Prediction Of Power Output Research Articles

Offshore wind power output prediction based on convolutional attention mechanism

ABSTRACT To solve the problem that offshore wind power generation is susceptible to extreme weather conditions that result in low accuracy of monitoring data, this work proposes a method for predicting offshore wind power generation output by combining convolution and attention mechanisms. The method combines bidirectional long and short-term memory (BiLSTM) network with attention mechanism (AM). The input data are first weighted with the AM for reducing the predictive weight of the interference data, and then the attention mechanism-weighted data are fed into the BiLSTM network. The bi-directional propagation neural network can effectively utilize all the input information, resulting in higher prediction accuracy. The method of combining BiLSTM network with AM is compared with the method employing long and short-term memory network, gated recurrent unit, and bidirectional long and short-term memory network alone through simulation. The mean square error (MSE) of the BiLSTM combined with AM method is 41.28% smaller than the MSE of the best LSTM method among the compared methods and 58.43% smaller than the average of the compared methods. The R2-R-Square is 14.21% larger than the R2-R-Square of the LSTM and 32.91% larger than the average of the compared methods. The results show that the proposed method of combining convolution and attention mechanism for offshore wind power generation output prediction has higher prediction accuracy.

Resource efficient PV power forecasting: Transductive transfer learning based hybrid deep learning model for smart grid in Industry 5.0

This paper presents an innovative approach for enhancing power output forecasting of Photovoltaic (PV) power plants in dynamic environmental conditions using a Hybrid Deep Learning Model (DLM). The hybrid DLM employs a synergy of Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) network, and Bidirectional LSTM (Bi-LSTM), effectively capturing spatial and temporal dependencies within weather data crucial for accurate predictions. To optimize the DLM’s performance efficiently, a unique Kepler Optimization Algorithm (KOA) is introduced for hyperparameter tuning, drawing inspiration from Kepler’s laws of planetary motion. By leveraging KOA, the DLM attains optimal hyperparameter configurations, elevating power output prediction precision. Additionally, this study integrates Transductive Transfer Learning (TTL) with the deep learning models to enhance resource efficiency. By leveraging knowledge gained from previously learned tasks, TTL enables the DLM to improve its forecasting capabilities while minimizing resource utilization. Datasets encompassing environmental parameters and PV plant-generated power across diverse sites are employed for DLM training and testing. Three hybrid models, amalgamating KOA, CNN, LSTM, and Bi-LSTM techniques, are introduced and evaluated. Comparative assessment of these models across distinct PV sites yields insightful observations. Performance evaluation, focused on short-term PV power forecasting, underscores the hybrid DLM’s superiority over individual CNN and LSTM models. This hybrid approach achieves remarkable accuracy and resilience in predicting power output under varying weather conditions, showcasing its potential for efficient PV power plant management.

Analysis of power system scheduling operation mode based on multi-objective optimization algorithm

Abstract Exploring the majorization strategy of the power system (PS) dispatching operation is to achieve economic cost reduction and reduce environmental pollution. In this paper, starting from the PS dispatching model, the adaptive Corsi variance is introduced to get rid of the local optimum using particle swarm majorization procedure, and the adaptive Corsi variance multiple swarm coevolutionary procedure is constructed through coevolutionary strategy and information sharing strategy. The MCPSO-ACPM procedure is used to optimize the PS scheduling operation model, and experiments are conducted on both load and unit for the optimized scheduling model. From the load majorization results, the peak-to-valley variance is concentrated from 176.02KW to 110.51KW compared with the original load, and the peak-to-valley ratio is reduced by 0.718, which saves customers 98.63 yuan in electricity purchase cost. From the scheduling majorization prediction, the PS output power prediction value of 1 min during the day is closest to the actual measured value of output power, and its prediction deviation is about 2.67%. This shows that the use of a multi-objective majorization procedure can realize the optimal dispatch of PS and achieve the reduction of economic cost.

Electrical power output prediction of combined cycle power plants using a recurrent neural network optimized by waterwheel plant algorithm

It is difficult to analyze and anticipate the power output of Combined Cycle Power Plants (CCPPs) when considering operational thermal variables such as ambient pressure, vacuum, relative humidity, and temperature. Our data visualization study shows strong non-linearity in the experimental data. We observe that CCPP energy production increases linearly with temperature but not pressure. We offer the Waterwheel Plant Algorithm (WWPA), a unique metaheuristic optimization method, to fine-tune Recurrent Neural Network hyperparameters to improve prediction accuracy. A robust mathematical model for energy production prediction is built and validated using anticipated and experimental data residuals. The residuals’ uniformity above and below the regression line suggests acceptable prediction errors. Our mathematical model has an R-squared value of 0.935 and 0.999 during training and testing, demonstrating its outstanding predictive accuracy. This research provides an accurate way to forecast CCPP energy output, which could improve operational efficiency and resource utilization in these power plants.

Intelligent solar photovoltaic power forecasting

This paper presents a day-ahead forecasting method for photovoltaic (PV) power plants in commercial sectors. The method is based on numerical weather prediction (NWP) models from open weather maps and power plant specifications. The output of the model is the predicted power output from the PV power plant, which is incorporated into an optimal control strategy of the PV plant using battery storage. The use of optimal algorithms assists in the PV power plant curtailment in cases of over-generation and reduces the dependence on conventional power sources such as generators in cases of under-generation by the PV plant. It was found that most forecasting methods do not incorporate PV power and storage systems for proper optimization and demand management. This can be seen as a gap for further research of forecasting models integrated with battery storage systems to improve PV power system performance. Results obtained show a good performance of the developed model. A root means square error (RMSE) of 425.79 W and 595.10 W and a mean absolute error (MAE) of 246.26 W and 238 W were achieved for a summer and winter day, respectively. Furthermore, an excellent positive correlation exists between the predicted output power and the observed results, with R2 values over 90%.

Estimation of Short-Term Power of Wind Turbines Using Artificial Neural Network (ANN) and Swarm Intelligence

The integration of wind power into the electricity grid faces a significant challenge due to the unpredictable nature of wind speed fluctuations. Therefore, ensuring precise short-term predictions of power output from wind turbines is vital for effectively incorporating wind energy into the grid and proficiently managing power systems. In this paper, an Artificial Neural Network (ANN)-based approach for the short-term power forecasting of wind turbines based on a swarm intelligence algorithm is proposed. Also, a simulation study of the wind power real system at different wind speeds is presented by using MATLAB/Simulink. The swarm intelligence algorithm is employed to optimize the forecasting model parameters. The performance of the proposed algorithm is evaluated using real data from a wind farm in Turkey. Three distinct methodologies are utilized to process the data efficiently: ANN, ANN with Firefly Algorithm (ANN-FA), and ANN with Particle Swarm Optimization (ANN-PSO). The results demonstrate that the swarm intelligence algorithm outperforms traditional forecasting methods, such as statistical approaches and machine learning techniques, in terms of accuracy and reliability. Furthermore, the computational efficiency of the algorithm is examined, and it is shown that the swarm intelligence-based approach offers a practical solution for real-time forecasting applications. The algorithm’s scalability and adaptability make it suitable for large-scale wind farms with multiple turbines, as it can handle the inherent variability and uncertainties associated with wind power generation. The proposed method offers an accurate and reliable forecasting tool that can assist power system operators and energy market participants in making informed decisions for the efficient utilization of wind energy resources.

Study on hydroturbine power trend prediction based on machine learning

The application of computational fluid dynamics combined with 3D modeling of the hydraulic model was discussed in this paper. Through the complex calculation of the internal low field, dynamic simulation was used to predict the characteristics of water energy index. Based on the collection of a large number of different types of turbine operation data in different working conditions, the improved multi-layer neural network power prediction model based on the adaptive inverse normalization strategy and the improved grey wolf algorithm were used to optimize the least squares support vector machine power prediction model. Rlue activation function was used in the improved multi-layer neural network model uses to improve the computation speed and prevent gradient disappearance. Combined with L2 regularization to prevent data overfitting, Adam gradient optimizer was applied to achieve fast convergence of network output. The simulation results shown that the method can realize the high precision prediction of turbine output power.

Elman short-term wind power prediction based on the optimized seagull algorithm

Accurate prediction of wind farm power output can relieve the pressure of grid frequency regulation and peak regulation and improve grid stability. With the goal of improving power prediction accuracy and reducing overall prediction error, this paper proposes an Elman short-term wind power prediction model on the basis of an optimized seagull algorithm. Firstly, the Elman network is used as the base prediction model, and the seagull algorithm is applied to seek the best values for its weights. Secondly, the chaotic circle mapping with better initial characteristics is improved to equalize its sequence distribution for optimizing the population initialization. Then, to address the lack of local search capability, an optimized iterative approach using the sine cosine operator is used to achieve a balance between local exploitation ability and global search capability. Finally, after simulation and analysis of the actual data set, it is verified that the model has a better prediction effect.

Novel Cycling Assessment Protocol Accurately Predicts Power Output At Lactate Breakpoint And Maximum Oxygen Consumption

Novel Cycling Assessment Protocol Accurately Predicts Power Output At Lactate Breakpoint And Maximum Oxygen Consumption

A Study on Short‐Term Wind Power Forecasting Method Based on Wind Speed Spatio‐Temporal Calibration and Power Self‐adaptive Correction

Accurately predicting wind power is premise for the large‐scale integration of wind energy into the power grid and realizing reliable power grid scheduling and efficient wind farm operation. The spatial–temporal differences between the acquisition points of numerical weather prediction and wind farm sites should be considered because they significantly affect the accuracy of wind power output. Meanwhile, since there are different actual constraints on wind farm electrical networks and units, and the weather conditions may vary, there are no universal methods for identifying and correcting wind power output. To this end, this paper proposes a wind speed calibration model based on an Attention‐Bi‐LSTM network and a self‐adaptive recognition‐correction model based on SHAP‐CART classification for wind power output prediction. First, a spatial–temporal dataset is established by using real weather data from wind farm sites and data from the surrounding NWP points. Then, wind speed is accurately calibrated using the Attention‐Bi‐LSTM model. Next, considering the practical constraints on wind farm electrical networks and units, the self‐adaptive recognition‐correction model based on SHAP‐CART classification is employed to effectively recognize and predict the operating states and power output of the units, thus improving the accuracy and stability of wind power output prediction. Finally, the effectiveness and superiority of the proposed method is verified by using a wind farm in Eastern China as an example and analyzing the prediction results in various approaches. © 2023 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC.

Optimization of MEMS-based Energy Scavengers and output prediction with machine learning and synthetic data approach

Micro-electromechanical systems (MEMS) based energy scavengers can serve as perpetual small-scale power sources. In this context, this work presents a novel approach for optimizing MEMS-based energy scavengers by predicting output power and natural frequency of operation through Synthetic Data and Machine Learning (ML) algorithms. For optimization, the work considers a bimorph cantilever-based energy scavenger, which consists of an elastic layer sandwiched between two piezoelectric layers. Various design parameters and their effect on the performance of the energy scavengers were evaluated through different ML models. The exploratory data analysis and preprocessing steps were carried out over the generated data, and subsequently, prediction of frequency and power was performed; 80% of the data was used for training purposes, and 20% of the data was used for testing purposes. Extreme Gradient Boost (XGBoost) was observed to be the best among the ML models for predicting frequency and power - with the R2 score of 97% and 86%, respectively. The optimized scavenger can generate an output power of 5.06 mW with an output voltage of 11.02 V. This approach can help the researchers by predicting the output power and resonant frequency, opening new research opportunities in the optimization of MEMS-based Energy Scavengers.

Optimization of BP Neural Network for Wind Power Output Prediction Based on Particle Swarm Optimization Algorithm

Because wind power has strong uncertainty and uncontrollability, there are great difficulties in the process of grid connection and power dispatching. The existing prediction models have some shortcomings, such as insufficient precision and poor convergence. In this paper, an improved short-term prediction model is proposed. The correlation coefficient is determined by the grey correlation method, and then the optimal coefficient is calculated by optimizing the backpropagation network using a particle swarm optimization algorithm. Comparing the prediction results of the new model with those of a single BP network and the POS optimization BP network, it can be concluded that the new model is more precise than the other models.

Unit Combination Scheduling Method Considering System Frequency Dynamic Constraints under High Wind Power Share

Power systems with a high wind power share are characterized by low rotational inertia and weak frequency regulation, which can easily lead to frequency safety problems. Providing virtual inertia for large-scale wind turbines to participate in frequency regulation is a solution, but virtual inertia is related to wind power output prediction. Due to wind power prediction errors, the system inertia is reduced and there is even a risk of instability. In this regard, this article proposes a unit commitment model that takes into account the constraints of sharp changes in frequency caused by wind power prediction errors. First, the expressions of the equivalent inertia, adjustment coefficient, and other frequency influence parameters of the frequency aggregation model for a high proportion wind power system are derived, revealing the mechanism of the influence of wind power prediction power and synchronous machine start stop status on the frequency modulation characteristics of the system. Second, the time domain expression of the system frequency after the disturbance is calculated by the segment linearization method, and the linear expressions of “frequency drop speed and frequency nadir” constraints are derived to meet the demand of frequency regulation in each stage of the system. Finally, a two-stage robust optimization model based on a wind power fuzzy set is constructed by combining the effects of wind power errors on power fluctuation and frequency regulation capability. The proposed model is solved through affine decision rules to reduce its complexity. The simulation results show that the proposed model and method can effectively improve the frequency response characteristics and increase the operational reliability of high-share wind power systems.

Photovoltaic Power Output Prediction Based on TabNet for Regional Distributed Photovoltaic Stations Group

With the increasing proportion of distributed photovoltaic (DPV) installations in county-level power grids, to improve the centralized operation and maintenance of the stations and to meet the needs of power grid dispatching, the output of the county-level regional DPV stations group needs to be predicted. In this paper, the weather prediction information is used to predict the output based on the model input average strategy. To eliminate the effect of the selected non-optimal training sample collection period on the prediction accuracy, an ensemble prediction method based on the minimum redundancy maximum relevance criterion and TabNet model is carried out. To reduce the influence of weather prediction errors on the power output prediction, a modified model based on error prediction is proposed. The ensemble prediction model is used to predict the day-ahead output, and a combination prediction model based on the proposed ensemble prediction model and the proposed modified model is established to predict the hour-ahead output. The experimental results verify the effectiveness of the proposed models. Compared with the corresponding reference models, the proposed ensemble prediction method reduces the normalized mean absolute errors (nMAEs) and the normalized root mean square errors (nRMSEs) of the day-ahead output prediction results by 2.86% and 5.51%, respectively. The combination prediction model reduces the nMAE and nRMSE of the hour-ahead output prediction results by 3.05% and 3.05%, respectively. Therefore, the prediction accuracy can be improved by the proposed models.

Advanced Machine Learning Techniques for Accurate Very-Short-Term Wind Power Forecasting in Wind Energy Systems Using Historical Data Analysis

Accurate wind power forecasting plays a crucial role in the planning of unit commitments, maintenance scheduling, and maximizing profits for power traders. Uncertainty and changes in wind speeds pose challenges to the integration of wind power into the power system. Therefore, the reliable prediction of wind power output is a complex task with significant implications for the efficient operation of electricity grids. Developing effective and precise wind power prediction systems is essential for the cost-efficient operation and maintenance of modern wind turbines. This article focuses on the development of a very-short-term forecasting model using machine learning algorithms. The forecasting model is evaluated using LightGBM, random forest, CatBoost, and XGBoost machine learning algorithms with 16 selected parameters from the wind energy system. The performance of the machine learning-based wind energy forecasting is assessed using metrics such as mean absolute error (MAE), mean-squared error (MSE), root-mean-squared error (RMSE), and R-squared. The results indicate that the random forest algorithm performs well during training, while the CatBoost algorithm demonstrates superior performance, with an RMSE of 13.84 for the test set, as determined by 10-fold cross-validation.

A Review of Modern Wind Power Generation Forecasting Technologies

The prediction of wind power output is part of the basic work of power grid dispatching and energy distribution. At present, the output power prediction is mainly obtained by fitting and regressing the historical data. The medium- and long-term power prediction results exhibit large deviations due to the uncertainty of wind power generation. In order to meet the demand for accessing large-scale wind power into the electricity grid and to further improve the accuracy of short-term wind power prediction, it is necessary to develop models for accurate and precise short-term wind power prediction based on advanced algorithms for studying the output power of a wind power generation system. This paper summarizes the contribution of the current advanced wind power forecasting technology and delineates the key advantages and disadvantages of various wind power forecasting models. These models have different forecasting capabilities, update the weights of each model in real time, improve the comprehensive forecasting capability of the model, and have good application prospects in wind power generation forecasting. Furthermore, the case studies and examples in the literature for accurately predicting ultra-short-term and short-term wind power generation with uncertainty and randomness are reviewed and analyzed. Finally, we present prospects for future studies that can serve as useful directions for other researchers planning to conduct similar experiments and investigations.

Wind power regression prediction based on stacked LSTMs with attention mechanisms for evaluating technological improvement effects of wind turbines

Accurately evaluating the technological improvement effects of wind turbines is crucial for wind farm operators. To this end, this paper proposes an innovative approach that employs a wind power regression model which leverages external environmental information to predict the output power of wind turbines. The effectiveness of technological improvements can be evaluated by comparing the predicted output power with the measured output power. In this paper, a model called stacked LSTM networks with attention mechanisms is designed. In the proposed model, the stacked LSTM networks are used to enhance the nonlinear fitting ability and capture deeper features of the input sequence. Furthermore, temporal attention mechanisms are employed to make the model focus on important time-series information of the data. In addition, a hierarchical attention mechanism is designed to explore the correlation among the outputs of the stacked LSTM networks and enrich the model’s output information. The experiments on the data from a wind farm show that the proposed method outperforms various wind power prediction benchmarks, achieving lower RMSE, MAE, and MAPE values of 142.82, 104.2, and 4.85%, respectively.

A design of ultra-short-term power prediction algorithm driven by wind turbine operation and maintenance data for LSTM-SA neural network

Due to factors such as meteorology and geography, the generated power of wind turbines fluctuates frequently. In this way, power changes should be predicted in grid connection to take control measures in time. In this paper, an operation and maintenance data-driven LSTM-SA (long short-term memory with self-attention) prediction algorithm is designed to predict the ultra-short-term power of wind turbines. First, the wind turbine operation and maintenance data, including wind speed, blade deflection angle, yaw angle, humidity, and temperature, are subjected to feature selection by using the Pearson correlation coefficient method and the Lasso algorithm, thereby establishing the correlation between wind speed, blade deflection angle, and out power. Then, full-connect neural network is trained to establish a mapping model of wind speed, blade deflection angle, and out power. The power change rate k is calculated by the derivative of output power to wind speed. Finally, based on the historical power data and the power change rate k, the LSTM neural network power prediction model is trained to calculate the output power prediction value. In order to increase the training efficiency and reduce the delay, the self-attention mechanism is used to optimize the hidden layer of the LSTM model. The test results show that, compared with similar prediction algorithms, this algorithm has higher prediction accuracy, faster convergence speed, and better stability, which can solve the problem of accurately predicting ultra-short-term power when wind power training data is inadequate.

SOLAR ENERGY FORECASTING WITH DEEP LEARNING TECHNIQUE

The increasing reliance on renewable energy sources, such as solar power, necessitates accurate forecasting to ensure efficient grid integration and stability. This study explores the application of deep learning techniques, particularly deep neural networks (DNNs), in predicting solar irradiance and power output. By leveraging advanced algorithms and large datasets, this research aims to enhance the precision of solar energy forecasts, thereby optimizing grid management and resource allocation. Deep learning techniques, characterized by their ability to model complex nonlinear relationships, offer significant advantages over traditional statistical methods in forecasting solar energy. This study focuses on the development and evaluation of deep neural networks for predicting solar irradiance, which directly influences the power output of photovoltaic (PV) systems. The models are trained on extensive historical weather data, satellite imagery, and real-time solar output measurements to capture temporal and spatial variations in solar energy. Key components of the research include data preprocessing to handle missing values and noise, feature extraction to identify relevant patterns, and model training using various architectures of DNNs such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs). The study also employs ensemble learning techniques to combine predictions from multiple models, further enhancing forecast accuracy. Results demonstrate that deep neural networks significantly outperform traditional forecasting methods in terms of accuracy and reliability. The improved forecasts enable better scheduling of energy generation and distribution, reducing reliance on fossil fuels and minimizing grid imbalances. Additionally, accurate solar energy predictions support the efficient integration of solar power into the grid, enhancing overall system stability and reducing operational costs. The study also addresses challenges in implementing deep learning models, such as computational requirements and the need for large, high-quality datasets. Solutions include leveraging cloud computing resources and developing standardized data collection protocols to facilitate broader application and scalability of the models. Application of deep learning techniques in solar energy forecasting represents a promising advancement for the renewable energy sector. By improving the accuracy of solar irradiance and power output predictions, deep neural networks contribute to more reliable grid integration and efficient energy management. This research advocates for the continued exploration and adoption of advanced machine learning methods to support the transition to a sustainable energy future. Keywords: Deep Learning Techniques; Solar Energy; Forecasting; Grid Integration; Power Output.

Evaluation of Machine Learning Methods on Large-Scale Spatiotemporal Data for Photovoltaic Power Prediction

The exponential increase in photovoltaic (PV) arrays installed globally, particularly given the intermittent nature of PV generation, has emphasized the need to accurately forecast the predicted output power of the arrays. Regardless of the length of the forecasts, the modeling of PV arrays is made difficult by their dependence on weather. Typically, the model projections are generated from datasets at one location across a couple of years. The purpose of this study was to compare the effectiveness of regression models in very short-term deterministic forecasts for spatiotemporal projections. The compiled dataset is unique given that it consists of weather and output power data of PVs located at five cities spanning 3 and 6 years in length. Gated recurrent unit (GRU) generalized the best for same-city and cross-city predictions, while long short-term memory (LSTM) and ensemble bagging had the best cross-city and same-city predictions, respectively.