Abstract
Highly accurate short-term load forecasting (STLF) is essential in the operation of power systems. However, the existing predictive methods cannot achieve an effective balance between prediction accuracy and computational cost. Furthermore, the prediction residual is rarely used to improve the predictive accuracy in STLF. This paper proposes a novel decomposition-based ensemble model for the STLF task. First, an optimized empirical wavelet transform (OEWT) is developed to rationally decompose the STLF load by combining the approximate entropy method with the empirical wavelet transform. Particularly, OEWT improves both prediction accuracy and computational cost in STLF. Second, a new hybrid machine learning method (named master learner) is proposed by rationally combining long short-term memory networks (LSTMs) with broad learning system (BLS) in STLF, effectively strengthening the predictive accuracy without significantly increasing the computational cost. Third, a residual learning model (named residual learner) is developed in the master learner to extract the effective predictive information from residual results, further improving the prediction accuracy in STLF. Fourth, an auxiliary learner is proposed by introducing another BLS to connect the input and output of the proposed model, enhancing the predictive robustness. The proposed decomposition-based ensemble model is compared with state-of-the-art and traditional models in STLF. Experimental results show that the model not only has high predictive accuracy and robustness but also low computational cost.
Similar content being viewed by others
Abbreviations
- AL:
-
Auxiliary learner
- ANN:
-
Artificial neural network
- APEN:
-
Approximate entropy
- BLS:
-
Broad learning system
- BLSTM:
-
Hybrid model composed of BLS and LSTM
- CD:
-
Critical distance of the Nemenyi test
- DBN:
-
Deep neural network
- DNN:
-
Deep neural network
- DWT:
-
Discrete wavelet transform
- EMD:
-
Empirical mode decomposition
- EWT:
-
Empirical wavelet transform
- FT:
-
Fourier transform
- LSTM:
-
Long short-term memory
- MAE:
-
Means absolute error
- ML:
-
Master learner
- NSW:
-
New South Wales, Australia
- OEWT:
-
Optimized empirical wavelet transform
- RL:
-
Residual learner
- RMSE:
-
Root means square error
- SLTF:
-
Short-term load prediction
- SVR:
-
Support vector regression
- SSA:
-
Singular spectrum analysis
- VMD:
-
Variational mode decomposition
- WT:
-
Wavelet transforms
- a f :
-
The activation function
- B + :
-
The pseudo-inverse of B
- DL :
-
Whether to have the direct link between the input layer and output layer
- d :
-
The distance between two reconstructed m dimensional vectors En
- E k :
-
The enhancement nodes of BLS
- E n :
-
The reconstructed m-dimensional vector by APEN
- eta :
-
The learning rate
- e n :
-
The components decomposed by EWT
- f RBF :
-
The radial basis functions
- f t :
-
The forget gate of LSTM
- g t :
-
The input node of LSTM
- h t :
-
The intermediate output of LSTM
- I :
-
The identity matrix
- i t :
-
The input gate of LSTM
- J i :
-
The feature nodes of BLS
- M :
-
The maximum value of the Fourier spectrum
- m :
-
The length of the comparison vector in APEN
- m i :
-
The maximum number of iterations
- N :
-
The number of components of EWT
- n :
-
The total number of test samples
- n ds :
-
The number of data sets in Nemenyi test
- n e :
-
The number of enhancement nodes
layer and output layer
- n f :
-
The number of feature nodes
- n h :
-
The number of hidden nodes
- n l :
-
The number of learning algorithms in Nemenyi test
- o t :
-
The output gate of LSTM
- q α :
-
The critical value based on the Studentized Range statistic
- r :
-
The similarity in APEN
- r b :
-
The random batch size of each time
- r m :
-
The randomization methods
- std :
-
The standard deviation of the time series
- s t :
-
The memory cell state of LSTM
- s RBF :
-
The spread of radial basis functions
- v m :
-
The momentum
- W :
-
The weight matrix between the feature nodes, enhancement nodes, and output of BLS
- X :
-
The input data of BLS
- \(\hat{Y}\) :
-
The final prediction value of BLS
- Y :
-
The actual value
- y i :
-
The load data at a certain moment
- y max :
-
The maximum value in the load data
- y min :
-
The minimum value in the load data
- \(\tilde{y}_{i}\) :
-
The normalized result of yi
- \({\hat{y}}_i\) :
-
The predicted data of yi
- κ :
-
The relative amplitude ratio of EWT
- ω n :
-
The midpoint of the corresponding frequencies of the two adjacent maximum values above the threshold
- τ n :
-
The transition phase of EWT
- ρ :
-
The interval for reorganizing the components en(t)
- λ:
-
The regularization parameter of BLS
- F −1(∙):
-
The inverse Fourier transform
- η(∙):
-
The linear transformation
- ξ(∙):
-
The nonlinear activation function
- σ(∙):
-
The sigmoid activation function
- φ(∙):
-
The tanh function.
- g(t):
-
The signal to be decomposed
- \({\hat{T}}_n\left(\omega \right)\) :
-
The scaling function
- \({\hat{P}}_n\left(\omega \right)\) :
-
The wavelet function
- β(x):
-
An arbitrary function that satisfiesCk([0, 1])
- \({K}_g^{\varepsilon}\left(n,t\right)\) :
-
The detail coefficient
- \({K}_g^{\varepsilon}\left(0,t\right)\) :
-
The approximate coefficient
- \({\bar{P}}_n\) :
-
The complex conjugate of Pn
- \({\bar{T}}_1\) :
-
The complex conjugate of T1
References
Hong T, J. I. J. o. S, Fan F (2016) Probabilistic electric load forecasting: A tutorial review. Int J Forecast 32(3):914–938
Bessani M, Massignan J, Santos T et al (2020) Multiple households very short-term load forecasting using bayesian networks. Electr Power Syst Res 189:106733
Sun JX, Wang JN, Yu WX et al (2020) Power load disaggregation of households with solar panels based on an improved long short-term memory network. J Electr Eng Technol 15(5):2401–2413
Dudek G, Peka P (2021) Pattern similarity-based machine learning methods for mid-term load forecasting: a comparative study. Appl Soft Comput 104(2):107223
Yin L, Xie J (2021) Multi-temporal-spatial-scale temporal convolution network for short-term load forecasting of power systems. Appl Energy 283(6):116328
Raza MQ, Mithulananthan N, Li J, Lee KY (2020) Multivariate ensemble forecast framework for demand prediction of anomalous days. IEEE Trans Sustain Energy 11(1):27–36
Han L, Peng Y, Li Y et al (2018) Enhanced deep networks for short-term and medium-term load forecasting. IEEE Access 7:4045–4055
Tang X, Dai Y, Wang T et al (2019) Short-term power load forecasting based on multi-layer bidirectional recurrent neural network. IET Gener Transm Distrib 13(17):3847–3854
Khwaja AS, Zhang X, Anpalagan A et al (2017) Boosted neural networks for improved short-term electric load forecasting. Electr Power Syst Res 143:431–437
Dosiek L (2020) The effects of forced oscillation frequency estimation error on the LS-ARMAS mode meter. IEEE Trans Power Syst 35(2):1650–1652
Moon J, Hossain MB, Chon KH (2021) AR and ARMA model order selection for time-series modeling with ImageNet classification. Sig Process 183(10):108026
Ertuğrul ÖF, Tekin H, Tekin R (2021) A novel regression method in forecasting short-term grid electricity load in buildings that were connected to the smart grid. Electr Eng 103:717–728
Xu W, Peng H, Zeng X et al (2019) A hybrid modeling method for time series forecasting based on a linear regression model and deep learning. Appl Intell 49:3002–3015
Munawar U, Wang Z (2020) A framework of using machine learning approaches for short-term solar power forecasting. J Electr Eng Technol 15(2):561–569
Xu C, Gordan B, Koopialipoor M et al (2019) Improving performance of retaining walls under dynamic conditions developing an optimized ANN based on ant colony optimization technique. IEEE Access 7:94692–94700
Elattar EE, Sabiha NA, Alsharef M et al (2020) Short term electric load forecasting using hybrid algorithm for smart cities. Appl Intell 50:3379–3399
Dedinec A, Filiposka S, Dedinec A et al (2016) Deep belief network based electricity load forecasting: An analysis of Macedonian case. Energy 115:1688–1700
Chen CLP, Liu Z (2018) Broad learning system: an effective and efficient incremental learning system without the need for deep architecture. IEEE Trans Neural Netw Learn Syst 29(1):10–24
Zhao F, Zeng GQ, Lu KD (2019) EnLSTM-WPEO: short-term traffic flow prediction by ensemble LSTM, NNCT weight integration and population extremal optimization. IEEE Trans Veh Technol 99:1–1
Le T, Vo B, Fujita H et al (2019) A fast and accurate approach for bankruptcy forecasting using squared logistics loss with GPU-based extreme gradient boosting. Inf Sci 494:294–310
Gang Shi C, Qin J, Tao C, Liu (2021) A VMD-EWT-LSTM-based multi-step prediction approach for shield tunneling machine cutterhead torque. Knowl-Based Syst 228:107213
Zhu L, Lian C (2019) Wind Speed forecasting based on a hybrid EMD-BLS method. 2019 Chinese Automation Congress (CAC), pp 2191–2195
Tan M, Yuan S, Li S et al (2020) Ultra-short-term industrial power demand forecasting using LSTM based hybrid ensemble learning. IEEE Trans Power Syst 35(4):2937–2948
Yan K, Li W, Ji Z et al (2019) A hybrid LSTM neural network for energy consumption forecasting of individual households. IEEE Access 7:1–1
Chen Y, Luh PB, Guan C et al (2010) Short-Term load forecasting: similar day-based wavelet neural networks. IEEE Trans Power Syst 25(1):322–330
Yang Y, Li W, Gulliver TA et al (2019) Bayesian deep learning based probabilistic load forecasting in smart grids. IEEE Trans Industr Inf 99:1–1
Ospina J, Newaz A, Faruque MO (2019) Forecasting of PV plant output using hybrid wavelet-based LSTM-DNN structure model. IET Renew Power Gener 13(7):1087–1095
Sheng Z, Wang H, Chen G et al (2021) Convolutional residual network to short-term load forecasting. Appl Intell 51:2485–2499
He Y, Yang Q, Wang S et al (2019) Electricity consumption probability density forecasting method based on LASSO-Quantile Regression Neural Network. Appl Energy 233–234:565–575
He Y, Li H, Wang S et al (2020) Uncertainty analysis of wind power probability density forecasting based on cubic spline interpolation and support vector quantile regression. Neurocomputing 430:121–137
Ye F, Zhang L, Zhang D et al (2016) A novel forecasting method based on multi-order fuzzy time series and technical analysis. Inf Sci 367–368:41–57
Md M, Alam S, Rehman LM, Al-Hadhrami JP, Meyer (2014) Extraction of the inherent nature of wind speed using wavelets and FFT. Energy Sustain Dev 22:34–47
Ujjwal Kumar K, De Ridder (2010) GARCH modelling in association with FFT–ARIMA to forecast ozone episodes. Atmos Environ 44(34):4252–4265
Huang NE, Shen Z, Long SR et al (1998) The empirical mode decomposition method and the Hilbert spectrum for non-stationary time series analysis. Proc R Soc A: Math Phys Eng Sci 454:903–995
Zhang X, Wang J (2018) A novel decomposition-ensemble model for forecasting short‐term load‐time series with multiple seasonal patterns. Appl Soft Comput 65:478–494
Moreno SR, Mariani VC, dos Santos Coelho L (2021) Hybrid multi-stage decomposition with parametric model applied to wind speed forecasting in Brazilian Northeast. Renew Energy 164:1508–1526
Moreno SR, Gomes Ramon, da Silva Viviana, Mariani Cocco, dos Santos Coelho Leandro (2020) Multi-step wind speed forecasting based on hybrid multi-stage decomposition model and long short-term memory neural network. Energy Convers Manag 213:112869
He F, Zhou J, Mo L, Feng K, Liu G (2020) Day-ahead short-term load probability density forecasting method with a decomposition-based quantile regression forest. Appl Energy 262:114396
Jatin Bedi D (2020) Energy load time-series forecast using decomposition and autoencoder integrated memory network. Appl Soft Comput 93:106390
Gilles J (2013) Empirical wavelet transform. IEEE Trans Signal Process 61(16):3999–4010
Salkuti SR (2018) Short-term electrical load forecasting using hybrid ANN–DE and wavelet transforms approach. Electr Eng 100:2755–2763
Qiu X, Ren Y, Suganthan PN et al (2017) Empirical Mode Decomposition based ensemble deep learning for load demand time series forecasting. Appl Soft Comput 54:246–255
Qiu X, Suganthan PN, Amaratunga GAJ (2018) Ensemble incremental learning Random Vector Functional Link network for short-term electric load forecasting. Knowl-Based Syst 145:182–196
Li Y, Wu H, Liu H (2018) Multi-step wind speed forecasting using EWT decomposition, LSTM principal computing, RELM subordinate computing and IEWT reconstruction. Energy Convers Manag 167:203–219
He Y, Wang Y (2021) Short-term wind power prediction based on EEMD-LASSO-QRNN model. Appl Soft Comput 105:107288
da Silva RG, Dal Molin Ribeiro MH et al (2021) A novel decomposition-ensemble learning framework for multi-step ahead wind energy forecasting. Energy 216:119174
Jaseena KU, Binsu C, Kovoor (2021) Decomposition-based hybrid wind speed forecasting model using deep bidirectional LSTM networks. Energy Convers Manag 234:113944
Shao Z, Fu C, Yang SL et al (2017) A review of the decomposition methodology for extracting and identifying the fluctuation characteristics in electricity demand forecasting. Renew Sustain Energy Rev 75:123–136
Pincus SM (1911) Approximate entropy as a measure of system complexity. Proc Natl Acad Sci USA 88(6):2297–2301
Daubechies I, Heil C (1992) TTen lectures on wavelets. Cbms-nsf Regional Conference Series in Applied Mathematics: Society for Industrial & Applied Mathematics
Friedman M (1939) The use of ranks to avoid the assumption of normality implicit in the analysis of variance. Publ Am Stat Assoc 32(200):675–701
Friedman M (1940) A comparison of alternative tests of significance for the problem of m rankings. Ann Math Stat 11:86–92
Sheskin DJ (2000) Handbook of parametric and nonparametric statistical procedures. Chapman & Hall, London
Acknowledgements
This work was supported in part by the National Natural Science Foundation of China under Grants 61863028, 81660299, and 61503177, and in part by the Science and Technology Department of Jiangxi Province of China under Grants 20161ACB21007, 20171BBE50071, and 20171BAB202033.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Liao, Z., Huang, J., Cheng, Y. et al. A novel decomposition-based ensemble model for short-term load forecasting using hybrid artificial neural networks. Appl Intell 52, 11043–11057 (2022). https://doi.org/10.1007/s10489-021-02864-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-021-02864-8