Hybrid Hourly Solar Energy Forecasting Using BiLSTM Networks with Attention Mechanism, General Type-2 Fuzzy Logic Approach: A Comparative Study of Seasonal Variability in Lithuania
Source abstract
Abstract
This research introduces a novel hybrid forecasting framework for solar energy prediction in high-latitude regions with extreme seasonal variations. This approach uniquely employs General Type-2 Fuzzy Logic (GT2-FL) for data preprocessing and uncertainty handling, followed by two advanced neural architectures, including BiLSTM and SCINet with Time2Vec encoding and Variational Mode Decomposition (VMD) signal processing. Four configurations are systematically evaluated: BiLSTM-Time2Vec, BiLSTM-VMD, SCINet-Time2Vec, and SCINet-VMD, each tested with GT2-FL preprocessed data and raw input data. Using meteorological data from Lithuania (2023–2024) with extreme seasonal variations where daylight hours range from 17 h in summer to 7 h in winter, F-BiLSTM-Time2Vec achieved exceptional performance, with nRMSE = 1.188%, NMAE = 0.813%, and WMAE = 3.013%, significantly outperforming both VMD-based variants and SCINet architectures. Comparative analysis revealed that Time2Vec encoding proved more beneficial than VMD preprocessing, especially when enhanced with fuzzification. The results confirm that fuzzification, BiLSTM architecture, and Time2Vec encoding provide the most robust forecasting capability under various seasonal conditions.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Solar Radiation and Photovoltaics · Photovoltaic System Optimization Techniques · Energy Load and Power Forecasting
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Smart Grids and Renewable Energy Laboratory, Lithuanian Energy Institute, Breslaujos St. 3, 44403 Kaunas, Lithuania
Funding
Research Council of Lithuania
References (61)
Show source-provided references
1. The role of environmental technology for energy demand and energy efficiency: Evidence from OECD countries10.1016/j.rser.2021.111735 ↗
2. Contribution of renewable energy (hydro, wind, solar and biomass) to decarbonization and transformation of the electricity generation sector for sustainable development10.1016/j.egyr.2023.07.024 ↗
3. A critical evaluation of grid stability and codes, energy storage and smart loads in power systems with wind generation10.1016/j.energy.2020.117671 ↗
4. An analysis of additional energy requirement to meet the sustainable development goals10.1016/j.jclepro.2020.122646 ↗
5. Eurostat (2025, July 04). Renewable Energy Statistics. Available online: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Renewable_energy_statistics.
6. Sharma, N., Sharma, P., Irwin, D., and Shenoy, P. (2011, January 17–20). Predicting solar generation from weather forecasts using machine learning. Proceedings of the IEEE International Conference on Smart Grid Communications (SmartGridComm), Brussels, Belgium.10.1109/SmartGridComm.2011.6102379 ↗
7. Integrating solar and wind energy into the electricity grid for improved power accessibility
8. A hybrid CSP–CPV system for improving the dispatchability of solar power plants10.1016/j.enconman.2016.02.015 ↗
9. Predicting solar energy generation through artificial neural networks using weather forecasts for microgrid control10.1016/j.renene.2018.03.070 ↗
10. Moser Baer Solar (2025, July 04). Solar PV Monitoring: Maximizing Performance Through Real-Time Analytics. Available online: https://www.moserbaersolar.com/maintenance-and-performance-optimization/solar-pv-monitoring-maximizing-performance-through-real-time-analytics/.
11. Rodriguez, J., Pontt, J., Correa, P., Lezana, P., and Cortes, P. (2005, January 2–6). Predictive power control of an AC/DC/AC converter. Proceedings of the 40th IAS Annual Meeting Conference Record of the Industry Applications Conference, Hong Kong, China.10.1109/IAS.2005.1518458 ↗
12. Narkutė, V. (2022). Forecast of Energy Produced by Solar Power Plants. [Ph.D. Thesis, Vilniaus Universitetas].
13. Mellit, A., Massi Pavan, A., Ogliari, E., Leva, S., and Lughi, V. (2020). Advanced methods for photovoltaic output power forecasting: A review. Appl. Sci., 10.10.3390/app10020487 ↗
14. Solar photovoltaic generation forecasting methods: A review10.1016/j.enconman.2017.11.019 ↗
15. Forecasting electric power generation in a photovoltaic power system for an energy network10.1002/eej.20755 ↗
16. Fuzzy algorithm for estimation of solar irradiation from sunshine duration10.1016/S0038-092X(98)00043-7 ↗
17. A model for the estimation of global solar radiation using fuzzy random variables10.1175/1520-0450(2002)041<1267:AMFTEO>2.0.CO;2 ↗
18. Fuzzy logic-based modeling and estimation of global solar energy using meteorological parameters10.1016/j.energy.2014.04.057 ↗
19. Evaluating the campus climate factors using an interval type-2 fuzzy ANP
20. Type 2-Fuzzy sets made simple10.1109/91.995115 ↗
21. A fast learning algorithm for deep belief nets10.1162/neco.2006.18.7.1527 ↗
22. Jlidi, M., Hamidi, F., Barambones, O., Abbassi, R., Jerbi, H., Aoun, M., and Karami-Mollaee, A. (2023). An artificial neural network for solar energy prediction and control using Jaya-SMC. Electronics, 12.10.3390/electronics12030592 ↗
23. A hybrid model for financial time-series forecasting based on mixed methodologies10.1111/exsy.12633 ↗
24. A study of deep learning-based multi-horizon building energy forecasting10.1016/j.enbuild.2023.113810 ↗
25. Bai, S., Kolter, J.Z., and Koltun, V. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv.
26. A TCN-based hybrid forecasting framework for hours-ahead utility-scale PV forecasting10.1109/TSG.2023.3236992 ↗
27. Predicting solar radiation for space heating with thermal storage system based on temporal convolutional network-attention model10.1016/j.applthermaleng.2022.119574 ↗
28. Short-term load forecasting based on improved TCN and DenseNet10.1109/ACCESS.2022.3218374 ↗
29. Accurate one step and multistep forecasting of very short-term PV power using LSTM-TCN model. Renew
30. WOA-VMD-SCINet: Hybrid model for accurate prediction of ultra-short-term Photovoltaic generation power considering seasonal variations10.1016/j.egyr.2024.09.025 ↗
31. Time series forecasting for hourly photovoltaic power using conditional generative adversarial network and Bi-LSTM10.1016/j.energy.2022.123403 ↗
32. A novel integrated photovoltaic power forecasting model based on variational mode decomposition and CNN-BiGRU considering meteorological variables10.1016/j.epsr.2022.108796 ↗
33. Short-term PV power forecast using hybrid deep learning model and Variational Mode Decomposition10.1016/j.egyr.2023.05.154 ↗
34. A hybrid deep learning model with an optimal strategy based on improved VMD and transformer for short-term photovoltaic power forecasting10.1016/j.energy.2024.131071 ↗
35. Prediction of wind and PV power by fusing the multi-stage feature extraction and a PSO-BiLSTM model10.1016/j.energy.2024.131345 ↗
36. Machine learning methods for solar radiation forecasting: A review10.1016/j.renene.2016.12.095 ↗
37. Solar power forecasting: A review
38. Multiple steps ahead solar photovoltaic power forecasting based on univariate machine learning models and data re-sampling10.1016/j.segan.2019.100286 ↗
39. A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network10.1016/j.apenergy.2019.113315 ↗
40. Day-ahead power forecasting in a large-scale photovoltaic plant based on weather classification using LSTM10.1016/j.energy.2019.07.168 ↗
41. Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM10.1016/j.energy.2018.01.177 ↗
42. Forecasting solar energy production in Spain: A comparison of univariate and multivariate models at the national level10.1016/j.apenergy.2023.121645 ↗
43. Hourly forecasting of solar irradiance based on CEEMDAN and multi-strategy CNN-LSTM neural networks10.1016/j.renene.2020.09.141 ↗
44. Scinet: Time series modeling and forecasting with sample convolution and interaction
45. Long term rolling prediction model for solar radiation combining empirical mode decomposition (EMD) and artificial neural network (ANN) techniques10.1063/1.4999240 ↗
46. Evaluation of investment opportunities with interval-valued fuzzy TOPSIS method10.2478/amns.2020.1.00044 ↗
47. A review on Type 2-Fuzzy logic applications in clustering, classification and pattern recognition10.1016/j.asoc.2014.04.017 ↗
48. A comprehensive review on type 2 fuzzy logic applications: Past, present and future10.1016/j.engappai.2020.103916 ↗
49. A recurrent self-evolving interval Type 2-Fuzzy neural network for dynamic system processing10.1109/TFUZZ.2009.2021953 ↗
50. A hybrid photovoltaic/wind power prediction model based on Time2Vec, WDCNN and BiLSTM10.1016/j.enconman.2023.117342 ↗
51. Two-dimensional compact variational mode decomposition: Spatially compact and spectrally sparse image decomposition and segmentation10.1007/s10851-017-0710-z ↗
52. Guermoui, M., Fezzani, A., Mohamed, Z., Rabehi, A., Ferkous, K., Bailek, N., and Ghoneim, S.S.M. (2024). An analysis of case studies for advancing photovoltaic power forecasting through multi-scale fusion techniques. Sci. Rep., 14.10.1038/s41598-024-57398-z ↗
53. Increasing the sine-cosine transformation accuracy in signal approximation and interpolation10.3103/S8756699008030047 ↗
54. A review on the attention mechanism of deep learning10.1016/j.neucom.2021.03.091 ↗
55. Short-term load forecasting based on LSTM networks considering attention mechanism10.1016/j.ijepes.2021.107818 ↗
56. Adaptability of financial time series prediction based on BiLSTM10.1016/j.procs.2022.01.003 ↗
57. VMD-SCINet: A hybrid model for improved wind speed forecasting10.1007/s12145-023-01169-3 ↗
58. Dong, P., Zuo, Y., Chen, L., Li, S., and Zhang, L. (2023, January 21–23). VSNT: Adaptive Network Traffic Forecasting via VMD and Deep Learning with SCI-Block and Attention Mechanism. Proceedings of the 2023 IEEE International Conference on Parallel and Distributed Processing with Applications, Big Data and Cloud Computing, Sustainable Computing and Communications, Social Computing and Networks (ISPA/BDCloud/SocialCom/SustainCom), Haikou, China.10.1109/ISPA-BDCloud-SocialCom-SustainCom59178.2023.00046 ↗
59. Analysis and validation of 24 hours ahead neural network forecasting of photovoltaic output power10.1016/j.matcom.2015.05.010 ↗
60. Selection of Unit Root Test on the Basis of Length of the Time Series and Value of AR (1) Parameter
61. A proposed novel adaptive DC technique for non-stationary data removal10.1016/j.heliyon.2023.e13903 ↗