Next Article in Journal
Web Site Fingerprint Attack Generation Technology Combined with Genetic Algorithm
Next Article in Special Issue
Wireless Device with Energy Management for Closed-Loop Deep Brain Stimulation (CLDBS)
Previous Article in Journal
Friction Feedforward Compensation Composite Control of Continuous Rotary Motor with Sliding Mode Variable Structure Based on an Improved Power Reaching Law
Previous Article in Special Issue
An Artificial Neural Network for Solar Energy Prediction and Control Using Jaya-SMC
 
 
Article
Peer-Review Record

Intelligent Approaches to Fault Detection and Diagnosis in District Heating: Current Trends, Challenges, and Opportunities

Electronics 2023, 12(6), 1448; https://doi.org/10.3390/electronics12061448
by Jonne van Dreven 1,2,3,*, Veselka Boeva 1, Shahrooz Abghari 1, HÃ¥kan Grahn 1, Jad Al Koussa 2,3 and Emilia Motoasca 2,3
Reviewer 1:
Reviewer 2:
Reviewer 3:
Reviewer 4: Anonymous
Electronics 2023, 12(6), 1448; https://doi.org/10.3390/electronics12061448
Submission received: 1 February 2023 / Revised: 4 March 2023 / Accepted: 13 March 2023 / Published: 18 March 2023
(This article belongs to the Special Issue Smart Energy Systems Using AI and IoT Solutions)

Round 1

Reviewer 1 Report

The article is well presented, I accept it as it is

Author Response

Dear reviewer, thank you for your time and review. We appreciate your feedback!

Reviewer 2 Report

The authors  presented a state of the arts methods for fault detection and diagnosis in district heating. 

 

The explanation of existing  works in terms of sentences is not a review or state of the arts,  you have to discuss  various techniques as well  as the current  requirements of the district heating

There in no new contribution  in terms   experimental,  technology,  So it can not be treated  as a new work.

Author Response

Dear reviewer,

Thank you for your time and review.

We have provided a literature survey. A literature survey is a comprehensive review of the literature on a particular topic to identify gaps, trends, and research directions, unlike a (systematic) literature review which is a more rigorous approach to a specific research question.

While we “just explained what other researchers do in previous research”. It is the exact purpose of a literature survey and was our aim, i.e., to summarise past work and include an analysis to identify trends, gaps, and research directions, which currently is missing in fault detection and diagnosis in district heating.

While “there is no new contribution in terms of experimental or technology”, it is not the aim and purpose of our literature survey to contribute in terms of experimental or technology. However, our literature survey contributes to knowledge in the specific field through a comprehensive summary of the topic. Additionally, we provide new work by including a SWOT analysis to highlight and clearly present the field's challenges, opportunities, and limitations (both minor or major limitations for machine learning). We provide a clear overview of current trends, as seen in, for example, Figure 1, to discuss the advantages and disadvantages and point out research directions with recommendations.

Reviewer 3 Report

This manuscript presents a comprehensive survey of state-of-the-art intelligent fault detection and diagnosis in district heating systems. An in-depth analysis of FDD in DH is presented by giving a survey in two aspects: Fault Detection and Fault Diagnosis, which are the first two steps of the automatic FDD process. At the end of the manuscript, you highlight the advantages and disadvantages of the techniques, as well as the recent trends and key challenges in FDD. Furthermore, you give suggestions for future research. However, there are some problems, such as spelling and grammar, etc. Detailed comments are listed as follows. Hope they can provide help in improving the quality of this paper.

1) It is noted that your manuscript needs careful editing because several mistakes occur in it. You should pay particular attention to English grammar, spelling, and sentence structure so that the details of the study are clear to the reader. Some of your mistakes are as follows:

(1) Some verbs require the singular form but some do not. For example, in line 29, you should use ‘consists’ instead of ‘consist’. Similar problems also occur in lines 564, 615, 837, 856, 872, 908, etc.

(2) Some prepositions in some phrases are wrong. For example, in line 262, you should use ‘a certain period of time’ instead of ‘a certain of time period’. Similar problems also occur in lines 794, 814, 925, etc.

(3) The spelling errors appear in lines 354, 905, 1015, etc.

(4) You need to use the Definite Article as the following example ‘an MST’. Similar problems also occur in lines 422, 499, 503, 539, 757, 778, 785, 850, 852, 872, 992, etc. Some other mistakes like ‘an two-stage’ also occur in lines 469, 1034, 1055, etc.

(5) Wrong pronouns are used in lines 596, 1202, etc.

2) You should pay attention to the usage of punctuation marks in your manuscript. In lines 35 and 130, the sentence ends without a full stop. Before submitting a revision be sure that your material is properly prepared and formatted.

3) In line 237, the ‘Fault diagnosis is divided into three subcategories’, I think you just divide it into two subcategories.

4)  The following references are recommended to be added.

(1) Jian Yu, Yu Wen, Lei Yang*, Zhibin Zhao**, Yanjie Guo, Xiao Guo, Monitoring on Triboelectric Nanogenerator and Deep Learning Method, Nano Energy, 2022, 92, 106698.

(2) Yanjie Guo, Jiafeng Tang, Lei Yang*, Zhibin Zhao**, Miao Wang, Peng Shi. RobustFlow: An unsupervised paradigm toward real-world wear detection and segmentation with normalizing flow, Tribology International, 2022, 179, 108173.

(3) Miao Wang, Lei Yang*, Zhibin Zhao, Yanjie Guo, Intelligent prediction of wear location and mechanism using image identification based on improved Faster R-CNN model, Tribology International, 2022, 169, 107466.

Author Response

Dear reviewer, thank you for your in-depth and helpful feedback. We have carefully addressed all points which are specified below:

1) mistakes

   1) 

    • Line 29: changed “consist” to “consists”.
    • Line 564: changed “proposes” to “propose”.
    • Line 615: changed “presents” to “present”.
    • Line 837: changed “evaluate” to “evaluates”.
    • Line 856: changed “and” to “,”.
    • Line 872: changed “proposes” to “propose”.
    • Line 908: changed “presents” to “present”.

 

   2)

    • Line 262: changed to “certain period of time”
    • Line 794: no change
    • Line 814: changed “with” to “by”
    • Line 925: changed “lacks” to “lags”

 

   3)

    • Line 354: no change.
    • Line 905: changed to “strengths”.
    • Line 1015: changed to “solely”.

 

   4)

    • Line 422: changed to “an”.
    • Line 469: changed to “a”.
    • Line 499: changed to “an”.
    • Line 503: changed to “an”.
    • Line 539: changed to “an”.
    • Line 757: changed to “an”.
    • Line 778: changed to “an”.
    • Line 785: changed to “an”.
    • Line 850: changed to “an”.
    • Line 852: changed to “an”.
    • Line 872: changed to “an”.
    • Line 992: changed to “an”.
    • Line 1034: changed to “an”.
    • Line 1055: changed to “an”.

 

   5)

    • Line 596: added “a circulate pump breakdown” and changed “they” to “the approach”.
    • Line 1202: changed “they” to “its"

 

 

 

2) punctuation

  • Line 35: added “.”
  • Line 130: added “.”

 

3) Mistake

  • Line 237: changed to “two”

4) Recommended references

We have considered the suggested references; however, as our survey focuses on literature related to district heating, we have not added them.

 

Additionally, we have carefully reread the manuscript and made some spelling changes seen below.

Other spelling mistakes/changes:

  • Line 75: added “pipes”.
  • Line 81: added abbreviation for domestic hot water (DHW).
  • Line 115: removed “thus” for clarity.
  • Line 148: capitalized “deep”.
  • Line 190: changed “solution” to “solutions”.
  • Line 215: added abbreviation HVAC.
  • Line 220: changed “till” to “until”.
  • Line 255: removed “it” for clarity.
  • Line 262: added spacing and “,”.
  • Line 500: removed “and” and separated sentence for clarity: “Observations which fall outside a specified threshold are considered outliers.”
  • Line 742: changed “types” to “types”.
  • Line 910: added “,” after 7.4.
  • Line 920: added “,” after “reveals”.
  • Line 926: changed “for” to “to”.
  • Line 941: removed “,” after “issues”.
  • Line 944: removed “moreover”
  • Line 945: changed “contributes to” to “improve”.
  • Line 1027: removed improper spacing.
  • Line 1052: removed improper “,”.
  • Line 1082: removed improper “,”.
  • Line 1104 and 1110: removed “the”.
  • Line 1125: changed “to utilize” to “utilizing”.
  • Line 1180 and 1185: added “,”.
  • Changed all “dataset” to “data set” for consistency.

Reviewer 4 Report

Very good article. Literature review done with really good care. Of course, not all relevant publications have been quoted, but this is not possible in real terms. Probably other reviewers may accuse the lack of reference to any work, but in my opinion you always have to find the end point and in my opinion this work meets the criteria of a review! It should be accepted for publication! Congratulations to the whole team.

Author Response

Dear reviewer, thank you for your time and review. We very much appreciate your kind words! Thank you.

Round 2

Reviewer 2 Report

I am not satisfied with the author response. when writing a literature review, we have considered all the prospective future application.

Most important part is data acquisition, which is missing in the manuscript, How these data is ac quired in various operating condition,

How environmental factor can affect the diagnosis process.

What are the major challenges faced by district heating operators when it comes to fault detection and diagnosis, and how can intelligent approaches help address these challenges? 

Machine learning and deep learning are the black box, how authors are relating the physics of component failure with the current machine learning or intelligent methods.

Authors can write some of the key opportunities for improving fault detection and diagnosis in district heating systems, and how can operators take advantage of these opportunities to improve system reliability and efficiency?

What are the most promising research directions in the field of intelligent fault detection and diagnosis in district heating systems.

 

The novelty of the paper should be better stressed. It is not easy to understand what has been really achieved and what is new with respect to literature.

Author Response

Dear Reviewer, 

We would like to thank you for providing valuable feedback allowing us to improve on our manuscript. 

We have been able to incorporate changes to reflect your comments. Please find below our detailed response to the comments. We have included a second manuscript file which represents the changes in red. 

Most important part is data acquisition, which is missing in the manuscript, How these data is ac quired in various operating condition, 

We fully agree with the reviewer's comment. Data acquisition is one of the most important aspects for FDD in any domain including DH. This is the reason why we have devoted an entire section to the current data collection process (Section 5: District Heating Data Collection). As the focus of our study is mainly on the customers side at the building level, we have extensively described the data acquisition on the demand side, as this data is primarily used for intelligent fault detection and diagnosis in our survey. With respect to your comments, we have expanded the section with an additional paragraph on data acquisition on the DH distribution network to cover the faults concerning the supply side, e.g., leakages. 

 

How environmental factor can affect the diagnosis process. 

Thank you for bringing up this to our attention. We have added a brief discussion on it in Section 7: Discussion, Subsection 7.3 Weaknesses (see Recommendation 3) explaining issues regarding lack of labeled data and how environmental factors can help the diagnosis process.  

  

What are the major challenges faced by district heating operators when it comes to fault detection and diagnosis, and how can intelligent approaches help address these challenges?  

We addressed the major challenges for operators in the DH domain in Section 1: Introduction. We have added additional text in the third paragraph to provide a more comprehensive view for the readers. 

  

Machine learning and deep learning are the black box, how authors are relating the physics of component failure with the current machine earning or intelligent methods. 

Physics of component failure is important in the labelling process of data. We mention the shortcomings of simulations and propose that emulations may hold some benefits by inducing physical faults, which help in generating realistic faulty data. 

Machine learning approaches, on the other hand, are data-driven, they try to learn underlying physics/component failure and extract hidden patterns using historical data, without preliminary knowledge about the physics nature of the studied phenomenon. Although, the performance of the ML model is highly related to the quality of the input of the data.  

In addition, in ML and DL research, many efforts are recently made in the direction of increasing the explainability and interpretability of the provided solutions. We have expanded on this topic in Section 2: Background, Subsection 3.2: Machine Learning, iv) deep learning.

We have extended the topic on predictive maintenance further in Section 7: Discussion, Subsection 7.1 Strengths, to refer to the role of physics of component failure in developing FDD models. Additionally, we added an emerging topic called physics-based ML. 

 

Authors can write some of the key opportunities for improving fault detection and diagnosis in district heating systems, and how can operators take advantage of these opportunities to improve system reliability and efficiency? 

We have presented key opportunities in Section 7: Discussion. We have provided a SWOT analysis to present a clear overview of opportunities but also severe limitations or strengths in the DH field for researchers in developing. intelligent FDD. While operators can take advantage of the solutions developed by researchers, they are not the main focus of our work, as operators are the end users. The opportunities refer to ML directions, thus assisting researchers and practitioners in developing better solutions with current constraints to support the end users in decision-making. 

 

how can operators take advantage of these opportunities to improve 

Operators can take advantage of intelligent approaches, as they assist in detecting faults early, such that corrective measures can be taken. We describe some of the benefits in Section 1: Introduction. However, our primary focus is on bringing the researchers' and stakeholders' attention to challenges in the DH domain and the possible solutions to improve different aspects, such as data collection, which can lead to improvements in intelligent FDD approaches using ML and AI. 

 

What are the most promising research directions in the field of intelligent fault detection and diagnosis in district heating systems. 

We present and discuss the most promising research directions in Section 7: Discussion. These are analyzed using the SWOT analysis and provide detailed recommendations (1-8), see the bold text in the same section. In addition, we summarize our findings in Section 8: Conclusion. We have added those as bullet points to make it more clear. Research directions cover both the short terms (what can we do now) as well as the long terms (directions which take more time). 

  

The novelty of the paper should be better stressed. It is not easy to understand what has been really achieved and what is new with respect to literature. 

We describe our contributions in Section 1: Introduction. In Section 3: Related Work, we additionally discuss several reviews and their limitations. To further clarify the novelty and contribution of our work we have added our contributions as bullet points in the Introduction. 

Round 3

Reviewer 2 Report

no further comments

 

 

Back to TopTop