Fault Management in Speed Control Systems of Hydroelectric Power Plants Through Petri Nets Modeling: Case Study of the Alazán Power Plant, Ecuador

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The aim of the presented work is the consideration of the usage of Petri nets in a hydraulic power plant SCADA system in order to decrease the intervention time during operation when a high-severity or mid/low severity fault occurs. The paper presents that the main problems in the specific power plant when such a fault occurs is the identification of the source of the fault and its following removal. 

I was unable to determine general weak areas of the article and can say that the presented work is very detailed and presents a very good overview of the topic. It also gives a more than adequate methodology for the mitigation of presented faults in the hydraulic power plant and the decreasing of the intervention time.

The results are well presented and are compared to real time data with a suggestion for the implementation of the presented work in the real life HPP.

My suggestions and comments are as follows:

1. The shown workflow diagram on figure 5 has differences from the accepted standards for flowchart diagrams that should be corrected. When a using a decision block (rhombus) you need to have one input and two outputs. There should be arrows showing what will happen when no signal is available or furthermore what will happen when the Injector control module is unavailable. The decision block can roll back to the PID controller waiting the availability of the module but this is needed as information. Also when talking about the Injector position signal, Speed of synchronization and the Load take decision blocks it should be noted which decision path is taken when (with a simple Yes/No or High/Low paths, etc.)
2. On figures from 8 through 11 it is good to show the dimension of the original time so it can be easier for a person overviewing the article to know that the time savings are in minutes which is significant for HPPs. Also it's probably better if the vertical axis is called "Intervention time" instead of "Original time".
3. It should be noted somewhere in the text when talking about the specific HPP what type of water turbine it uses - in this case a Pelton turbine. Therefore it will be nice to have in the conclusion a small comment about if the proposed solution can be implemented in HPPs with different type of water turbines or it can be used only for different HPPs with Pelton turbines.

Author Response

Comments 1: 

The aim of the presented work is the consideration of the usage of Petri nets in a hydraulic power plant SCADA system in order to decrease the intervention time during operation when a high-severity or mid/low severity fault occurs. The paper presents that the main problems in the specific power plant when such a fault occurs is the identification of the source of the fault and its following removal. 

I was unable to determine general weak areas of the article and can say that the presented work is very detailed and presents a very good overview of the topic. It also gives a more than adequate methodology for the mitigation of presented faults in the hydraulic power plant and the decreasing of the intervention time.

The results are well presented and are compared to real time data with a suggestion for the implementation of the presented work in the real life HPP.

My suggestions and comments are as follows:

1. The shown workflow diagram on figure 5 has differences from the accepted standards for flowchart diagrams that should be corrected. When a using a decision block (rhombus) you need to have one input and two outputs. There should be arrows showing what will happen when no signal is available or furthermore what will happen when the Injector control module is unavailable. The decision block can roll back to the PID controller waiting the availability of the module but this is needed as information. Also when talking about the Injector position signal, Speed of synchronization and the Load take decision blocks it should be noted which decision path is taken when (with a simple Yes/No or High/Low paths, etc.)

Response 1: We have thoroughly revised the diagram to align with standard flowchart conventions. Specifically, we ensured that all decision blocks now have one input and two labeled output paths (e.g., Yes/No or High/Low), as appropriate.

Additionally, we included explicit flow paths for the cases when signals are unavailable or when the Injector Control Module fails to respond. These paths now indicate that the system rolls back to the PID controller to await module availability. The logic branches for the Injector Position Signal, Speed of Synchronization, and Load Take Decision have also been clarified with clear labels to indicate the outcome of each decision condition.

The updated version of Figure 5 reflects these corrections and provides a more precise and informative visualization of the fault-handling workflow. We thank the reviewer again for helping improve the clarity and technical accuracy of this element.

 

Comments 2: On figures from 8 through 11 it is good to show the dimension of the original time so it can be easier for a person overviewing the article to know that the time savings are in minutes which is significant for HPPs. Also it's probably better if the vertical axis is called "Intervention time" instead of "Original time".

Response 2: We have revised the figures to explicitly indicate that the time values are expressed in minutes, which highlights the practical relevance of the time savings achieved in hydropower plant operations.

Additionally, we have renamed the vertical axis label from "Original time" to "Intervention time" to more accurately reflect the data being presented and avoid ambiguity. These changes enhance both the interpretability and professional presentation of the results.

Comments 3: It should be noted somewhere in the text when talking about the specific HPP what type of water turbine it uses - in this case a Pelton turbine. Therefore it will be nice to have in the conclusion a small comment about if the proposed solution can be implemented in HPPs with different type of water turbines or it can be used only for different HPPs with Pelton turbines.

Response 3: 

Introduction (penultimate paragraph): We have included a clarification specifying that the case study analyzes a horizontal-axis Pelton turbine with two injectors at the Alazán Hydroelectric Power Plant. This information is now presented explicitly to contextualize the technical focus of the model.

Conclusions (new paragraph added): A new paragraph has been added to explain that the proposed Petri net model can be generalized to other types of hydroelectric power plants, since the speed governor (Governor) is a core component present in all turbine types. The text highlights that, although the current implementation was developed for a Pelton turbine, the model can be adapted—with minor adjustments—for use with Francis, Kaplan, or other turbine systems.

 

In general, the manuscript has been reviewed and revised to correct language issues in English as well as inconsistencies in the reference formatting.

Reviewer 2 Report

Comments and Suggestions for Authors

1. Abstract: Although it is mentioned that “novel methodology” is proposed, the fundamental difference with the traditional methodology is not specified, and it is suggested to briefly mention the specific limitations of the traditional methodology and emphasize the innovations.
2. When describing the use of simulation modeling to assess the reliability and availability of hydropower plant systems and mentioning the advantages of Petri nets, it is appropriate to add some specific research cases or existing data of relevant research results. For example, it is illustrated how much Petri nets have specifically improved the fault detection accuracy in other similar hydropower systems or industrial systems, or the error range of system reliability assessment under the application of simulation modeling.
3. when presenting the background of the study, be more specific about the reasons for choosing the case of the Alazán hydroelectric power plant in Ecuador, highlighting its representativeness or unique challenges. Cite the official report data to visualize the problem of frequent failures and highlight the practical significance of the study.
4. The paper mentions the use of WoPeD software to design Petri nets and construct models based on system components and operational relationships, but does not explain in detail how to map the physical systems to the “repositories” and “transformations” of Petri nets. It is recommended that a short description of the mapping rules be added.
5. Table 3 lists common failures and their intervention times, but does not explain how these failures are categorized. It is recommended that the basis for categorization be stated.
6. the literature cited in the literature review spans from 2006 to 2023, and it is suggested to add some more recent literature.

Author Response

Comments 1: Abstract: Although it is mentioned that “novel methodology” is proposed, the fundamental difference with the traditional methodology is not specified, and it is suggested to briefly mention the specific limitations of the traditional methodology and emphasize the innovations.

Response 1: The abstract has been revised to explicitly clarify the limitations of traditional fault diagnosis methodologies, such as reliance on manual procedures, lack of formal modeling for concurrent system behavior, and dependence on expert interpretation. Additionally, the revised version emphasizes the core innovations of the proposed methodology—namely, the use of Petri nets to provide a structured, dynamic model of the Governor system, enabling more accurate and faster fault detection. These enhancements aim to improve clarity for readers regarding the contributions and practical advantages of our approach.

Comments 2: When describing the use of simulation modeling to assess the reliability and availability of hydropower plant systems and mentioning the advantages of Petri nets, it is appropriate to add some specific research cases or existing data of relevant research results. For example, it is illustrated how much Petri nets have specifically improved the fault detection accuracy in other similar hydropower systems or industrial systems, or the error range of system reliability assessment under the application of simulation modeling.

Response 2: We have revised the corresponding paragraph in the introduction to include recent and specific studies that illustrate the impact of Petri nets on reliability assessment and fault detection accuracy.

Specifically, we now reference the work of Mehdi et al. (2024), who applied colored timed Petri nets and Monte Carlo simulation to a mechatronic system and reported an MTBF of approximately 50,000 hours with component availability exceeding 98%. Additionally, Singh et al. (2025) developed a Petri-net–based RAM model for hydropower systems and quantified availability values between 95% and 97%, integrating fault tree and failure mode analysis to identify critical components and optimize maintenance intervals. We also include the case study by Melani et al. (2016), which demonstrated that the use of Petri nets reduced false positive rates in fault diagnosis by 30% and significantly shortened fault localization time compared to traditional SCADA-based systems.

Comments 3: When presenting the background of the study, be more specific about the reasons for choosing the case of the Alazán hydroelectric power plant in Ecuador, highlighting its representativeness or unique challenges. Cite the official report data to visualize the problem of frequent failures and highlight the practical significance of the study.

Response 3: We have expanded the background section to include the institutional history, strategic importance, and unique operational challenges of the Alazán power station. This contextual information clarifies the reasons for its selection as the case study. Additionally, we noted that due to confidentiality restrictions, the data used in the study was obtained from internal records and the operational experience of on-site technical staff.

Comments 4: The paper mentions the use of WoPeD software to design Petri nets and construct models based on system components and operational relationships, but does not explain in detail how to map the physical systems to the “repositories” and “transformations” of Petri nets. It is recommended that a short description of the mapping rules be added.

Response 4: We have added a detailed paragraph after the description of the WoPeD modeling process to explain the mapping rules used in the Petri net construction. This includes how physical components and operational conditions were represented as places, transitions, tokens, and arcs, based on their functional roles in the governor system. This addition clarifies how the physical system was systematically translated into the formal Petri net structure.

Comments 5: Table 3 lists common failures and their intervention times, but does not explain how these failures are categorized. It is recommended that the basis for categorization be stated.

Response 5: We have added a detailed explanation in the manuscript clarifying the basis for the failure categorization presented in Table 3. The classification was developed based on a combination of factors, including the type of component affected, the origin of the failure (electrical, mechanical, or control-related), the operational impact on the Governor system, and the estimated intervention time. For example, failures such as speed sensor malfunction were considered high priority due to their immediate impact on system stability, whereas other events like minor injector response deviations were treated as lower priority.

The categorization was constructed using field observations, historical maintenance records, and the technical judgment of operational and supervisory personnel. While not derived from a formal standard, the classification is consistent with common industrial practices that prioritize failure management based on safety, availability, and reliability criteria. This addition now appears directly after the introduction of Table 3 to enhance the clarity and contextual relevance of the data presented.

Comments 6: The literature cited in the literature review spans from 2006 to 2023, and it is suggested to add some more recent literature.

Response 6: We have incorporated two recent and relevant studies from 2024 into the literature review section:

Huang et al. (2024) proposed a novel fault localization method for hydroelectric units operating under conditions of limited fault samples. Their work introduced a hybrid deep learning framework (SG-WMBDL) combining sparse autoencoders, generative adversarial networks, wavelet-based noise reduction, and AdaBoost, demonstrating high fault detection accuracy using real-world sensor data.

Ge et al. (2024) presented a comprehensive analysis of Petri net applications in smart grids, focusing on fault detection, load balancing, and energy management. Their survey classified Petri net models used across various energy systems and discussed the potential and limitations of these models in real-world implementations.

 

In general, the manuscript has been reviewed and revised to correct language issues in English as well as inconsistencies in the reference formatting.

Reviewer 3 Report

Comments and Suggestions for Authors

This paper proposes a fault management method based on Petri nets, which is innovative and has the potential for practical application, but further improvements are needed.

1. The logical coherence between sections of the paper (introduction, methodology, results, discussion) is relatively loose. The authors need to enhance the logical flow.
2. The authors mention in the abstract that the method can "significantly reduce diagnostic and intervention times compared to traditional methods." How is this "significant reduction" demonstrated?
3. What does the parenthesis indicate in "speed governor (Governor)"? Why is "Governor" sometimes italicized and sometimes not? This is particularly concerning in the figures.
4. In the Abbreviations section, what do "ACEI Governor oil pressure" and "PACIE Governor oil pressure" mean? Are "AC AC" and "DC DC" meaningful?
5. The formulas need additional annotations. The reference format is inconsistent and needs further optimization.

Author Response

Comments 1: 

This paper proposes a fault management method based on Petri nets, which is innovative and has the potential for practical application, but further improvements are needed.

1. The logical coherence between sections of the paper (introduction, methodology, results, discussion) is relatively loose. The authors need to enhance the logical flow.

Response 1: We revised the transitions between major sections to enhance narrative coherence. Specifically, we added linking paragraphs at the end of the Introduction, Methodology, and Results sections to clearly guide the reader through the logical progression of the study. These additions emphasize how the objectives introduced are addressed through the modeling approach, how the simulations relate to the methodology, and how the results are interpreted in the Discussion. This restructuring ensures a more integrated and logically connected presentation of the study’s contribution.

Comments 2: The authors mention in the abstract that the method can "significantly reduce diagnostic and intervention times compared to traditional methods." How is this "significant reduction" demonstrated?

Response 2: We have clarified in the manuscript how the reduction in diagnostic and intervention times was demonstrated. Specifically, the proposed Petri net–based model was applied to simulate fault scenarios, and the resulting intervention times were compared to those recorded under conventional procedures used at the Alazán power station. The results showed an average time reduction of approximately 66% in diagnostic response and 38% in overall intervention time. Additionally, we revised the abstract to explicitly reference these quantitative results and provide a clearer basis for the claim of significant improvement.

Comments 3: What does the parenthesis indicate in "speed governor (Governor)"? Why is "Governor" sometimes italicized and sometimes not? This is particularly concerning in the figures.

Response 3: We have revised the manuscript to clarify the terminology and ensure consistency throughout. Specifically, we now introduce the term as "speed governor (hereafter referred to as the Governor)" upon its first appearance in the text. This clarifies that "Governor" is used subsequently as a concise technical reference to the speed regulation subsystem of the hydroelectric plant.

In addition, we have reviewed the entire manuscript, including all figures and tables, to apply a uniform formatting style for the term "Governor", using regular (non-italicized) font consistently. This revision ensures clarity and eliminates any ambiguity related to typographic variation.

Comments 4: In the Abbreviations section, what do "ACEI Governor oil pressure" and "PACIE Governor oil pressure" mean? Are "AC AC" and "DC DC" meaningful?

Response 4: "ACEI Governor oil pressure” and “PACIE Governor oil pressure” refer to distinct hydraulic pressure parameters within the governor system of the Alazán Hydroelectric Power Plant. Specifically, ACEI corresponds to the oil tank pressure, representing the base pressure available in the hydraulic reservoir. In contrast, PACIE refers to the operational working pressure, which indicates whether the pressure is within the expected functional range of 3.4 to 4.0 MPa. These terms are part of the plant’s internal instrumentation terminology and are used for monitoring different aspects of the governor’s hydraulic subsystem. The manuscript has been updated to clarify these definitions in the Abbreviations section and in the corresponding technical descriptions.

Additionally, the entries “AC AC” and “DC DC” were formatting errors. They have been corrected.

Comments 5: The formulas need additional annotations. The reference format is inconsistent and needs further optimization.

Response 5: 

As the manuscript contains a single formula related to the percentage improvement in intervention time, we have revised it to include a more detailed description of each term, including the unit of measurement and contextual relevance. This ensures clarity for the reader when interpreting the performance comparison. Additionally, we have reviewed all references to ensure formatting consistency throughout the manuscript, in accordance with the journal’s guidelines.

 

In general, the manuscript has been reviewed and revised to correct language issues in English as well as inconsistencies in the reference formatting.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

I have no further comments, the paper is acceptable.