Solar and Wind 24 H Sequenced Prediction Using L-Transform Component and Deep LSTM Learning in Representation of Spatial Pattern Correlation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please refer to the review document for details

Comments for author File: Comments.pdf

Author Response

Thank you very much for your comments. Please kindly refer to the attached documents. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper provides solar and wind predictions using evolution differentiated modular and LSTM long-term modelling based on pattern correlations. The content of the paper is very substantial, but there are still some issues that need to be addressed. My detailed comments are listed as follows:

1. Regarding the solar and wind prediction results in this paper, in what specific meteorological conditions (e.g., extreme weather events, highly variable irradiance/wind speed scenarios) does the proposed method demonstrate superior accuracy compared to other intelligent methods? Please elaborate with quantitative evidence and scenario-specific analysis.
2. The resolution of some figures in the manuscript should be furtherly improved.
3. The conclusion needs to be refined, and the main contributions and conclusions are drawn.
4. The analysis relies exclusively on data from 2015. Given that wind and solar predictions typically require long-term datasets to account for interannual variability and climatic cycles, please justify the omission of multi-year data.

Author Response

Thank you very much for your comments. Please kindly refer to the attached documents.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The paper proposes a novel hybrid approach combining Differential Learning (DfL) and Long Short-Term Memory (LSTM) deep learning for mid- to long-term solar and wind energy forecasting, demonstrating practical significance. However, several limitations should be addressed to enhance the robustness and generalizability of the findings:

1 The study relies exclusively on data from a specific geographic region (Tajoura, Libya) and a short time frame (November 2015), lacking diversity in climatic conditions and seasonal variations. This restricts the model’s generalizability and raises questions about its applicability to broader scenarios. The study appears more like a localized case study rather than a universally validated solution. It is suggested to expand the dataset to include multiple regions and extended time periods (e.g., multi-year data across different seasons) to ensure the model’s adaptability to diverse environmental conditions.

2 The model’s performance under extreme weather events (e.g., storms, sudden meteorological shifts) remains unverified. Additionally, its adaptability to long-term climate variability is not discussed.  It is suggested to incorporate extreme weather case studies and evaluate the model’s resilience. Further, explore adaptive strategies (e.g., dynamic retraining) to enhance its response to climate change-induced variability.

3 The experimental validation is based on only a 10-day testing window, which may not provide statistically significant results or reflect long-term forecasting stability. It is suggested to extend the evaluation period to several months or across different seasons to assess the model’s consistency and reliability over time.

4 While the methodological innovation is promising, the paper does not sufficiently discuss real-world deployment potential. For instance, how could this model integrate into microgrid energy management or storage optimization? Could it be combined with numerical weather prediction models for enhanced accuracy? Please provide concrete case studies in DISCUSSION to demonstrate practical utility. 

Author Response

Thank you very much for your comments. Please kindly refer to the attached documents.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The paper is an interesting study of the application of hybrid AI methods (in particular DfL and LSTM) to medium-term solar and wind power forecasting. I would like to highlight: The use of proprietary differential neural networks (D-PNN) and deep recurrent networks makes this paper stand out from the others. However, the paper suffers from being overloaded, difficult to read, and has stylistic shortcomings.
Content Notes:
1. There is no comparison with other modern AI methods (e.g. GRU, Transformer, etc.).
2. Only one dataset (2015) is used, which limits the generalizability of the findings.
3. The results are presented dryly, without a critical discussion of the biases.
4. There is no deeper discussion of the limitations of the models (e.g. overfitting, robustness to noise).

Comments on the Quality of English Language

Design notes:
1. The text is overloaded with formulas and technical details without sufficient explanatory base.
2. Too much "noisy" terminology: phrases like "evolutionary produced optimal PDE components related to the pattern complexity" are hard to read.
3. The article suffers from excessively long paragraphs and "dumping" ideas in one place.
4. Some figures (e.g. Figures 4 and 5) are not cited in the text.
5. Lack of articles: "Modelling boundary conditions is vital" → "The modelling of boundary conditions is vital".
6. Sentence structure: many sentences are excessively long and overloaded with phrases, for example: "Differential learning (DfL) is an unconventional recently developed biologically inspired strategy..." Better: "Differential learning (DfL) is a recently developed, biologically inspired strategy..."
7. Commas before subordinate clauses are completely missing:
"The prediction models were finally tested in the latest observation times in processing input..."
better: "...times, in processing input..."
8. Phrases like "are inevitable in the evolution of robust models" sound cumbersome and can be replaced with clearer ones: "are crucial for developing robust models".
9. The use of "chaotic fluctuation", "break anomalies", "overbreak situations" is not always clear - clarification or reformulation is required.
10. Graphs: The graphs lack units of measurement and explanations (for example, it is not immediately obvious what "DLT" means).

Conclusion. The article is interesting, but the formatting is terrible. Extensive linguistic and stylistic revisions are required.

Author Response

Thank you very much for your comments. Please kindly refer to the attached documents.

Author Response File: Author Response.pdf

Reviewer 5 Report

Comments and Suggestions for Authors

The revised manuscript demonstrates significant improvement in clarity and methodological articulation. Your integration of differential learning with deep learning offers a compelling framework for forecasting renewable energy potential under complex spatiotemporal dynamics. The inclusion of executable software for model validation is a strong contribution to reproducibility and applied research. The conceptual depth and practical orientation of the study make it a valuable addition to the field. I believe the manuscript is now ready for publication.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

My concerns have been satisfactorily addressed.

Reviewer 4 Report

Comments and Suggestions for Authors

no