Investigation of a Hybrid LSTM + 1DCNN Approach to Predict In-Cylinder Pressure of Internal Combustion Engines

Round 1

Reviewer 1 Report

The paper is well written and the subject is an interesting one. 

The paper presents in a coherent way an evaluating the forecasting performance of a LSTM+1DCNN architecture.  The aim of this research is to insert virtual sensors in the onboard control system, or more precisely, predicting the cylinder pressure of internal combustion engines

The paper contains important information and makes some contributions through the results section (Section 3).

I consider that the paper can be published in Information Journal, but some clarifications must be made before:

1. Why is P_intake missing from Figure 4 (b)?

2. In Figure 5 (b) the text cannot be read.

3. At line 237 and relation (1) should be the same index (predicted / precited); 4. At line 272, should be R².

Author Response

Dear Reviewer, Thank you for the important suggestions provided. Attached are responses to your comments

Author Response File: Author Response.docx

Reviewer 2 Report

1. In the line 19 the authors wrote "The aim of the present line of research is to insert virtual sensors in the onboard control system", and in line 322-323 the authors wrote "Our research goal is to explore the potential of using advanced machine learning technologies to replace physical sensors". What is the main aim of this article? 

2. There are no sources given under the tables and figures in the article.

3. In line 131 is written table 1 and should be written table 2. 

4. Above Table 2 is a very small description of this table.

5. In line 174, there should be a parenthesis and a dot after the word CAD.

6. There is no description of Figure 4.

7. 32 items of literature is not enough. There should be at least 50.

Author Response

Dear Reviewer, Thank you for the important suggestions provided. Attached are responses to your comments

Author Response File: Author Response.docx

Reviewer 3 Report

This article will be particularly useful for researchers, engineers, and professionals in automotive engineering, machine learning, and internal combustion engine optimization.

The paper investigates a hybrid approach combining Long Short-Term Memory (LSTM) with a one-dimensional Convolutional Neural Network (1DCNN) to predict the in-cylinder pressure of internal combustion engines. The objective is to enhance the control of internal combustion engines by utilizing advanced machine learning techniques. I have several remarks that should be taken into consideration by the authors:

 

1. Clarity and Organization:

The introduction should provide a clearer background on the existing methods and the gap the current research aims to fill.

 

2. Methodological Concerns:

The choice of LSTM+1DCNN needs a justification. Why was this specific combination chosen over other potential machine-learning models?

Detailed hyperparameters used for the LSTM and 1DCNN models should be provided to ensure reproducibility.

 

3. Data and Validation:

Information about the dataset, such as its source, size, and features, is missing. This needs to be included for clarity.

 

4. Results and Discussion:

It would be helpful to provide a section discussing the practical implications of the research findings for the automotive industry.

Potential limitations of the study should be addressed.

 

5. Technical and Logical Concerns:

The paper should address potential sources of bias or error in the measurements or predictions.

 

6. References:

It might be useful to compare and contrast the findings with existing literature on the topic. Here are some possible references: https://doi.org/10.4271/2023-01-0102; https://doi.org/10.17531/ein.2022.2.19

 

 7. Miscellaneous:

Consider adding a section discussing potential future work or how this research can be extended or applied in other contexts.

 

Conclusion:

 

The manuscript presents a novel approach that could significantly affect the automotive industry. However, several sections require further refinement and clarity to meet the journal's standards. Addressing the aforementioned critical comments before considering it for publication is recommended.

Author Response

Dear Reviewer, Thank you for the important suggestions provided. Attached are responses to your comments

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Accept. The reviewer's comments have been taken into account in the article.

Author Response

Thanks for the comment

Reviewer 3 Report

The authors attached an article without marked changes, which makes the analysis of changes a fficult. After reviewing the revised version of the manuscript, I appreciate the efforts to address the comments raised in my initial review. However, there are still some areas that require further attention:

1.      Clarity and Organization:

While the manuscript has tried to elucidate the research objective, it still lacks a clear background on the existing methods in the introduction section. It is crucial to provide a more comprehensive overview of the current state of the art and how the present research fits into this landscape. The authors reply that mentioned by the reviewer papers have been added to the article, but in the revised version of the manuscript, they are not present.

2.      Methodological Concerns:

I found references to the LSTM+1DCNN combination in the article, which suggests an attempt to justify this specific choice. However, comparing this choice to other potential machine-learning models, a more detailed justification would provide greater clarity to readers.

I recommend the authors address these remaining concerns to ensure the manuscript offers a comprehensive and clear contribution to the field.

 

 

Author Response

attached is the response to the new revisions

Author Response File: Author Response.docx