Probabilistic electric load forecasting through Bayesian Mixture Density Networks

Brusaferri, Alessandro; Matteucci, Matteo; Spinelli, Stefano; Vitali, Andrea

doi:10.1016/j.apenergy.2021.118341

Cited by 26 publications

(8 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the non-linear and non-smooth nature of the power load, the ideal prediction accuracy is difficult to be achieved with a single prediction model due to structural limitations. Therefore, the second type of 'decomposition-prediction-reconstruction' model, which combines the advantages of signal processing and multiple prediction methods, has become the focus of research at this stage [13][14][15][16][17]. Wavelet decomposition, empirical modal decomposition, and variational modal decomposition algorithms have received much attention in short-term load forecasting [14].…”

Section: Introductionmentioning

confidence: 99%

“…However, parameter tuning in the VMD decomposition procedure is time-consuming and subjective, which impacts the decomposition's accuracy. References [15,16] proposed a prediction method using an optimization algorithm for the automatic optimization of key parameters of VMD, which achieved good prediction accuracy, but the model was complex and not adapted to time series. The permutation entropy (PE) method was suggested in [17] to analyze the complexity of each modal function of VMD and reorganize the modal function to obtain subsequences, effectively increasing the efficiency of VMD in processing time series.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Short‐term electrical load forecasting model based on multi‐dimensional meteorological information spatio‐temporal fusion and optimized variational mode decomposition

Wang,

Zhou,

et al. 2023

IET Generation Trans & Dist

View full text Add to dashboard Cite

This paper proposes a method to enhance the accuracy of power load forecasting by considering the variability in the impact of multi‐dimensional meteorological information on power load in diverse regions. The proposed method employs spatio‐temporal fusion (SF) of multi‐dimensional meteorological information and applies the Copula theory to analyze the non‐linear coupling of meteorological information from multiple stations with power load to achieve SF in the spatial dimension. To enhance the accuracy of load forecasting in the time dimension, this paper improves the core parameters of the variational mode decomposition (VMD) using the marine predators algorithm (MPA) and utilizes the weighted permutation entropy (WPE) to construct the MPA‐VMD fitness function for the adaptive decomposition of the load sequence. Moreover, this paper constructs input sets for the long short‐term memory model and the MPA‐LSSVM model by combining each component of the time dimension and each meteorological information of the spatial dimension to obtain the prediction results of each component. The prediction model corresponding to each component is selected according to the evaluation index and reconstructed to obtain the overall prediction results. The analysis results demonstrate that the proposed forecasting method outperforms the traditional forecasting method and effectively enhances the accuracy of power load forecasting.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Short‐term electrical load forecasting model based on multi‐dimensional meteorological information spatio‐temporal fusion and optimized variational mode decomposition

Wang,

Zhou,

et al. 2023

IET Generation Trans & Dist

View full text Add to dashboard Cite

show abstract

“…Due to the nonlinear and non-smooth nature of the electric load, the ideal prediction accuracy is difficult to be achieved with a single prediction model due to structural limitations. Therefore, the second type of "decomposition-prediction-reconstruction" model, which combines the advantages of signal processing and multiple prediction methods, has become the focus of research at this stage [13][14][15][16][17]. Wavelet decomposition, empirical modal decomposition, and variational modal decomposition algorithms have received much attention in short-term load forecasting [14].…”

mentioning

confidence: 99%

“…Ref. [15][16] proposed a prediction method using an optimization algorithm for the automatic optimisation of key parameters of VMD, which achieved good prediction accuracy, but the model was complex and not adapted to time series. The permutation entropy method was suggested in the Ref.…”

mentioning

confidence: 99%

Short Term Electrical Load Combination Forecasting Model based on Multi-dimensional Meteorological Information Spatio-Temporal Fusion and MPA-VMD

Wang

Zhou

et al. 2023

Preprint

View full text Add to dashboard Cite

This paper considers the variability in the impact of multi-dimensional meteorological information on power load in different regions. To improve the accuracy of load forecasting in the spatial dimension, the method of spatio-temporal fusion (SF) of multi-dimensional meteorological information is proposed. The Copula theory is applied to analyze the nonlinear coupling of meteorological information such as wind speed, rainfall, temperature, and sunshine intensity from multiple meteorological stations with the power load and to achieve spatio-temporal fusion. In the time dimension, the core parameters of the variational mode decomposition (VMD) are improved by the marine predators algorithm (MPA). The weighted permutation entropy (WPE) is used to construct the MPA-VMD fitness function for the adaptive decomposition of the load sequence. In addition, the input sets of the LSTM model and MPA-LSSVM model are constructed by combining each component of the time dimension and each meteorological information of spatial dimension to obtain the prediction results of each component. The prediction model corresponding to each component is selected according to the evaluation index, and reconstructed to obtain the overall prediction results. The analysis results show that the proposed forecasting method is better than the traditional forecasting method and effectively improves the accuracy of power load forecasting.

show abstract

“…In order to prevent local optimization in the training process,Liu et al (2021) proposes a hybrid CNN model by integrating genetic algorithm (GA) and a particle swarm optimization (PSO) to collaboratively optimize the network hyperparameters and weights. A data-driven load probability density forecasting method is developed based on Orthogonal Maximum Correlation Coefficient feature selection and Convolutional Gated Recurrent Unit (CGRU) quantile regression to deal with the uncertainties of wind power and load demand(Liu et al, 2022) Brusaferri et al (2022). proposed a probabilistic load forecasting framework using Bayesian Mixture Density Networks to enhance the mapping capabilities of neural networks and developed an end-to-end training method to discover the latent functional relation to conditioning variables and characterize the inherent load stochasticity and parameters uncertainty.…”

mentioning

confidence: 99%

A novel forward operator-based Bayesian recurrent neural network-based short-term net load demand forecasting considering demand-side renewable energy

Wen

Xian

Chen³

et al. 2022

Front. Energy Res.

View full text Add to dashboard Cite

Currently, traditional electricity consumers are now shifting to a new role of prosumers since more integration of renewable energy to demand side. Accurate short-term load demand forecasting is significant to safe, stable, and reliable operation of a renewable energy-dominated power system. In this paper, a short-term load forecasting model based on a bidirectional long short-term memory network (Bi-LSTM) using kernel transfer operator is proposed to achieve short-term load demand forecasting. To consider the influence of seasonality, holiday effects and weather on load demand forecasting, and simultaneously to improve the accuracy and performance of the forecasting model, this paper implements the dimensionality reduction of the input data by introducing an improved kernel transfer operator based on the Perron-Frobenius method. On this basis, a Bayesian Bi-LSTM for short-term load demand forecasting is formulated to obtain the probability prediction interval of load demand. To verify the validity of the proposed method, the actual historical load demand in a certain region in China was used for model training and verification. The forecasting results are compared with several conventional load demand methods using probability prediction technics. Simulation analysis shows that the proposed method can effectively reflect the short-term uncertainty of load demand, and is superior to the conventional methods in terms of forecasting accuracy and computational performance.

show abstract

Probabilistic electric load forecasting through Bayesian Mixture Density Networks

Cited by 26 publications

References 43 publications

Short‐term electrical load forecasting model based on multi‐dimensional meteorological information spatio‐temporal fusion and optimized variational mode decomposition

Short‐term electrical load forecasting model based on multi‐dimensional meteorological information spatio‐temporal fusion and optimized variational mode decomposition

Short Term Electrical Load Combination Forecasting Model based on Multi-dimensional Meteorological Information Spatio-Temporal Fusion and MPA-VMD

A novel forward operator-based Bayesian recurrent neural network-based short-term net load demand forecasting considering demand-side renewable energy

Contact Info

Product

Resources

About