Effect of Pulsed Electric Field on the Drying Kinetics of Apple Slices during Vacuum-Assisted Microwave Drying: Experimental, Mathematical and Computational Intelligence Approaches

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The research article presents a comprehensive study on the impact of pulsed electric field (PEF) treatment on the drying kinetics of apple slices using a vacuum-assisted microwave dryer. The article utilizes a combination of experimental methods, mathematical modelling, and machine learning techniques to analyze and predict moisture ratios during the drying process. The study is well-structured, and the use of advanced modelling techniques is commendable. However, several areas need improvement, clarification, and correction.

Comments:

1)The abstract could specify the percentage reduction in drying time and the exact quality parameters that were maintained or improved due to PEF treatment.

2)The introduction could be enhanced by a more in-depth discussion of the challenges and gaps in current drying technologies that this study aims to address.

3)The description of sample preparation is clear, but it would be helpful to provide a rationale for choosing the specific apple variety (cv. Gravenstein).

4)The method for calculating specific energy should include more details on the justification for the selected parameters.

5)The description of the drying equipment and setup (Figure 1) is adequate, but more information on the calibration and validation of the equipment would enhance credibility.

6)The selection of the Page and Weibull models is well-justified. Discuss why these models were chosen over others and their limitations in more detail.

7)The choice of transfer functions and the structure of the neural networks should be better justified with references to previous studies or theoretical considerations.

8)The drying rate and moisture ratio results are well-presented. The discussion could benefit from a comparison with other similar studies to contextualize the findings.

9)The discussion on the limitations of the models should be expanded, including potential sources of error and how they were mitigated.

10)The impact of PEF and microwave power on specific compounds (e.g., antioxidants, phenolics) should be discussed in more depth, particularly regarding the mechanisms behind these changes.

11)The performance of the ANN and SVR models is fine. It would be beneficial to include a discussion on the computational efficiency of these models and their potential for real-time application in industrial settings.

12) The sensory analysis could be expanded to include more detailed feedback from the panellists, including any specific comments on texture and flavour changes.

 13)Some figures (e.g., Figure 3) could be made clearer with better labelling and larger text. Additionally, including error bars in graphical representations of data would provide a better understanding of variability and reliability.

14)Suggestions for future research directions could be provided.

 

Comments on the Quality of English Language

Moderate editing of the English language is required.

Author Response

Dear Reviewer #1

I would like to appreciate you for revision of the paper, I studied all considered comment and made corrections as you have noted. At following responses of all comments are separately presented in detail;

Comment1: The abstract could specify the percentage reduction in drying time and the exact quality parameters that were maintained or improved due to PEF treatment.

Response1: percentage reduction in drying time was added in the abstract (The  Based on the findings, implementing PEF reduced the drying time from 4.2 to 31.4% compared to the untreated sample). In addition, the improved quality parameters was mentioned ( Drying apple slices using PEF treatment and 100W of microwave energy not only reduces drying time but also maintains the chemical properties such as total phenolic content, total flavonoid content, antioxidant activity), vitamin C, color and sensory qualities of the product).

 

Comment2: The introduction could be enhanced by a more in-depth discussion of the challenges and gaps in current drying technologies that this study aims to address.

Reponse2: The exist challenges and gap was discussed in the L56-62 (the sections is highlighted. Although the MV method ameliorates the economy and efficiency of the drying process, preserving retained water-soluble components of some fruit such as apples is essential. Thus, the extent of the changes produced on the solid matrix of these kinds of fruit should be reduced, and to achieve this purpose, pretreatment method such as pulsed electric field [3,10]  has been recommended. Pretreatment methods not only preserve product quality during drying but also represent energy-efficient and economical manner). Also, the application of PEF to address the mentioned challenges was described in the L64-70 (the sections is highlighted. It not only influences drying rate and energy consumption but also mitigates excessive temperature elevation, safeguarding against undesirable color alterations [11,12]. PEF treatment induces an electroporation phenomenon in the cell membranes of the fruit, resulting in enhanced mass transfer and improved water permeability throughout the drying process [13]. The utilization of PEF as a pretreatment for various drying tech-niques and its advantageous outcomes, encompassing decreased drying times and improved color preservation, have been observed in several studies).

 

Comment3: The description of sample preparation is clear, but it would be helpful to provide a rationale for choosing the specific apple variety (cv. Gravenstein).

Reponse3: The variety of cv. Gravenstein was chosen based on its availability in the Danish market as well as its popularity for consumers (Apples (cv. Gravenstein) that is one of the available apple in the Danish market and popular for consumers were purchased from a local market (Copenhagen, Denmark) and stored in a refrigerator at 4°C for 48 hours)

 

Comment4: The method for calculating specific energy should include more details on the justification for the selected parameters.

Response4: Specific energy equates to kinetic energy, and is the amount of energy required to move a unit of mass [18]. The electric field strength (1.2 KV/cm) and specific energy (15.2-26.8 kJ/kg) were calculated by equations 1 and 2, respectively [18]. The parameters chosen for calculating specific energy were based on standard practices in the field and literature [11, 12, 45].

 

Comment5: The description of the drying equipment and setup (Figure 1) is adequate, but more information on the calibration and validation of the equipment would enhance credibility.

Reponse5: The calibration and validation process of applied temperature sensor was added to the text (The temperature sensors in the MV dryer should be calibrated and validated to give proper measurements. First, it was made sure that the microwave was off to avoid microwave interference. The calibration was done in a controlled environment, after which the temperature sensor was placed at specified points within the dryer where they are to be used during operation. The sensor was evaluated across a variety of temperatures that reached the entire operating range of the MV dryer. Readings were taken at each point for the temperature from both the sensors on the dryer and the reference thermometer. After calibration, the microwave dryer was run under normal operating conditions and the calibrated reference thermometer periodically checked the temperature readings. Finally, we verified consistency and accuracy of sensor readings under all operating conditions, maintaining those capabilities over time)

 

Comment6: The selection of the Page and Weibull models is well-justified. Discuss why these models were chosen over others and their limitations in more detail.

Reponse6: The superiorities of the Page and Weibull models compare to some mathematical models were described (In fact, various mathematical models have been applied to evaluate the drying kinetic in the literature [22,23,38] and based on the desired outcome, some of them was selected as appropriate mathematical models. In this paper, Page and Weibull were applied because they presnet better flexibility and fit for non-linear and complex drying behaviors. Furthermore, although some of mathematical models fit drying data well but may lack the flexibility and interpretability of the Page and Weibull model, particularly in describing the entire drying process).

 

Comment7: The choice of transfer functions and the structure of the neural networks should be better justified with references to previous studies or theoretical considerations.

Reponse7: The applied transfer function consist of elliotosig (non-linear activation function for ANN ), tansig (normalizing activation function in ANN), logsig (binary classification activation in ANN),radbas ( local activation for function approximation), poslin (non-linear activation for deep networks) and tribas (Piecewise linear activation for interpolation) were chosen based on the mentioned application that were described completely in the literature [25,33,34,35,36]. Based on the test and error we found that one hidden layer did not present acceptable accuracy and three hidden layers spent much time compared two hidden layer with-out less error.

 

Comment8: The drying rate and moisture ratio results are well-presented. The discussion could benefit from a comparison with other similar studies to contextualize the findings.

Reponse8: The authors try to discuss more about the comparison of this study with similar studies ( The influence of the PEF treatment on the drying rate of the current research was in agreement with [16], [19], and [41]. Impact of pre-treatment with PEF on drying rate and changes in spinach quality dur-ing hot air drying was investigated [16]. They reported the drying time for PEF samples were shortened for 1.5 h by air drying. They supposed that moisture supply to the sur-face was accelerated after PEF treatment, so the drying rate were increased in the early stage of drying. Also, influence of PEF pre-treatment on the convective drying kinetics of onions was assessed by [19]. The drying time and the diffusion coefficient for onion in this study were optimal at a medium PEF treatment intensity (4–6 kJ/kg) giving a Z value in the range of 0.53–0.60. They reported that applying PEF treatment resulted in an enhancement of the drying rate throughout the drying process. The sharp incline observed in the middle phase of the drying curve also indicated that liquid moisture diffusion was the primary mechanism for eliminating water from the sample's interior.

Similar behavior of PEF treatment on the MR trends over the drying time was reported for onion [19], carrot [3], spinach [16], and potato [18]. Effects of PEF on vacuum dry-ing and quality characteristics of dried carrot was evaluated by [3] under different drying temperatures (25, 50, 75, and 90 °C). When the temperature increased from 25 to 90 °C, the total drying time decreased from 21,600 s to 5400 s and from 9720 s to 3600 s for untreated and PEF-pretreated samples, respectively. The application of PEF treatment caused a noticeable decrease of drying time (by 33–55%) and accelera-tion drying kinetics even at a higher temperature (90 °C); it evidently reflected the ef-fects of electroporation by the PEF treatment).

 

Comment9: The discussion on the limitations of the models should be expanded, including potential sources of error and how they were mitigated.

Reponse9: Although the overall accuracy and errors of Page and Weibull models were acceptable, it appears that all the models assumed a form of drying kinetics that may not capture all the complexities for the drying process. Thus, it should be combined with other additional empirical models. Further, the models have the assumption that the drying characteristics are homogeneous, which may not be the case for sliced apple samples; thus, this model could prove to be useful for materials with relatively uniform properties, or perhaps be integrated together with other models accounting for heterogeneity.

 

Comment10: The impact of PEF and microwave power on specific compounds (e.g., antioxidants, phenolics) should be discussed in more depth, particularly regarding the mechanisms behind these changes.

Reponse10: The influence of PEF on the chemical properties was discussed (During treatment with PEF, it can potentially disrupt cell membranes and eventually lead to the release of intracellular components. The disruption may result in degradation or oxidation of certain sensitive antioxidants during the process when otherwise protected in cell structures[47]. Polyphenols are very sensitive to heat and drying is done at high temperatures, so the degradation of polyphenols can easily occur by oxidation or simple thermal degrada-tion. The applied PEF may enhanced the cell membrane permeability, facilitating easy access of polyphenols and hence their bioavailability. Based on the literature, PEF treatment could induce reversible breakdowns in the cell membrane at the moment of application, called electroporation. As a result, there was an increase in the permeability of the cell membrane, and intracellular flavonoids were released into the surrounding medium for preservation [47]. It seems PEF decreased generation of some reactive oxygen species, induce oxidative stress, and thus oxidize and could preserve the vitamin C content. Ascorbic acid is easily oxidized, and high oxidative stress due to PEF treatment could maintain the vitamin C content [24,41]).

 

Comment11: The performance of the ANN and SVR models is fine. It would be beneficial to include a discussion on the computational efficiency of these models and their potential for real-time application in industrial settings.

Response11: Based on results of this paper and literature [6,25,26,27,28,36,37], it can be claimed that ANN and SVR models are prominent ML which can be used for evidence predicting and optimizing drying process in industrial settings. ANN models, as complex, non-linear relationship learners, are characterized by their layered architecture approach. Its major advantage is high computational performance and adaptability to real-time applications, which are extremely necessary under equipped modern hardware with optimized algorithms. They can be used to process volumes of large datasets from drying operations quickly for providing correct predictions and adjustments in real-time[26,28]. On the other side, whereas in general less computationally intensive and quite robust with smaller datasets, SVR models may not scale to achieve real-time performance, because the computational complexity in dealing with large feature spaces and nonlinear mappings may be an issue[37]. In this sense, in the industrial field, where dryers work against variable conditions and quick, adaptive responses to their actions are required, always an important choice will be the ANN model, due to better handling of dynamic and complex data, while the SVR models can be very useful in the cases with constrained dimensions of data and simpler relationships. While both models help improve operational efficiency, ANN's computational efficiency closely relates it as being the instant model of choice for large industrial models.

 

Comment12: The sensory analysis could be expanded to include more detailed feedback from the panellists, including any specific comments on texture and flavour changes.

Response: The author appreciate the helpful comment but Because the experiments were done and repeating the experiment is time-consuming, the sensory analysis by panellists is impossible at this moment.

 

Comment13: Some figures (e.g., Figure 3) could be made clearer with better labelling and larger text. Additionally, including error bars in graphical representations of data would provide a better understanding of variability and reliability.

Response13: The quality of figures (especially figure.3) is enhanced for more clear and also the size of the text was increased.

 

Comment14: Suggestions for future research directions could be provided.

Reponse14: Helpful suggestion were added to the conclusion section as future research directions ( However, some challenges such as achieving more accuracy and versatility of developed MV dryer and applied ML models were found in this study. To overcome these kind of challenge it is suggested the dyer integrate with some novel technologies such as machine vision systems. This involves aligning the imaging setup to monitor the product consistently and possibly synchronizing it with the drying parameters. In addition, Implement the trained models within the drying system’s control software can be helpful action to achieve more accuracy. The models should process image processing data in real-time to predict moisture content and other relevant parameters. Moreover, closing a feedback loop in which the predictions of the model can enable a dynamic adjustment of the drying condition).

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

(1) Write keywords in alphabetic order.

(2) The citation format of the literature in the article is not consistent, such as Lines 48, 127, 195, 241, etc. There are multiple similar errors throughout the entire manuscript. Please carefully checked and corrected.

(3) The unit of “mg mL−1” was written incorrectly in Line 211.

(4) How many times are each experiment repeated in this article?

(5) Tables and figures do not use abbreviations, and there is no need to add periods after words, such as Lines 142, 144, 222, 233, 264, 268, etc. There are multiple similar errors throughout the entire manuscript. Please carefully checked and corrected.

(6) Please add the general name of Figure 2.

(7) Please add the name of the subgraph in Figure 6.

(8) The font format of some proprietary terms should be italicized in the manuscript, such as Lines 105, 688, etc. Suggest the author to check and make modifications.

(9) Please cite a reference for the section “2.7.2. Mechanical properties” in Line 290, and “2.7.3. Sensory analysis” in Line 299.

(10) In order to conduct scientific discussions, more reference materials are needed to support your results, please clearly explain your findings, emphasize important or novel findings, and mention how they compare to the latest research results in the section “3.5. Characterization of the dried samples”.

(11) The format of the references is different. The initial letter of a sentence is capitalized in some titles of reference article, while the initial letter of a word is capitalized in others, for example, in Lines 622, 644, etc.

(12) There are many details and errors in the article that need to be carefully checked and corrected.

Author Response

Dear Reviewer #2

I am grateful for your constructive criticism and helpful comments. Responses of comments are presented following:

Comment1: Write keywords in alphabetic order.

Response1: The key words were organized  in alphabetic order

 

Comment2: The citation format of the literature in the article is not consistent, such as Lines 48, 127, 195, 241, etc. There are multiple similar errors throughout the entire manuscript. Please carefully checked and corrected.

Response2: The citation format of the literature were checked through the text and revised.

 

Comment3: The unit of “mg mL−1” was written incorrectly in Line 211.

Reponse3: The author appreciate attention of reviewer; the unit was amended.

 

Comment4: How many times are each experiment repeated in this article?

Reponse4: all of the experiments were replicated three times in this paper (the statement was added to the text-material and methods section).

 

Comment5: Tables and figures do not use abbreviations, and there is no need to add periods after words, such as Lines 142, 144, 222, 233, 264, 268, etc. There are multiple similar errors throughout the entire manuscript. Please carefully checked and corrected.

Reponse5: To prevent any confusion in the tables and figures, some abbreviations has been used and their definition were described in their captions. As reviewer mentioned, the periods after words through the text were checked and revised.

 

Comment6: Please add the general name of Figure 2.

Reponse6: The general name of Figure2 (Applied ML model) is added.

 

Comment7: Please add the name of the subgraph in Figure 6.

Response7: The subgraph of Figure 6.a and Figure 7.b are  added.

 

Comment8: The font format of some proprietary terms should be italicized in the manuscript, such as Lines 105, 688, etc. Suggest the author to check and make modifications.

Response8:  The font format of proprietary terms are changed to italicized.

 

Comment9: Please cite a reference for the section “2.7.2. Mechanical properties” in Line 290, and “2.7.3. Sensory analysis” in Line 299.

Response9: A reference for Mechanical properties [49] and Sensory analysis [14] are considered.

 

Comment10: In order to conduct scientific discussions, more reference materials are needed to support your results, please clearly explain your findings, emphasize important or novel findings, and mention how they compare to the latest research results in the section “3.5. Characterization of the dried samples”.

Response10: The author discussed more about the application and novelty of the current research compare the literature. As reviewer mentioned, more reference material are described to support the current results.

 

Comment11: The format of the references is different. The initial letter of a sentence is capitalized in some titles of reference article, while the initial letter of a word is capitalized in others, for example, in Lines 622, 644, etc.

Response11: The format of reference list was double checked and the mentioned errors are revised.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The article is fine now.

Accept in its present form

Reviewer 2 Report

Comments and Suggestions for Authors

no