Time series forecasting for hourly photovoltaic power using conditional generative adversarial network and Bi-LSTM

Huang, Xiaoqiao; Li, Qiong; Tai, Yonghang; Chen, Zaiqing; Liu, Jun; Shi, Junsheng; Liu, Wu-Ming

doi:10.1016/j.energy.2022.123403

Cited by 103 publications

(32 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The modern photovoltaic power generation forecasting methods start to use artificial intelligence algorithms such as support vector machines [12][13][14], random forests [15,16] and other machine learning algorithms [17], convolutional neural networks [18,19], long short-term memory artificial neural networks [20][21][22] and other deep learning algorithms [23].…”

Section: Modern Photovoltaic Power Generation Forecasting Methodsmentioning

confidence: 99%

Photovoltaic Power Generation Forecasting Based on the ARIMA-BPNN-SVR Model

Fan¹,

Hui-zhen²,

Chen³

et al. 2022

Glob. J. Energ. Technol. Res. Updat.

View full text Add to dashboard Cite

With the continuous expansion of the capacity of photovoltaic power generation systems, accurate power generation load forecasting can make grid dispatching more reasonable and optimize load distribution. This paper proposes a combined forecasting model based on Auto Regression Integrate Moving Average (ARIMA), back propagation neural network (BPNN), and support vector regression (SVR), namely ARIMA-BPNN-SVR model, aiming at the problem of low accuracy of a single model and traditional forecasting model. Through the complementary advantages of ARIMA, BPNN, and SVR models, the model has good anti-noise ability, nonlinear mapping, and adaptive ability when processing photovoltaic power generation data. Data experiments are carried out on solar photovoltaic power generation in the United States, and the accuracy of model forecasting is evaluated according to MAE, MSE, RMSE, and MAPE. The experimental results show that the proposed ARIMA-BPNN-SVR outperforms the forecasting performance of the single models ARIMA, BPNN, and SVR. Its MAE, MSE, RMSE and MAPE are 0.53, 0.41, 0.64 and 0.84 respectively. In the Wilcoxon sign-rank test, the p-value of the proposed model reached 0.98, indicating the effectiveness of the ARIMA-BPNN-SVR model.

show abstract

Section: Modern Photovoltaic Power Generation Forecasting Methodsmentioning

confidence: 99%

Photovoltaic Power Generation Forecasting Based on the ARIMA-BPNN-SVR Model

Fan¹,

Hui-zhen²,

Chen³

et al. 2022

Glob. J. Energ. Technol. Res. Updat.

View full text Add to dashboard Cite

show abstract

“…Due to meet the demand of power and maintain a balance between the supply and demand, always prediction process is carried out for the constructed solar farms so as to have a complete analysis on solar output power production and supply to the end users. Under this scenario, machine learning (ML) models are widely employed as black box models for performing the forecast mechanism of the solar PV output power [5][6][7][8][9][10][11][12][13] and this section of this research paper presents a detailed survey on different techniques and ML models applied over the years for predicting the PV output power.…”

Section: Plos Onementioning

confidence: 99%

“…In view of the literature study made on the related works as above in the prediction of solar PV output power, it is lucid that several researchers has developed and analysed the machine learning based predictor models for the said application. Among the machine learning models, few feed forward models and their variants, recurrent neural predictors and memory based models has been widely used [11][12][13][14][15][16][17][18]. Also, with the growth of deep learning based techniques, researchers has initiated in developing predictor models for solar PV output power forecasting using various deep learning models for the said application [1,14,19,20,26,27,43,47].…”

Section: Challengesmentioning

confidence: 99%

“…• Occurrences of global minima and stagnation issues [3][4][5][6][7] • Scalability problems on the normalization procedures adopted [2,8,[12][13][14][15][16][17] • Over-fitting and under-fitting issues [5, 6, 9-11, 23, 48, 51] • Dimensionality constraints of the solar farm data and data handling issues [18][19][20][21][22][23][24] • Elapsed training time [29,31,37] • Data extraction problems in regression based ML models [10][11][12][13][14][15] • Higher number of trainable parameters in DL models [1, 14, 19-20, 26, 27, 43, 47] • Repetitive training of deep neural networks [19,20,26,27] • High computational overhead due to repetitive process [29][30][31][32][33][34][35][36] • Few predictor models with high complexity and data redundancy [45][46][47]…”

Section: Challengesmentioning

confidence: 99%

See 1 more Smart Citation

Variational mode decomposition combined fuzzy—Twin support vector machine model with deep learning for solar photovoltaic power forecasting

et al. 2022

View full text Add to dashboard Cite

A novel Variational Mode Decomposition (VMD) combined Fuzzy-Twin Support Vector Machine Model with deep learning mechanism is devised in this research study to forecast the solar Photovoltaic (PV) output power in day ahead basis. The raw data from the solar PV farms are highly fluctuating and to extract the useful stable components VMD is employed. A novel Fuzzy–Twin Support Vector Machine (FTSVM) model developed acts as the forecasting model for predicting the solar PV output power for the considered solar farms. The twin support vector machine (SVM) model formulates two separating hyperplanes for predicting the output power and in this research study a fuzzy based membership function identifies most suitable two SVM prediction hyperplanes handling the uncertainties of solar farm data. For the developed, new VMD-FTSVM prediction technique, their optimal parameters for the training process are evaluated with the classic Ant Lion Optimizer (ALO) algorithm. The solar PV output power is predicted using the novel VMD-FTSVM model and during the process multi-kernel functions are utilized to devise the two fuzzy based hyperplanes that accurately performs the prediction operation. Deep learning (DL) based training of the FTSVM model is adopted so that the deep auto-encoder and decoder module enhances the accuracy rate. The proposed combined forecasting model, VMD-ALO-DLFTSVM is validated for superiority based on a two 250MW PV solar farm in India. Results prove that the proposed model outperforms the existing model in terms of the performance metrics evaluated and the forecasted PV Power.

show abstract

“…These can impact the regular operation of the grid inverter and introduce considerable problems to large‐scale PV power grid connections (Wang, Zhang, et al, 2019). Therefore, accurate PV power forecasts can provide effective prior information to support local planning and thereby reduce the severe “abandonment” phenomenon (Huang et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Photovoltaic power combination prediction system based on improved multi‐objective optimization algorithm and nonlinear weighting strategy

2022

View full text Add to dashboard Cite

The large-scale connection of photovoltaic (PV) power generation to the power grid introduces considerable challenges to the grid's stable operation. To ensure grid stability, the key is to improve the accuracy of PV power prediction. Currently, PV power prediction is primarily dependent on single-model construction, which ignores the role of data processing technologies, thus resulting in inadequate prediction performance. Accordingly, in this study, a combined prediction system is developed by integrating advanced deep learning algorithms and data processing techniques to improve the accuracy of PV power prediction. In addition, a nonlinear weighting strategy based on an improved multi-objective dragonfly optimization algorithm (IMODA) is proposed to determine the final prediction result. In the IMODA, cloud model generator, tent mapping, and Pareto solution selection strategy based on knee points are introduced to resolve the defects of the original algorithm. The performance of the proposed combined system was scientifically evaluated and analysed by considering the PV power datasets of four seasons in Belgium as an example. The mean absolute percentage errors of the proposed system on the four datasets are 4. 0198, 4.7943, 4.3587, and 5.9286. These values indicate an improvement range of 5%-30% compared with the errors of other benchmark models.

show abstract

Time series forecasting for hourly photovoltaic power using conditional generative adversarial network and Bi-LSTM

Cited by 103 publications

References 35 publications

Photovoltaic Power Generation Forecasting Based on the ARIMA-BPNN-SVR Model

Photovoltaic Power Generation Forecasting Based on the ARIMA-BPNN-SVR Model

Variational mode decomposition combined fuzzy—Twin support vector machine model with deep learning for solar photovoltaic power forecasting

Photovoltaic power combination prediction system based on improved multi‐objective optimization algorithm and nonlinear weighting strategy

Contact Info

Product

Resources

About