Fault-Detection-Based Machine Learning Approach to Multicellular Converters Used in Photovoltaic Systems
Round 1
Reviewer 1 Report
The paper addresses a fault detection method based on machine learning for a multicellular converter used in PV systems. I consider that the paper presents an interesting idea, but I have some issues that I think can help to improve the article.
The paper addresses a fault detection method based on machine learning for a multicellular converter used in PV systems. I consider that the paper presents an interesting idea, but I have some issues that I think can help to improve the article.
Improve the quality of Fig 1 and 2.
Which values can take Si in (2)? are states (0 or 1) or duty cycles ([0,1]).
Does Vdc is the output for the DC/DC boost converter?
Is despised the dynamic for the capacitors in Vdc?
Does the math model consider a load resistor in the output instead power grid?
The use of Vdc as constant and Rl in the model must be justified because in Fig 1 are shown as input an mppt converter and output the power grid.
Explain why (20) fulfill the stability criteria
How is fulfilled the stability criteria in the Exact Linearization Control?
How must the parameters be selected in Kr?
What considerations allow the model in Figure 3 to become the model in Figure 4?
For the test shown in Fig. 5, Are the controller's parameters adequately syntonized? It affects the dynamic response directly.
Does the simulation consider the control for the output current?
What kind of failure is considered for the tests?
How do the control algorithms respond to deviations in parameters and Vdc ? to ensure robustness
As the work focuses on Machine learning, I think that section 7 can present more in detail the evaluation of the proposed algorithms for fault detection. Additionally, knowing about the computational complexity of the algorithm's implementation would be interesting.
Author Response
All response to comments are in attached file
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1. Research motivation and research significance should be elaborated in Introduction.
2. Has the fit of the research method and the research subjects been demonstrated? Can a comparison with other methods be made?
3. The logic of the literature review needs to be improved.
4. There should be more discussion in the results.
Author Response
All response to comments are in attached file
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Add the name for the first figure in section 4.
Fix table 3.
As time processing is added in section 7, the capacities for the used PC must be added (Processor, Ram, ...).
Author Response
All responses to reviewer comments are in attached file
Author Response File: Author Response.pdf
Reviewer 2 Report
Agree to publish
Author Response
The authors would like to thank the reviewer for his useful comments
