Short-term Load Forecasting Using Deep Belief Network with Empirical Mode Decomposition and Local Predictor

Yan, Yu; Ji, Tianyao; Li, M. S.

doi:10.1109/pesgm.2018.8586129

Cited by 9 publications

(7 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generally, the architecture of the DL system comprises more layers than a traditional MLP, and it mimics the human brain function [88]. The DL concept's unique feature is helped widely due to its non‐linearity modelling capacity for various forecasting applications [89]. The various DL sub methods applied widely for STLFs application in ‘ Micro‐ grid ’ are ‘ Convolution neural networks (CNN), Deep belief networks (DBN), and Deep auto‐encoder (DAE) ’.…”

Section: Ai Methodsmentioning

confidence: 99%

A review on short‐term load forecasting models for micro‐grid application

Kondaiah

Saravanan

Sanjeevikumar

et al. 2022

The Journal of Engineering

View full text Add to dashboard Cite

Load forecasting (LF), particularly short‐term load forecasting (STLF), plays a vital role throughout the operation of the conventional power system. The precise modelling and complex analyses of STLF have become more significant in advanced microgrid (MG) applications. Several models are proposed for STLF and tested successfully in the literature. The selection of a forecasting method is mostly based on data availability and its objectives. This article presents a survey of the latest analytical and approximation techniques reported in the literature to model STLF in an MG environment. This article mainly focusses on the review on important methods applied to forecast renewable energy availability, energy demand, and price and load demand. Different models, their main objectives, methodology, error percentage, and so forth, are critically reviewed and analysed. For quick reference, we have highlighted the important points in the form tables. The researchers can quickly identify and frame their research problem related to the LF area by reading this review paper.

show abstract

Section: Ai Methodsmentioning

confidence: 99%

A review on short‐term load forecasting models for micro‐grid application

Kondaiah

Saravanan

Sanjeevikumar

et al. 2022

The Journal of Engineering

View full text Add to dashboard Cite

show abstract

“…A data-driven deep learning framework to load forecasting is proposed in [26], Box-Cox transformation is used to process data, and deep belief network is used for the load forecasting. A user-side load prediction method that integrates empirical mode decomposition and deep belief network is proposed in [27]; by decomposing the original load data into several eigenmode functions with different frequencies and amplitudes, the deep belief network is used to feature extraction and time-series prediction for each modal function. A load forecasting method based on ensemble empirical mode decomposition (EMMD) and Elman network is proposed in [28], the EEMD sample entropy was used to decompose the original power load sequence into a series of subsequences.…”

Section: Introductionmentioning

confidence: 99%

Short-Term Electrical Load Forecasting Based on VMD and GRU-TCN Hybrid Network

Cai

et al. 2022

Applied Sciences

View full text Add to dashboard Cite

With the continuous increase in user-side flexible controllable resources connected into a distribution system, the components of electrical load become too diverse and difficult to be accuracy forecasted. A short-term load forecast method that integrates variational modal decomposition (VMD), gated recurrent unit (GRU) and time convolutional network (TCN) into a hybrid network is proposed in this paper. Firstly, original electrical load sequence data with noise are decomposed into intrinsic IMF components with different frequencies and amplitudes based on the VMD method. Secondly, a combined load forecasting method based on the GRU and TCN network is proposed for the high and low-frequency load subsequent signals, respectively. Finally, the high and low-frequency signals forecasting results of the GRU and TCN network are rebuilt for the final load forecasting. The experiment results based on actual operation data (data set 1) and simulation data (data set 2), which show that the proposed method can reduce the forecasting error by 36.20% and 10.8%, respectively, in comparison with VMD-GRU. The reliability and accuracy of the proposed method is verified through the comparison with other methods such as LSTM, Prophet and XG Boost.

show abstract

“…The DL architecture includes different models, such as long short-term memory networks (LSTMN), convolutional neural networks (CNN), deep belief networks, and deep Boltzmann machine networks. Among them, LSTMN and CNN are popular in the problem of power load forecasting [10][11]. The main feature of the DL model is that the accuracy of the load-forecasted results highly depends on its hyperparameters.…”

Section: Introductionmentioning

confidence: 99%

Effects of Data Standardization on Hyperparameter Optimization with the Grid Search Algorithm Based on Deep Learning: A Case Study of Electric Load Forecasting

Ngoc

Đại

Minh

2022

Adv. technol. innov.

View full text Add to dashboard Cite

This study investigates data standardization methods based on the grid search (GS) algorithm for energy load forecasting, including zero-mean, min-max, max, decimal, sigmoid, softmax, median, and robust, to determine the hyperparameters of deep learning (DL) models. The considered DL models are the convolutional neural network (CNN) and long short-term memory network (LSTMN). The procedure is made over (i) setting the configuration for CNN and LSTMN, (ii) establishing the hyperparameter values of CNN and LSTMN models based on epoch, batch, optimizer, dropout, filters, and kernel, (iii) using eight data standardization methods to standardize the input data, and (iv) using the GS algorithm to search the optimal hyperparameters based on the mean absolute error (MAE) and mean absolute percent error (MAPE) indexes. The effectiveness of the proposed method is verified on the power load data of the Australian state of Queensland and Vietnamese Ho Chi Minh city. The simulation results show that the proposed data standardization methods are appropriate, except for the zero-mean and min-max methods.

show abstract

Short-term Load Forecasting Using Deep Belief Network with Empirical Mode Decomposition and Local Predictor

Cited by 9 publications

References 10 publications

A review on short‐term load forecasting models for micro‐grid application

A review on short‐term load forecasting models for micro‐grid application

Short-Term Electrical Load Forecasting Based on VMD and GRU-TCN Hybrid Network

Effects of Data Standardization on Hyperparameter Optimization with the Grid Search Algorithm Based on Deep Learning: A Case Study of Electric Load Forecasting

Contact Info

Product

Resources

About