To ensure the efficient operation and healthy development of photovoltaic power generation systems, reliable distributed photovoltaic power prediction is significant. However, the lack of modeling ability of existing ...
To ensure the efficient operation and healthy development of photovoltaic power generation systems, reliable distributed photovoltaic power prediction is significant. However, the lack of modeling ability of existing distributed photovoltaic power forecasting methods for dynamic spatiotemporal correlation of distributed photovoltaic power data limits the further improvement of prediction accuracy. Therefore, a distributed photovoltaic ultra-short-term power forecasting method based on spatial-temporal attention mechanism and graph convolutional networks is proposed in this paper. Firstly, distributed photovoltaic cluster is divided into sub-regions and graph structured data can be generated with each sub-region being a node on the graph. Then, spatial-temporal attention mechanism is used to adaptively capture spatiotemporal correlation information and dynamically update adjacency matrix. Finally, spatial-temporal convolution is used to extract deep spatiotemporal features of distributed photovoltaic power data and achieve regional power prediction. The validity and progressiveness of the model have been verified in real world data sets.
As the proportion of wind power in grid continues to increase, more and more newly built or expanded wind farms are facing problems such as lack of historical samples and insufficient training of prediction models. Wi...
As the proportion of wind power in grid continues to increase, more and more newly built or expanded wind farms are facing problems such as lack of historical samples and insufficient training of prediction models. With the increasing frequency of extreme weather events, most wind farms also face problems such as scarce samples of extreme weather scenarios and large power forecasting errors. In response to the above-mentioned issues, this paper proposes an wind power interval prediction method based on conditional generative adversarial networks (CGAN) and kernel extreme learning machines (KELM). With the historical samples of extreme weather scenarios, a large number of new samples are generated through CGAN. Based on KELM, an interval prediction model under extreme weather is constructed. Compared to point prediction, this model can provide richer probability information, and compared to other interval prediction methods, the results show an effective improvement in prediction performance.
With the rise of the application of sharing economy in various fields of powersystem, As a typical application of shared economy in the field of energy storage, the optimal allocation of shared energy storage on the ...
With the rise of the application of sharing economy in various fields of powersystem, As a typical application of shared economy in the field of energy storage, the optimal allocation of shared energy storage on the source-network-load side has been a great topic. The problem dealt with in this paper is the configuration result of the source-grid-load energy storage system under the same control strategy. This paper designs an optimization method for the source-network-load side configuration of generalized shared energy storage in regional power grid: Firstly, according to the extensional usage scenario and demand of energy storage, the alternative set of shared energy storage configuration is constructed. Secondly, based on the profit model of shared energy storage, a powersystem source-network-load optimization scheduling strategy is proposed. Finally, a numerical example is carried out to prove the feasibility and efficacy of the suggested method.
The loss of electricity during transmission cannot be ignored, and a large number of power consumers are seeking flexible and affordable distributed generation methods to reduce electricity costs. However, a large num...
The loss of electricity during transmission cannot be ignored, and a large number of power consumers are seeking flexible and affordable distributed generation methods to reduce electricity costs. However, a large number of distributed photovoltaic (PV) connections pose challenges to electric power dispatch and the security of powersystem. Accurate PV power forecast can solve these problems, which requires a large amount of accurate historical data to build forecast model. This article proposes a spatiotemporal interpolation method to complement historical data of PV sites. Wavelet packet transform (WPT) is applied to decompose original time sequence to stable sequence and fluctuant sequence. To eliminate the fluctuations caused by cloud movement, dynamic time warping (DTW) is used to regularize fluctuant sequences of adjacent PV sites. Finally, through spatial interpolation, a precise historical data can be obtained. The case study shows that under cloudy conditions, proposed interpolation method reduces the NRMSE by 0.51% compared to the direct interpolation algorithm.
This paper proposes a renewable energy prediction method adapted to the climate characteristics of plateau mountains and oceans in the southern region. The proposed approach contains four parts: data clearing, time-pe...
This paper proposes a renewable energy prediction method adapted to the climate characteristics of plateau mountains and oceans in the southern region. The proposed approach contains four parts: data clearing, time-period data clustering, modeling for renewable forecasts, and feature-prediction accuracy correlation matrix calculation. Therein, the forecast algorithm is improved mainly from two aspects: 1) Time-period clustering for data is implemented before training the forecast models. Different forecast models are trained with clustered data sets for achieving overall higher performance in forecasts compared with the one directly trained based on the whole data set. 2) We analyze the impact of complicated climate features, such as plateau mountains and ocean climate, on model prediction results, and select critical climate features for different prediction tasks based on the feature-prediction accuracy correlation matrix, indicating great improvements in renewable energy prediction.
Climate change and global warming threaten both nature and humanity. Renewable energy is an effective way to solve this problem. Distributed photovoltaics (DPV) have attracted much attention due to their low environme...
Climate change and global warming threaten both nature and humanity. Renewable energy is an effective way to solve this problem. Distributed photovoltaics (DPV) have attracted much attention due to their low environmental impact. But the uncertainty of DPV output is putting pressure on the distribution network. Therefore, DPV power prediction is very important. In this paper, the factors affecting DPV yield are reviewed and the forecasting methods are summarized. The results show that geographical location, weather parameters, photovoltaic panel characteristics, and data noise will affect the prediction results. The main forecasting methods are statistical forecasting and principle forecasting. Statistical forecasting methods perform better in forecasting. In the future, prediction algorithms will need to be improved to adapt to different conditions, such as weather and time scales. Data preprocessing algorithms also help to improve the accuracy and stability of predictions.
The measurement system in UHVDC systems mainly includes DC voltage measurement and DC current measurement, and the accuracy of its results is the key to ensure the reliable operation of the DC control and protection s...
The measurement system in UHVDC systems mainly includes DC voltage measurement and DC current measurement, and the accuracy of its results is the key to ensure the reliable operation of the DC control and protection systems. In this paper, based on the analysis of the principle of the UHVDC transmission system, the ±800kV UHVDC transmission system model is constructed using PSCAD/EMTDC simulation software. The effects of rectifier-side DC current and DC voltage measurement faults on the response characteristics of the DC control and protection systems are investigated under different scenarios by superimposing different abnormal measurement values, and the simulation results are analyzed in conjunction with the control strategy and protection principles. The simulation results show that the DC system exhibits different response characteristics under different measurement anomalies. The deviation of rectifier-side DC current measurement affects the DC voltage, DC current and power transmission of the two stations and causes the protections to operate. And the abnormal DC voltage measurement causes the protections to operate falsely.
When partial discharge occurs in the transformer, it will produce higher frequency pulse current, high frequency phenomenon will enhance the circuit capacitance effect, making the circuit bandwidth, in order to detect...
详细信息
ISBN:
(数字)9798331523558
ISBN:
(纸本)9798331523565
When partial discharge occurs in the transformer, it will produce higher frequency pulse current, high frequency phenomenon will enhance the circuit capacitance effect, making the circuit bandwidth, in order to detect the local discharge signal so as to realize the transformer for equipment monitoring, the establishment of the transformer broadband equivalent circuit model, the propagation law of partial discharge pulse current signal in the transformer is investigated, and the effects of four denoising methods are compared under different evaluation indexes; finally, a typical transformer partial discharge condition is studied, and the validity of the transformer partial discharge research based on the wide-band equivalent circuit is verified.
The traditional method for detecting and classifying different types of faults in Ultra High Voltage Direct Current (UHVDC) has low accuracy rate and does not sufficiently exploit the information of timing characteris...
The traditional method for detecting and classifying different types of faults in Ultra High Voltage Direct Current (UHVDC) has low accuracy rate and does not sufficiently exploit the information of timing characteristics in the electrical quantity signals, making it difficult to identify some system faults and measurement faults. This paper presents a fault diagnosis strategy which is based on Gated Recurrent Unit (GRU) to overcome these shortcomings. The study investigates three key GRU training factors, namely the amount of network layers, learning rate, and batch size, which may affect the training effect of GRU. And several sets of experiments are designed to determine the most suitable GRU network parameters. The classification effect of GRU is compared with that of RNN and LSTM, and the experimental results illustrate the high diagnostic accuracy of this approach. In addition, the method does not need to transform the data and can use the initial 1D time-series data as the network input, avoiding the loss of valid information and reducing the complexity of the fault diagnosis process.
As the core device in the field of communication, filter is developing towards miniaturization and high frequency. In this paper, a new dual frequency filter based on step impedance resonator is proposed. The resonant...
详细信息
暂无评论