Neural networks for pattern-based short-term load forecasting: A comparative study

Dudek, Grzegorz

doi:10.1016/j.neucom.2016.04.021

Cited by 93 publications

(61 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Dentre alguns exemplos estão os modelos auto regressivos integrados baseados em médias móveis (ARIMA) e os modelos de suavização exponencial de Holt-Winters de acordo com [1]. No entanto, com o crescimento das pesquisas em inteligência computacional, modelos baseados em redes neurais passaram a ser bastante utilizados como em [2,3,10].…”

Section: Introductionunclassified

Modelos de Previsão de Carga: Abordagens imunoinspiradas baseadas em Similaridade de Padrões

Silva¹,

Lisboa²,

Vieira³

et al. 2018

Proceedings XIII Brazilian Congress on Computational Inteligence

View full text Add to dashboard Cite

O problema de previsão de carga a curto prazo consiste em estimar a demanda futura de um período de até uma semana e possui fundamental importância nas operações de um sistema elétrico. Tal problema pode ser modelado em termos de similaridades de padrões, permitindo o uso de algoritmos de regressão baseados em aprendizado de máquina aplicados a análise de séries temporais. Neste trabalho, dois modelos imunoinspirados são avaliados e incorporados a um método de transformação. Os experimentos apresentam resultados interessantes, comprovando a aplicabilidade dos modelos e da estratégia adotada.A demanda de carga pode ser considerada como uma das informações mais importantes a serem utilizadas na operação de sistemas de energia. A operação dos sistemas de energiaé bastante dependente de tal informação, da qual decorre o funcionamento bem sucedido de geradores, subestações e linhas, assim como o planejamento de novas unidades no sistema de distribuição de energia.Desta forma, a realização da previsão da carga a curto prazo possui um papel fundamental, sobretudo nas tarefas de controle, agendamento e segurança dos sistemas elétricos. Considerando este fato, um grande número de modelos têm sido propostos, sejam estes baseados em regressão, análise estatística,

show abstract

Section: Introductionunclassified

Modelos de Previsão de Carga: Abordagens imunoinspiradas baseadas em Similaridade de Padrões

Silva¹,

Lisboa²,

Vieira³

et al. 2018

Proceedings XIII Brazilian Congress on Computational Inteligence

View full text Add to dashboard Cite

show abstract

“…ANNs can in this case perform better than, for example, a Markov chain predictor. ANNs are used for load forecasting for cities, utility plants, and buildings . ANNs can be used without load history as input data .…”

Section: Introductionmentioning

confidence: 99%

“…Instead, input data can be based on temperature information, which is often easier to obtain. Because of the modeling of patterns, the problem of forecasting multiple seasonal nonstationary time series can be simplified using ANNs . The degree of variation in demand data can be reduced by dividing data into classes before using them as input data to ANNs …”

Section: Introductionmentioning

confidence: 99%

“…Because of the modeling of patterns, the problem of forecasting multiple seasonal nonstationary time series can be simplified using ANNs. 19 The degree of variation in demand data can be reduced by dividing data into classes before using them as input data to ANNs. 20 Although ANNs are proved useful for load classification in many cases, there is 1 aspect that cannot easily be obtained; ie, their model parameters (or weight parameters) are not easily translated to real-world events.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Modeling electricity load curves with hidden Markov models for demand-side management status estimation

Andersson

2016

Int. Trans. Electr. Energ. Syst.

View full text Add to dashboard Cite

Summary The development of the smart grid will lead to increased possibilities to improve energy efficiency via demand‐side management (DSM) and ultimately contribute to a more sustainable and secure energy system. The smart grid will offer great opportunities for developing methods for automatic analyses of various states in the energy system. This paper proposes a method that aims to classify load curves and estimate current power reduction capacity from DSM. The method is based on the hidden Markov model where the hidden states represent customers' demand behaviors in an aggregated load curve. The method is tested on 3 aggregated load curves, based on measured data from 36 Swedish companies. Results show that the method can recognize when DSM has been utilized as well when DSM has not been utilized. Results also show that the energy system becomes more stable, from a system dynamical point of view, when DSM is utilized.

show abstract

“…As shown in paper [8], an artificial neural network (ANN) model with a back-propagation algorithm was presented for short-term load forecasting in micro-grid power systems. In paper [9], several univariate approaches based on neural networks were proposed and compared, which included multi-layer perceptron, radial basis function neural network, generalized regression neural network, fuzzy counter propagation neural networks, and self-organizing maps. Rana et al [10] introduced an advanced wavelet neural networks for very short-term load forecasting, and the complex electricity load data was decomposed into different frequencies by the proposed method to separately realize the load forecasting.…”

mentioning

confidence: 99%

The Short-Term Power Load Forecasting Based on Sperm Whale Algorithm and Wavelet Least Square Support Vector Machine with DWT-IR for Feature Selection

Liu

2017

Sustainability

View full text Add to dashboard Cite

Short-term power load forecasting is an important basis for the operation of integrated energy system, and the accuracy of load forecasting directly affects the economy of system operation. To improve the forecasting accuracy, this paper proposes a load forecasting system based on wavelet least square support vector machine and sperm whale algorithm. Firstly, the methods of discrete wavelet transform and inconsistency rate model (DWT-IR) are used to select the optimal features, which aims to reduce the redundancy of input vectors. Secondly, the kernel function of least square support vector machine LSSVM is replaced by wavelet kernel function for improving the nonlinear mapping ability of LSSVM. Lastly, the parameters of W-LSSVM are optimized by sperm whale algorithm, and the short-term load forecasting method of W-LSSVM-SWA is established. Additionally, the example verification results show that the proposed model outperforms other alternative methods and has a strong effectiveness and feasibility in short-term power load forecasting. regression analysis was introduced to forecast short-term power load. Zhao et al. [6] presented an improved grey model optimized by a new optimization algorithm Ant Lion Optimizer for annual power load forecasting and the rolling mechanism was introduced to improve the forecasting accuracy. Coelho et al. [7] used a novel hybrid evolutionary fuzzy model with parameter optimization for load forecasting in a MG scenario. As can be seen from the above references, this kind of forecasting method has a mature theoretical basis, and its calculation process is simpler and easier to understand; however, this kind of method is unable to deal with non-linear forecasting problem due to theoretical limitation, its application object is often relatively simple, and the prediction accuracy usually fails to meet the expectations.With the development of modern computer technology and the continuous improvement of mathematical theory, intelligent forecasting methods are favored by many scholars for deep research. Artificial neural network (ANN) and support vector machine (SVM) are usually attractive to many scholars for solving short-term power load forecasting problems. As shown in paper [8], an artificial neural network (ANN) model with a back-propagation algorithm was presented for short-term load forecasting in micro-grid power systems. In paper [9], several univariate approaches based on neural networks were proposed and compared, which included multi-layer perceptron, radial basis function neural network, generalized regression neural network, fuzzy counter propagation neural networks, and self-organizing maps. Rana et al. [10] introduced an advanced wavelet neural networks for very short-term load forecasting, and the complex electricity load data was decomposed into different frequencies by the proposed method to separately realize the load forecasting. In paper [11], an effective short-term load forecasting model based on the generalized regression neural network with decreasing step fru...

show abstract

Neural networks for pattern-based short-term load forecasting: A comparative study

Cited by 93 publications

References 30 publications

Modelos de Previsão de Carga: Abordagens imunoinspiradas baseadas em Similaridade de Padrões

Modelos de Previsão de Carga: Abordagens imunoinspiradas baseadas em Similaridade de Padrões

Modeling electricity load curves with hidden Markov models for demand-side management status estimation

The Short-Term Power Load Forecasting Based on Sperm Whale Algorithm and Wavelet Least Square Support Vector Machine with DWT-IR for Feature Selection

Contact Info

Product

Resources

About