Dynamic Adaptive Artificial Hummingbird Algorithm-Enhanced Deep Learning Framework for Accurate Transmission Line Temperature Prediction

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

The authors re-addressed all issues.

Author Response

3. Point-by-point response to Comments and Suggestions for Authors

4. Response to Comments on the Quality of English Language

5. 其他说明

我们衷心感谢审稿人对我们手稿的彻底审查和宝贵的反馈。他们的建设性意见显著提高了我们工作的清晰度和严谨性

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

Revision looks good for me.

Further revision is required; for instance, in Table 5, "ARIMA," includes an unnecessary extra ",".

Author Response

3. Point-by-point response to Comments and Suggestions for Authors

Comments 1: Further revision is required; for instance, in Table 5, "ARIMA," includes an unnecessary extra ",". Response 1: Thank you for pointing this out.  We agree with this comment.  Therefore, we have corrected the extra comma in Table 5. 4. Response to Comments on the Quality of English Language

5. Additional clarifications

We sincerely thank the reviewers for their thorough examination of our manuscript and their valuable feedback.  Their constructive comments have significantly improved the clarity and rigor of our work

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper proposes a novel deep learning model called DA-AHA-CLT for predicting transmission line temperatures, combining convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and temporal pattern attention (TPA) mechanisms. The model is optimized using a new Dynamic Adaptive Artificial Hummingbird Algorithm (DA-AHA). Experiments show DA-AHA-CLT significantly outperforms traditional models and other optimized models across multiple metrics, achieving an R2 score of 0.987 for full time-step prediction, compared to 0.878 for the base CLT model. 

 

However, I have the following concerns,

 

(1) Time series prediction has historically posed significant challenges for deep learning-based methods. For instance, in the renowned Makridakis (M) Competitions, it is only recently that hybrid models combining traditional statistical or machine learning techniques have emerged as top performers. Therefore, when the authors claim victory for their model, it is essential to also consider the performance of traditional time series prediction models such as ARIMA, SARIMA, and LGBM. For more detailed information, they can refer to this Arxiv paper: https://arxiv.org/abs/2401.13912.

 

(2) Secondly, I have reservations about the model the authors intend to use. Although it appears suitable for academic testing purposes, its practicality is questionable. For instance, as an end user seeking to predict line temperature for the next seven days, I would not have access to the ambient temperature or voltage data for that period. Therefore, while the model may be acceptable as a toy model for testing, its real-world applicability remains unproven. Please provide evidence of its practical utility.

 

(3) The authors have opted for the Hummingbird Algorithm, yet they also discuss other population-based optimization methods, such as Particle Swarm Optimization (PSO) and the Whale Optimization Algorithm. The question arises: why did they choose the Hummingbird Algorithm?

 

(4) The content of the paper should be presented as a coherent, descriptive narrative rather than a series of bullet points. For instance, lines 314 to 325 discuss the topic of "2.3.3 Foraging strategy selection." However, the current writing style does not form a descriptive paragraph. Instead, it resembles a list of points outlining Guided Foraging and Territorial Foraging. To enhance clarity and readability, it would be beneficial to transform these points into a more integrated and explanatory text.

 

(5) Meticulous proofreading is essential. There are numerous errors in the figure indexing. For instance, line 506 references Figure 1, but it should actually be Figure 7.

 

(6) Additionally, while the 10% outlier threshold may hold true in some cases, it oversimplifies the prediction. According to Figure 7, there is an actual periodic oscillation in line temperature. This observation brings me back to my initial question: Can a simpler method be effective in this scenario?

 

Comments on the Quality of English Language

Proofreading is needed

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting study. This paper proposes a deep learning model architecture based on the Dynamic Adaptive Artificial Hummingbird Algorithm (DA-AHA), named DA-AHA-CNN-LSTM-TPA (DA-AHA-CLT). However, some major issues need to be improved as follows:

1. In the abstract section, there are some abbreviations such as R², RMSE, MAE and MedAE. First, you need to use their longcase, followed by abbreviations.

2. The abstract section could have more numerical results.

3. Some sentences in the introduction need reference. For instance, "Traditional transmission line temperature prediction methods primarily rely on physical or statistical models. However, these approaches often face limitations when handling multivariate time series data and complex dynamic environments. Physical models depend heavily on extensive a priori knowledge and assumptions, making them difficult to adapt to changing external conditions. Statistical models, on the other hand, struggle with prediction accuracy when dealing with highly nonlinear and multivariate correlated time series. Moreover, traditional models struggle to balance the modeling requirements of short-term trends and long-term dependencies, leading to predictions that deviate from actual conditions and limiting their applicability in modern power systems".

4. Please avoid marginal subtitles such as "2.1 LSTM". You can say, "2.1 The implementation of LSTM method in the proposed method" like that.

5. Also, a subsection of 2.1 LSTM needs a reference. Please read this paper (Switched Auto-Regressive Neural Control (S-ANC) for Energy Management of Hybrid Microgrids). This paper is related to this study.

6. I can not see the explanations of some symbols in the equations. For example, what are W_f, W_c, W_i and so on? Please check all equations.

7. It would be better to put a table in the literature review section to compare the traditional and your methods.

8. You used some longcase of abbreviations more than once. If you use them in the introduction section, you do not need to use their longcase in another section. Please check all of them.

9. You do not need Figure 12 because of Table 6.

10. What are the limitations of your study? Please mention them in your study.