning models are established based on the weather type classification to predict the PV power output. The results show that weather type classification is vital to the selection of appropriate machine learning models and the accurate prediction of PV power output because the characteristic correlation between the meteorological data and PV power output always changes. In general, the Lasso Regressor, Random Forest Regressor, Gradient Boosting Regressor, and Support Vector Regressor models show better performances than the other models. Furthermore, all the models’ accuracy is relatively high when the local meteorological conditions are relatively stable, such as in October, November, and December, during which time the Mean Relative Error values are 2.07, 1.07, and 1.73, respectively. During the period when the weather is unstable, the performance of the SVR model is better than that of the other models. The prediction accuracy can be significantly improved with integrating the accurate weather classification into the model. With regards to each daytime period, the prediction accuracy in the morning and evening is relatively high and the MREs for these times are small. This study provides a theoretical basis for selecting appropriate machine learning models to predict photovoltaic power generation under different weather conditions.

Ultra‐short‐term photovoltaic power forecasting of multifeature based on hybrid deep learning

The output power of photovoltaic power has randomness and volatility, which poses new challenges to the peak shaving and dispatching of the power system. Therefore, the accurate prediction of photovoltaic power output is an effective way to maintain the security and stability of the power grid. In this article, a hybrid prediction model based on improved convolutional neural network and bidirectional gated recurrent unit is proposed to predict PV output power. Firstly, the concept of time relevance is introduced to fuse this feature with weather characteristics as the input of the prediction model. Secondly, the time series generated adversarial neural network is applied to the field of photovoltaic power generation prediction for the first time to enhance the dataset obtained. Thirdly, the model proposed in this article is trained and compared with other models under different characteristics, different weather conditions, and different forecast time steps. The experimental results show that the hybrid deep learning model proposed in this article has better prediction performance. Under different characteristics, the model proposed in this article has the highest prediction accuracy. After data enhancement, the prediction accuracy of each model has been improved by 1% to 3%. Among them, the prediction accuracy of the model proposed in this article is 0.977 and 0.960, respectively. Under different weather conditions and different time steps, the proposed model has the best stability. Under cloudy and rainy weather, the prediction accuracy of the proposed model only decreases by about 6%. In addition, the prediction accuracy of the proposed model only decreases by approximately 1.5% to 4% from 15 to 180 minutes.

Solar Photovoltaic Forecasting of Power Output Using LSTM Networks

The penetration of renewable energies has increased during the last decades since it has become an effective solution to the world’s energy challenges. Among all renewable energy sources, photovoltaic (PV) technology is the most immediate way to convert solar radiation into electricity. Nevertheless, PV power output is affected by several factors, such as location, clouds, etc. As PV plants proliferate and represent significant contributors to grid electricity production, it becomes increasingly important to manage their inherent alterability. Therefore, solar PV forecasting is a pivotal factor to support reliable and cost-effective grid operation and control. In this paper, a stacked long short-term memory network, which is a significant component of the deep recurrent neural network, is considered for the prediction of PV power output for 1.5 h ahead. Historical data of PV power output from a PV plant in Nicosia, Cyprus, were used as input to the forecasting model. Once the model was defined and trained, the model performance was assessed qualitative (by graphical tools) and quantitative (by calculating the Root Mean Square Error (RMSE) and by applying the k-fold cross-validation method). The results showed that our model can predict well, since the RMSE gives a value of 0.11368, whereas when applying the k-fold cross-validation, the mean of the resulting RMSE values is 0.09394 with a standard deviation 0.01616.

Accurate prediction of photovoltaic power output based on long short‐term memory network

An accurate power output prediction of the photovoltaic system is pivotal to eliminate the extra cost and the negative impact in the utility grid integrated with photovoltaic power sources. The power output of a photovoltaic system is predicted by introducing a long short-term memory method. Moreover, the influence of noise data on prediction results is eliminated with the empirical mode decomposition. To further improve the accuracy and stability of the prediction method, the parameters of long short-term memory neural networks are determined with a sine cosine algorithm. The performances of the long short-term memory method in terms of root mean square error, mean absolute error, and coefficient of determination in January and August are analysed, respectively. Compared with other prediction schemes, the long short-term memory method provides superior accuracy for photovoltaic power output prediction.

Prediction of photovoltaic power output based on similar day analysis, genetic algorithm and extreme learning machine

Recently, many machine learning techniques have been successfully employed in photovoltaic (PV) power output prediction because of their strong non-linear regression capacities. However, single machine learning algorithm does not have stable prediction performance and sufficient generalization capability in the prediction of PV power output. In this work, a hybrid model (SDA-GA-ELM) based on extreme learning machine (ELM), genetic algorithm (GA) and customized similar day analysis (SDA) has been developed to predict hourly PV power output. In the SDA, Pearson correlation coefficient is employed to measure the similarity between different days based on five meteorological factors, and the data samples similar to those from the target forecast day are selected as the training set of ELM. This operation can effectively increase the number of useful samples and reduce the time consumption on training data. In the ELM, the optimal values of the hidden bias and the input weight are searched by GA to improve the prediction accuracy. The performance of the proposed forecast model is evaluated with coefficient of determination (R2), mean absolute error (MAE) and normalized root mean square error (nRMSE). The results show that the SDA-GA-ELM model has higher accuracy and stability in day-ahead PV power prediction.

Study on impacts of dust accumulation and rainfall on PV power reduction in East China

The photovoltaic (PV) power plants’ power generation is affected obviously by the cleanliness of the photovoltaic modules. The dust is the primary source causing the pollution. Natural dust deposition is affected by human activities and meteorological factors such as temperature, humidity, wind speed and PM10 concentration in the region where PV modules are installed. The research efforts so far have been focused on predicting PV power outputs, however, in the real operations the PV outputs can significantly deviate from the prediction due to the dust pollution which, unfortunately, has not been fully investigated. This paper, to the authors’ best knowledge, is the first research to analyze the dust composition and evaluate its impacts on the PV power outputs with the real field test data in East China Collected through a series of experiments completed in the lab of the Institute of Distributed Generation and Microgrid, Zhejiang University of Technology, in East China. A PV output forecasting model was set up first based on the irradiance and the component’s temperature. Then a complete meteorological and electrical data online monitoring system was developed for data acquisition. The morphology of the dust particles on the investigated photovoltaic modules was observed by a scanning electron microscope, and the composition of the dust particles was measured by X-ray fluorescence. The research found that the dust accumulated on the surface of photovoltaic modules tend to form clusters under the influence of rainfall which is one major cause on fast dropping PV outputs. The analysis shows the dust deposit mainly comprises of SiO2 and CaCO3. Further analysis indicates in East China, the average dust density on PV modules is 0.644 g/m2 in a week and consequently the dust reduces the PV output power by 7.4% one week. The research is of significance not only in terms of giving the real-data based analysis on dust properties deposited on PV panels, but also on revealing the relationship between power output attenuation and dust deposition, which helps improve the PV power output prediction accuracy and develop efficient cleaning strategies for photovoltaic modules.

Integrating Gray Data Preprocessor and Deep Belief Network for Day-Ahead PV Power Output Forecast

Generation output forecasting is a crucial task for planning and sizing of a photovoltaic (PV) power plant. The purpose of this paper is to present an effective model for day-ahead forecasting PV power output of a plant based on deep belief network (DBN) combined with gray theory-based data preprocessor (GT-DBN), where the DBN attempts to learn high-level abstractions in historical PV output data by utilizing hierarchical architectures. Test results obtained by the proposed model are compared with those obtained by other five forecasting methods including autoregressive integrated moving average model, back propagation neural network, radial basis function neural network, support vector regression, and DBN alone. It shows that the proposed model is superior to other models in forecasting accuracy and is suitable for day-ahead PV power output prediction.

Time series forecasting of solar power generation for large-scale photovoltaic plants

Accurate solar power forecasting is essential for grid-connected photovoltaic (PV) systems especially in case of fluctuating environmental conditions. The prediction of PV power output is critical to secure grid operation, scheduling and grid energy management. One of the key elements in PV output prediction is time series analysis especially in locations where the historical solar radiation measurements or other weather parameters have not been recorded. In this work, several time series prediction methods including the statistical methods and those based on artificial intelligence are introduced and compared rigorously for PV power output prediction. Moreover, the effect of prediction time horizon variation for all the algorithms is investigated. Hourly solar power forecasting is carried out to verify the effectiveness of different models. The data utilized in the current work comprises 3640 h of operation data taken from a 20 MW grid-connected PV station in China.

Evaluation of Photovoltaic Power Generation by Using Deep Learning in Solar Panels Installed in Buildings

Southern Taiwan has excellent solar energy resources that remain largely unused. This study incorporated a measure that aids in providing simple and effective power generation efficiency assessments of solar panel brands in the planning stage of installing these panels on roofs. The proposed methodology can be applied to evaluate photovoltaic (PV) power generation panels installed on building rooftops in Southern Taiwan. In the first phase, this study selected panels of the BP3 series, including BP350, BP365, BP380, and BP3125, to assess their PV output efficiency. BP Solar is a manufacturer and installer of photovoltaic solar cells. This study first derived ideal PV power generation and then determined the suitable tilt angle for the PV panels leading to direct sunlight that could be acquired to increase power output by panels installed on building rooftops. The potential annual power outputs for these solar panels were calculated. Climate data of 2016 were used to estimate the annual solar power output of the BP3 series per unit area. The results indicated that BP380 was the most efficient model for power generation (183.5 KWh/m2-y), followed by BP3125 (182.2 KWh/m2-y); by contrast, BP350 was the least efficient (164.2 KWh/m2-y). In the second phase, to simulate meteorological uncertainty during hourly PV power generation, a surface solar radiation prediction model was developed. This study used a deep learning–based deep neural network (DNN) for predicting hourly irradiation. The simulation results of the DNN were compared with those of a backpropagation neural network (BPN) and a linear regression (LR) model. In the final phase, the panel of module BP3125 was used as an example and demonstrated the hourly PV power output prediction at different lead times on a solar panel. The results demonstrated that the proposed method is useful for evaluating the power generation efficiency of the solar panels.

Improvement of Short-Term BIPV Power Predictions Using Feature Engineering and a Recurrent Neural Network

The time resolution and prediction accuracy of the power generated by building-integrated photovoltaics are important for managing electricity demand and formulating a strategy to trade power with the grid. This study presents a novel approach to improve short-term hourly photovoltaic power output predictions using feature engineering and machine learning. Feature selection measured the importance score of input features by using a model-based variable importance. It verified that the normative sky index in the weather forecasted data had the least importance as a predictor for hourly prediction of photovoltaic power output. Six different machine-learning algorithms were assessed to select an appropriate model for the hourly power output prediction with onsite weather forecast data. The recurrent neural network outperformed five other models, including artificial neural networks, support vector machines, classification and regression trees, chi-square automatic interaction detection, and random forests, in terms of its ability to predict photovoltaic power output at an hourly and daily resolution for 64 tested days. Feature engineering was then used to apply dropout observation to the normative sky index from the training and prediction process, which improved the hourly prediction performance. In particular, the prediction accuracy for overcast days improved by 20% compared to the original weather dataset used without dropout observation. The results show that feature engineering effectively improves the short-term predictions of photovoltaic power output in buildings with a simple weather forecasting service.

Recurrent Neural Network-Based Hourly Prediction of Photovoltaic Power Output Using Meteorological Information

Recently, the prediction of photovoltaic (PV) power has become of paramount importance to improve the expected revenue of PV operators and the effective operations of PV facility systems. Additionally, the precise PV power output prediction in an hourly manner enables more sophisticated strategies for PV operators and markets as the electricity price in a renewable energy market is continuously changing. However, the hourly prediction of PV power outputs is considered as a challenging problem due to the dynamic natures of meteorological information not only in a day but also across days. Therefore, in this paper, we suggest three PV power output prediction methods such as artificial neural network (ANN)-, deep neural network (DNN)-, and long and short term memory (LSTM)-based models that are capable to understand the hidden relationships between meteorological information and actual PV power outputs. In particular, the proposed LSTM based model is designed to capture both hourly patterns in a day and seasonal patterns across days. We conducted the experiments by using a real-world dataset. The experimental results show that the proposed ANN based model fails to yield satisfactory results, and the proposed LSTM based model successfully better performs more than 50% compared to the conventional statistical models in terms of mean absolute error.

LSTM-Attention-Embedding Model-Based Day-Ahead Prediction of Photovoltaic Power Output Using Bayesian Optimization

Photovoltaic (PV) output is susceptible to meteorological factors, resulting in intermittency and randomness of power generation. Accurate prediction of PV power output can not only reduce the impact of PV power generation on the grid but also provide a reference for grid dispatching. Therefore, this paper proposes an LSTM-attention-embedding model based on Bayesian optimization to predict the day-ahead PV power output. The statistical features at multiple time scales, combined features, time features and wind speed categorical features are explored for PV related meteorological factors. A deep learning model is constructed based on an LSTM block and an embedding block with the connection of a merge layer. The LSTM block is used to memorize and attend the historical information, and the embedding block is used to encode the categorical features. Then, an output block is used to output the prediction results, and a residual connection is also included in the model to mitigate the gradient transfer. Bayesian optimization is used to select the optimal combined features. The effectiveness of the proposed model is verified on two actual PV power plants in one area of China. The comparative experimental results show that the performance of the proposed model has been significantly improved compared to LSTM neural networks, BPNN, SVR model and persistence model.

Evaluation of data-driven models for predicting solar photovoltaics power output

Abstract This research was undertaken to evaluate different inverse models for predicting power output of solar photovoltaic (PV) systems under different practical scenarios. In particular, we have investigated whether PV power output prediction accuracy can be improved if module/cell temperature was measured in addition to climatic variables, and also the extent to which prediction accuracy degrades if solar irradiation is not measured on the plane of array but only on a horizontal surface. We have also investigated the significance of different independent or regressor variables, such as wind velocity and incident angle modifier in predicting PV power output and cell temperature. We found wind velocity and module temperature measurements are too unreliable for use in inverse solar PV models. The inverse regression model forms have been evaluated both in terms of their goodness-of-fit, and their accuracy and robustness in terms of their predictive performance. Given the accuracy of the measurements, expected CV-RMSE of hourly power output prediction over the year varies between 3.2% and 8.6% when only climatic data are used. Depending on what type of measured climatic and PV performance data is available, different scenarios have been identified and the appropriate modeling pathways have been proposed. These models are to be implemented on a controller platform for optimum operational planning of microgrids and integrated energy systems.