A Machine Learning Approach to Predict Relative Residual Strengths of Recycled Aggregate Concrete after Exposure to High Temperatures

Response:

Thank you for your comment, we appreciate it and we would kindly appreciate keeping the figure as It is as the collected data in the figure is based on remarkable research articles and actual experience from the lab. As well as the colors have ben chosen to express the increase of the temperature.

Notes/actions:

-

Response:

Thank you for your comment, The temperature conditions for all the collected data were somehow in the same range and circumstances. 1. all collected data consider the behavior post high temperature exposure, 2. The cooling process for all samples is air cooling without using water, 3. The temperature rate of the oven is in a range that explained in the paper at lines 269-273, where its effect is not severely affect the performance.

Notes/actions:

Any data used in the paper from any research paper has been carefully investigated and any paper has different conditions of heating or cooling has been excluded from the beginning.

Response:

Thank you for your response. The problem is fixed.

Notes/actions:

Figure number has been corrected

Response:

Thank you for your comment. It should have been minus.

Notes/actions:

The equation has been corrected.

Response:

Thank you for your response the references has been fixed.

Notes/actions:

References have been added

Response:

Thank you for your recommendation. The recommended paper has been used in the paper as 23^rd reference.

Notes/actions:

Lines 60 to 64

“Initial models to achieve this goal were based on statistical techniques, however, in recent years, with advancements in the field of machine learning, many researchers have shown interest in the prediction of the mechanical properties of concrete [20-22] and other materials [23]”

Response:

Thank you for your insightful comments on the abstract of my paper. We have carefully considered your feedback and made revisions to address the points you raised.
In response to your suggestion about lacking specificity regarding the novel contributions of the research, We have explicitly stated in the revised abstract that our paper introduces innovative hybrid machine learning models. These models provide practical equations and model trees that are user-friendly for estimating the mechanical properties of recycled aggregate concrete after exposure to high temperatures. I have also emphasized the uniqueness of our contribution, highlighting that while the models are more human-friendly compared to traditional ML algorithms, they maintain a high level of accuracy.

Notes/actions:

The abstract was completely rewritten to include unique aspects and advancement mentioned by the machine learning models as follows

“In recent years, there has been a heightened focus among researchers and policymakers on assessing the environmental impact and sustainability of human activities. In this context, the reutilization of construction materials, particularly recycled aggregate concrete, has emerged as an environmentally friendly choice in construction projects, gaining significant traction. This study addresses the critical need to investigate the mechanical properties of recycled aggregate concrete under diverse extreme scenarios.

Conducting an extensive literature review, key findings were synthesized on the relative residual strength of recycled aggregate concrete following exposure to high temperatures. Leveraging these insights, innovative hybrid machine learning models were developed, offering practical equations and model trees for predicting the relative residual compressive strength, flexural strength, elasticity modulus, and splitting tensile strength of recycled aggregate concrete post high-temperature exposure. Uncertainty analysis was performed on each model to assess the reliability, while sensitivity analysis was performed to find out the significance of each input variable for each predictive model.

The unique contribution of the paper lies in the provision of easily applicable equations and model trees, enhancing accessibility for practitioners seeking to estimate mechanical properties of recycled aggregate concrete. Notably, our hybrid machine learning models stand out for their user-friendly nature compared to conventional ML algorithms, without compromising on accuracy. This paper not only advances our understanding of sustainable construction practices but also equips industry professionals with efficient tools for practical implementation.”

Response:

Thank you for your keen observation. The sample sizes are as follows: 403 for compressive strength, 65 for flexural strength, 323 for modulus of elasticity, and 85 for splitting tensile strength.

To address the limitations identified in the literature review and the models, a new section titled "Applicability and Limitations" has been incorporated after the Results and Discussion Section.

Notes/actions:

The first paragraph of the Methodology has been rewritten to include sample sizes as following
“This paper introduces a series of models developed to predict diverse mechanical properties of RCA concrete, including compressive strength, flexural strength, modulus of elasticity, and splitting tensile strength. The primary objective is to create models that are easily applicable for practitioners, engineers, and scientists to estimate these mechanical properties in real-world scenarios. The sample sizes for each property were as follows: 403 for compressive strength, 65 for flexural strength, 323 for modulus of elasticity, and 85 for splitting tensile strength. To ensure the robustness of our models, we employ the Synthetic Minority Over-sampling Technique with Edited Nearest Neighbors (SMOTER) in the initial step to assess data balance. In the event of imbalances, oversampling, as detailed in preceding sections, is utilized to rectify the data distribution.”

Also for limitations of the literature review two paragraphs has been assigned to a newly written Applicability and Limitations section as follows

“Note that the data has been gathered through a comprehensive literature review, introducing the possibility of certain biases. For instance, the tendency to publish positive results may skew the findings in favor of better performance for RCA concrete exposed to high temperatures. Additionally, the selected data are sourced from public databases of published papers, potentially introducing bias as private information is not considered. Consequently, caution is advised when utilizing the results of this research, taking into account the aforementioned biases.

Nevertheless, the proposed technique, when applied to a larger dataset, has the potential to yield models with a similar level of performance. In cases where more data is available, it is recommended to adopt the algorithm proposed in this study and thoroughly evaluate the models before implementation.”

Response:

Thank you for your thoughtful feedback on our paper. We appreciate the careful consideration you've given to the limitations of our study.

Regarding the machine learning models used for predicting the relative residual strength, uncertainty analysis, and sensitivity analysis, we acknowledge that the paper could benefit from a more explicit discussion of their limitations. While we did encounter certain challenges inherent to these models, we will ensure to incorporate a dedicated section in the revised manuscript to address this aspect comprehensively.

We welcome any additional suggestions or insights you may have and look forward to further discussions on this topic. Thank you again for your constructive feedback.

Notes/actions:

A new section called Applicability and Limitations has been added to the paper and the following part covers the limitation of the methods used in this study

“The proposed models exhibit certain limitations. While they leverage linear regression to offer user-friendly and interpretable results, their reliance on this approach restricts their ability to capture nonlinearity. Consequently, the models face constraints in terms of accuracy and generalization. A potential avenue for future research involves exploring the applicability of nonlinear regression techniques within the proposed framework.

Additionally, the use of SMOTER to address data imbalance raises concerns about potential biases. The synthesized database generated by SMOTER may deviate from the distribution of the actual database, impacting the reliability of the models. It is crucial to acknowledge that the training data is confined to that collected for the study, and real-world applications may pose challenges for materials with characteristics divergent from those present in the used database.

Importantly, uncertainty analysis is inherently tied to the assumptions made during both the modeling process and subsequent analysis. For instance, assuming a normal distribution for the error of the models influences the interpretation of confidence intervals. Caution is advised to prevent overestimation or underestimation of error. Additionally, sensitivity analysis is contingent on the chosen evaluation metrics and does not account for codependency among features, potentially leading to higher or lower sensitivity toward specific variables.”

Response:

Thank you for your comment. The practical application of our study is a pivotal aspect, and we aim to enhance clarity for the reader. To achieve this, we furnish a easily understandable example of real-world model usage in the Applicability and Limitations section. Additionally, a summary highlighting these practical applications is presented in the conclusion.

Notes/actions:

The following information has been added to the Applicability and Limitations section

“The proposed models can be valuable in various scenarios where RCA concrete structures are subjected to fire. For instance, following a fire incident in a building, engineers can evaluate the mechanical properties of the exposed concrete by understanding the mixture design and determining the maximum temperature reached during the fire. This approach is equally applicable to tunnel liners post-accidents. By considering these parameters, structural integrity and stability can be thoroughly examined through various structural analyses.”


The following information has been added to the Conclusion section

“Using the proposed models engineers, can assess the mechanical properties of exposed concrete by analyzing mixture design and max temperature in the aftermath of a fire. It enables a thorough examination of structural integrity and stability using diverse structural analyses.”

Response:

Thank you for taking the time to review our article and providing valuable feedback. We appreciate your diligence in studying our work. Based on your comments, we have carefully reconsidered the structure of the paper and made significant revisions to address the concerns raised.

We understand the importance of having clear and standard sections in a research article.

Notes/actions:

In response to your feedback, we have restructured the paper

-        Introduction

-        Theoretical Machine Learning Framework

-        Data Collection

-        Methodology

-        Results and Discussion

-        Applicability and Limitations

-        Conclusion

To make the purpose of the study and achievements more clear the abstract has been rewritten as follows

“In recent years, there has been a heightened focus among researchers and policymakers on assessing the environmental impact and sustainability of human activities. In this context, the reutilization of construction materials, particularly recycled aggregate concrete, has emerged as an environmentally friendly choice in construction projects, gaining significant traction. This study addresses the critical need to investigate the mechanical properties of recycled aggregate concrete under diverse extreme scenarios.

Conducting an extensive literature review, key findings were synthesized on the relative residual strength of recycled aggregate concrete following exposure to high temperatures. Leveraging these insights, innovative hybrid machine learning models were developed, offering practical equations and model trees for predicting the relative residual compressive strength, flexural strength, elasticity modulus, and splitting tensile strength of recycled aggregate concrete post high-temperature exposure. Uncertainty analysis was performed on each model to assess the reliability, while sensitivity analysis was performed to find out the significance of each input variable for each predictive model.

The unique contribution of the paper lies in the provision of easily applicable equations and model trees, enhancing accessibility for practitioners seeking to estimate mechanical properties of recycled aggregate concrete. Notably, our hybrid machine learning models stand out for their user-friendly nature compared to conventional ML algorithms, without compromising on accuracy. This paper not only advances our understanding of sustainable construction practices but also equips industry professionals with efficient tools for practical implementation.”

To make the purpose and achievements of the paper clearer the final part of the introduction has been rewritten as follows

“The increasing preference for recycled aggregate concrete highlights the necessity for user-friendly methods to predict its mechanical properties under diverse conditions, especially when exposed to elevated temperatures. This study addresses this need by aiming to provide engineers and researchers with a versatile and easily applicable tool. This tool aims to facilitate precise assessment of various mechanical characteristics of recycled aggregate concrete when subjected to high-temperature conditions.

One practical application of these predictive models is in assessing the functionality and structural integrity of buildings following fire incidents. This assessment relies on predicting the compromised mechanical attributes resulting from exposure to fire. Importantly, the existing literature lacks models for such predictions, and the utilization of machine learning techniques in predicting properties of recycled aggregate concrete under high temperatures has been significantly overlooked.

This investigation places emphasis on providing an optimal solution by initially exploring linearity within the dataset. If linear models prove insufficient, a proposed algorithm leverages the M5p decision tree technique and various linear regressions to provide a interpretable and accurate model for mechanical properties of RCA concretes. This combination aims to predict the residual compressive strength, flexural strength, elasticity modulus, and splitting tensile strength of recycled aggregate concrete. The models have been evaluated rigorously using evaluation metrics and cross-validation. An uncertainty analysis has been performed to find out reliability of the models and compare them to each other. A sensitivity analysis is performed to find out importance of the variables in prediction of each mechanical properties.”

Response:

Regarding your suggestion to include numerical results in the Abstract section, we agree that this would enhance the comprehensiveness of our abstract. We revised the abstract to incorporate key numerical findings obtained during our research process, providing readers with a succinct overview of our study's outcomes.

Notes/actions:

Following text has been added to the abstract.

“This paper presents interpretable models achieving high levels of performance, with $R^2$ values of 0.91, 0.94, 0.9, and 0.96 for predicting the relative residual compressive strength, flexural strength, modulus of elasticity, and splitting tensile strength of RCA concrete exposed to high temperatures, respectively.”

Response:

We appreciate your observation regarding the abundance of references in the Introduction section. We carefully revised the section to streamline and reduce the number of references, ensuring that only the most relevant and critical citations are included. This adjustment enhanced the clarity and focus of our introduction while maintaining its scholarly rigor.

Notes/actions:

Following citations has been removed from the paper

“Ahmad et al. \cite{ahmad2021comparative} used 207 data points to develop models based on an artificial neural network and decision tree that predicts the relative residual mechanical properties of regular concrete after exposure to high temperatures. Hoong et al. \cite{hoong2020determination} used a convolutional neural network to find the composition of RCA using images and compared the results to manual sorting. Chen et al. \cite{chen2021convolution} used a convolutional neural network to predict the relative residual compressive strength of concrete exposed to high temperatures. They used 19 features including mixture design and heating profile to develop their models. Xu et al. \cite{xu2021computation} used standalone and ensemble machine learning techniques for the prediction of the mechanical properties of high-performance concrete.”

Response:

We acknowledge your suggestion to reorganize the manuscript by moving the Theoretical Machine Learning Framework section and the Data Collection section to the Methodology section. We agree that this adjustment would enhance the logical flow and organization of our paper, ensuring that all methodological aspects are presented cohesively.

Notes/actions:

Theoretical Machine Learning Framework, and Data collection  section has been moved to Methodology section per recommendation of the reviewer.

Response:

Thank you for your comments. After careful equations designations has been found to be consistent.

Notes/actions:

The designations of the models in Fig 8 and Table 3 has been found to be the same.

The designations of the equations in Table 4 and Figure 10 have been found to be consistent

Response:

We acknowledge your suggestion to provide a more detailed description of Figure 7. We revised the figure caption to include additional information that clarifies its content and relevance to the study.

Furthermore, we appreciate your recommendation to relocate Table 2. We agree that this adjustment enhanced the organization and flow of the paper, and we implemented this change accordingly. Figure 10 and Figure 15 has been kept in the results and discussion sections since they are derived from the results of the developed models.

Notes/actions:

The following text has been added as description of figure 7

“Schematic view of steps used for model development and evaluation in this study including data importing, preprocessing, model development, evaluation, sensitivity analysis.”

Table 2 has been moved under the umbrella of the Methodology section.