Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Aug 24;12(1):14467.
doi: 10.1038/s41598-022-18820-6.

Deep belief rule based photovoltaic power forecasting method with interpretability

Peng Han  1 , Wei He  2   3 , You Cao  4 , YingMei Li  5 , YunYi Zhang  1
Affiliations

Deep belief rule based photovoltaic power forecasting method with interpretability

Peng Han et al. Sci Rep. .

Abstract

Accurate prediction of photovoltaic (PV) output power is of great significance for reasonable scheduling and development management of power grids. In PV power generation prediction system, there are two problems: the uncertainty of PV power generation and the inexplicability of the prediction result. The belief rule base (BRB) is a rule-based modeling method and can deal with uncertain information. Moreover, the modeling process of BRB has a certain degree of interpretability. However, rule explosion and the inexplicability of the optimized model limit the modeling ability of BRB in complex systems. Thus, a PV output power prediction model is proposed based on a deep belief rule base with interpretability (DBRB-I). In the DBRB-I model, the deep BRB structure is constructed to solve the rule explosion problem, and inefficient rules are simplified by a sensitivity analysis of the rules, which reduces the complexity of the model. Moreover, to ensure that the interpretability of the model is not destroyed, a new optimization method based on the projection covariance matrix adaptation evolution strategy (P-CMA-ES) algorithm is designed. Finally, a case study of the prediction of PV output power is conducted to illustrate the effectiveness of the proposed method.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
The DBRB-I model based on the PV power generation system.
Figure 2
Figure 2
The interpretability of the DBRB-I model.
Figure 3
Figure 3
Reasonable and unreasonable belief distributions.
Figure 4
Figure 4
The feasible region of DBRB-I model optimization.
Figure 5
Figure 5
The optimization process of the modified P-CMA-ES.
Figure 6
Figure 6
Data observation.
Figure 7
Figure 7
The relationship between each attribute and output power.
Figure 8
Figure 8
Initial PV output power prediction model based on DBRB-I.
Figure 9
Figure 9
The sensitivity analysis for Sub-BRB1 (x-axis represents rule weights, y-axis represents MSE).
Figure 10
Figure 10
The sensitivity analysis for Sub-BRB2 (x-axis represents rule weights, y-axis represents MSE).
Figure 11
Figure 11
The sensitivity analysis for Sub-BRB3 (x-axis represents rule weights, y-axis represents MSE).
Figure 12
Figure 12
Analysis of Sub-BRB1 rule activation weights (x-axis represents test data, y-axis represents activation weights).
Figure 13
Figure 13
Analysis of Sub-BRB2 rule activation weights (x-axis represents test data, y-axis represents activation weights).
Figure 14
Figure 14
Analysis of Sub-BRB3 rule activation weights (x-axis represents test data, y-axis represents activation weights).
Figure 15
Figure 15
The sensitivity analysis of trained Sub-BRB1 (x-axis represents rule weights, y-axis represents MSE).
Figure 16
Figure 16
The sensitivity analysis of trained Sub-BRB2 (x-axis represents rule weights, y-axis represents MSE).
Figure 17
Figure 17
The sensitivity analysis of trained Sub-BRB3 (x-axis represents rule weights, y-axis represents MSE).
Figure 18
Figure 18
Transparency and validity of the DBRB-I(20) model.
Figure 19
Figure 19
The belief distribution of Sub-BRB3.
Figure 20
Figure 20
Prediction results of the DBRB-I(20) model.
Figure 21
Figure 21
Prediction results of the DBRB-I(41) model.
Figure 22
Figure 22
Comparison of DBRB(41) and DBRB(48).
Figure 23
Figure 23
Interpretability analysis of Sub-BRB3.
Figure 24
Figure 24
The fitness value of the DBRB model over 400 generations.
Figure 25
Figure 25
Comparison of the number of effective rules of the DBRB-I(20) model.
Figure 26
Figure 26
Compare predictions from different models.
Figure 27
Figure 27
Sensitivity analysis of Sub-BRB3 of the DBRB-I(20) model (x-axis represents rule weights, y-axis represents MSE).
Figure 28
Figure 28
Belief distribution of each rule of DBRB-I(20).
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Wang KJ, Qi XX, Liu HD. Photovoltaic power forecasting based LSTM- convolutional network. Energy. 2020;189:116225.
    1. Antonanzas J, et al. Review of photovoltaic power forecasting. Sol. Energy. 2016;136:78–111.
    1. Eseye AT, Zhang JH, Zheng DH. Short-term photovoltaic solar power forecasting using a hybrid Wavelet-PSO-SVM model based on SCADA and Meteorological information. Renewable Energy. 2017;118:357–367.
    1. Halabi LM, Mekhilef S, Hossain M. Performance evaluation of hybrid adaptive neuro-fuzzy inference system models for predicting monthly global solar radiation. Appl. Energy. 2018;213:47–261.
    1. Chen XY, Du Y, Lim EG, Wen HQ, Jiang L. Sensor network based PV power nowcasting with spatio-temporal preselection for grid-friendly control. Appl. Energy. 2019;255:113760.