Convex Optimisation Model for Ship Speed Profile: Optimisation under Fixed Schedule

Round 1

Reviewer 1 Report

Please read carefully the attached pdf file, which contains all the recommendations.

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

Thank you for the thorough review of our manuscript. Below, you can find a point-by-point answer (bolded) to your comments and suggestions. We also indicated the changes made to the revised manuscript by line numbers where applicable. We also included the revised manuscript in PDF format, where changes are highlighted.

Preamble

I have accepted enthusiastically the review of the paper proposed by the editor for only the reason: this month, thirty seven years ago, I submitted a BEng thesis entitled “Ship optimisation speed from the ship efficiency point of view”. When I have accepted the task I have been extremely interested on how far the achievements in the field went since.

Disappointingly, reading the paper for review, I have found that after almost four decades the progress is very little…

General Remarks

1. Needless to say that authors of the paper propose a challenging topic that is of a wide interest in the field of ship operation;

2. In spite of the serious limitations imposed by the fixed schedule restrictions, the paper approached only gently the actual problem of reducing the ship emissions. No final remarks of the submitted paper conclude the savings in the carbon monoxide and dioxide emissions when applying the proposed investigation. The authors are only concerned with the fuel savings, which is actually not the scope of EEDI;

Answer: Emissions and their reduction with the proposed method would depend largely on the used fuel, engine and on any emission reduction technologies installed aboard. Since these were not in the scope of the research, we feel that discussing fuel consumption reduction as a proxy to emission reduction is justified. Similar logic is discussed in reference [9] in the manuscript. However, we did add calculation of CO2 emission reduction to the results section (Lines 484-486 in the revised manuscript) to act as an example of the magnitude of emission reductions. In the context of emission regulations, we feel that the proposed research would be more targeted towards fulfilling the future CII tiers, rather than EEDI. With CII as the frame of reference, the governed CO2 emissions and their reduction potential can be quite easily discerned from the reported fuel savings. We have added new content to lines 161-166 in the revised manuscript that explains this reasoning in the manuscript.

3. Although the authors avoided to motivate explicitly in the abstract the choice of the subject, the paper foresees a numerical description of the efficacy rather than to the ship efficiency. From this point of view I do wonder whether a ship owner, who is always a business-oriented person, who deals with crew salaries, port waiting and taxis, operation delays and so one may appreciate the overall usefulness of the scientific work reported in here;

Answer: The ship owner perspective was briefly explored in the discussion section. It is true that the ship owner may be more concerned with maximizing the time charter equivalent (TCE) of a vessel than minimizing the fuel consumption with the proposed method. The TCE problem would for example allow a flexible schedule for the voyage. However, TCE maximization forms a different research problem which we were not focusing on in this manuscript. This is definitely a subject for further research though. Furthermore, paying attention to fuel consumption is expected to become more important in the future due to the regulations getting more stringent. Computational efficiency of the developed method ultimately translates to savings accrued by the party providing the service of speed profile optimization. Furthermore, it is critical in case speed profile optimization is used a pre-requisite for subsequent calculations, which was one of the aims for this study. We added content to the introduction on lines 157-161 and discussion on lines 534-537 that should explain the motivation for this study further.

4. In my particular opinion, I may take this study as a first step of further researches that may probably need to be taken into consideration;

Answer: Indeed, this was our first speed optimization study and we plan to extend research in this field. The natural follow-up topic would be to assess speed profile optimization with a flexible schedule, as per your suggestion. Such a study would allow us to explore the topics of your previous comment. We added a section on this into the discussion (lines 591-595 in the revised manuscript).

5. The ship cruise regards multiple journeys, for different ships working under different navigation conditions, from Houston, TX, to London, UK. I would suggest the authors to mention the correct name of the destination port, i.e. London Gateway. This for avoiding any confusion, since London has more than one port and the reader may get confused;

Answer: Great suggestion. We specified the destination port more clearly in each occurrence where route is mentioned.

6. The Introduction chapter is fairly poor and should therefore it must be necessarily improved. Some similar papers dealing with the same subject should probably referenced as well;

Answer: We improved the introduction with multiple additions, including a more comprehensive literature review of similar topics. (Lines 27-29, 92-99, 114-132 and 157-166)

7. Lastly, I suggest the authors to slightly change the title of the paper from: Convex Optimisation Model for Ship Speed Profile Optimisation Under Fixed Schedule to: Convex Optimisation Model for Ship Speed Profile: Optimisation under Fixed Schedule

Answer: We agree that this is a more suitable title and changed the title to “Convex Optimization Model for Ship Speed Profile: Optimisation Under Fixed Schedule” as per your suggestion.

Particular Technical Remarks

1) The Harvald method of computing the ship resistance, which is already 38 years old is out of fashion;

2) Harvald method may only be used when certain geometric conditions are fulfilled i.e. main dimensions ratios such as L/B, B/t, or the hull coefficients such as CB or CP and others such as the longitudinal center of buoyancy position, or the geometry of the stern overhang. Indeed the Harvald method was conceived for cruiser sterns and empirical corrections (not fully confirmed!) were proposed as a remedy for this restriction. I therefore strongly suggest the authors to check once again whether or not they were under the assumptions of the method employed in their study;

Answer: Great remarks and definitely information that is good to know for the future. We would like to highlight that the main focus of this study was on developing the optimization framework for efficient and high-resolution calculation of the speed profile. For this reason, precise modelling of ship resistance was not the main focus. Furthermore, it is expected that the parties which might see value in the developed methods, i.e. voyage optimization solver providers, already employ accurate resistance models. On the topic of calm water resistance: various corrections in the residual resistance coefficient still result in a convex calm water resistance format. Thus, such additions would not interfere with the methodology as it was laid out in the manuscript. We did take hull properties into account when choosing the residual resistance coefficient. We added new content on this topic into the discussion chapter on lines 580-590 in the revised manuscript.

3) Dealing with a strong non-linear optimisation problem with restrictions, the authors completely miss to explicitly mention the restrictions for the considered flow cases, which is unquestionably unacceptable for such a numerical simulation nowadays;

Answer: We were left somewhat puzzled in regard to what was meant specifically by flow cases. If this refers to the considered resistance model, we believe that our previous answer regarding Harvald’s method, and related revisions made, apply to this comment as well. The wave resistance model is adjusted to different Froude number regions and this logic is embedded into the model itself. We would refer to reference [33] in the revised manuscript for more details on this. In case this comment refers to dangerous wave conditions, such as wave surfing or parametric rolling, these fall into the same category of conditions as very high wave heights. Operation in such conditions was not restricted in this study, although they could be restricted quite easily as discussed in the discussion section. We added a comment about these conditions to lines 561-562 in the revised manuscript. As for the constraints related to the convex optimization problem, we have presented all the constraints enforced in the optimization problem, and there are no relevant constraints left out either from the paper or the model itself, to the best of our knowledge.

4) The paper fails in convincing the reader that the overall accuracy of the solver is sustained by the V&V calculations.

Answer: We elaborated more clearly how the fuel consumption values were calculated on lines 443-448 in the revised manuscript. All of the fuel consumption values were calculated with the model described in chapter 2.1, using the actual conditions encountered with the selected speed profile. If this comment was in particular targeted towards the accuracy of the convex optimisation solver, we can say with confidence that the produced result is the global optimum of the declared optimisation problem. The used solver was ECOS, which can be considered reliable. The used solver is also now indicated on line 467 in the revised manuscript. ECOS is an interior point solver, which computes the optimal solution with a deterministic stopping criteria and with prescribed level of accuracy. Unless the problem is infeasible or unbounded, the iterative algorithm is guaranteed to converge to a solution. Due to the convex structure of the speed optimization problem, a local solution is also the global solution.

5) Several statements are redundant, which is unacceptable. They all should be revised;

Answer: We removed multiple redundant statements. Changes were made in particular to sections of the revised manuscript on lines 152, 175-181, 193, 379, 416-417 and 478.

6) Without a further deep discussion on the correctness of the solutions, based only on the items above mentioned, I do consider that the paper has to be substantially revised for being further considered for publication.

Linguistics remarks

English language is rather good and should be accepted only after a more careful proofreading. Some mistyping errors such as:

- P3, L120: Hydro-meteorological instead of hydrometeorological. The “hydrometeorogical” noun/adjective obsessively appears 37(!!) more times in the paper, a leitmotiv which might be tiresome for the reader. English language is very generous, therefore I suggest the authors to rephrase the text, wherever necessary, to avoid such inadequate repetitions for a scientific way of writing;

- P3, L121: NAPA Voyage optimisation vs NAPA Voyage Optimisation (P3, L122);

- P4, L173; P6, L214; P6, L219; P15, L464: ie. instead of i.e.;

- P23, Ref.17: Ocean Engineering instead of Ocean engineering; and so on, have to be corrected

Answer: Thank you for the careful assessment of the manuscript. We corrected the suggested linguistic remarks and made several other corrections as well. The comment on the term “hydrometeorological” is reasonable. We changed multiple occurrences of this term to avoid repetition. NAPA Voyage Optimization fixed to be capitalized due to it being the name of a product.

 

Reviewer 2 Report

The paper presents a novel convex optimisation model for ship speed profile optimisation under

varying environmental conditions, with a fixed schedule for the journey. The authors demonstrate the

efficacy of the proposed method, a combined speed profile optimisation model was developed and

 tested with 5 different ships for 20 journeys. Authors claim the fuel savings over speed profile optimisation 1.1 % for all of the studied journeys and 3.5 % on average for all journeys.

 

 

Some minor clarifications are needed and some recommendations for improvements are given below.

 

1. Line 145: Therefor => Therefore
2. Line 184, for Cf it is recommended to add ITTC reference
3. Table 3: It is unclear how the design speed for the selected ships is unknown. For the “Area opposed to wind” it is more reasonable that the data is missing but also not quite understandable. If one is using ships in the calculation he should take the ships with all known data.
4. Figure 8: For which ship the speed profile is given? This type of figure will be interested to be given for all considered ships.
5. It is suggested to couple figures 7 and 8.

 

Author Response

Dear reviewer,

Thank you for the thorough review of our manuscript. Below, you can find a point-by-point answer (bolded) to your comments and suggestions. We also indicated the changes made to the revised manuscript by line numbers where applicable. We also included the revised manuscript in PDF format, where changes are highlighted.

Some minor clarifications are needed and some recommendations for improvements are given below.

1. Line 145: Therefor => Therefore

Answer: We corrected the grammatical error.

1. Line 184, for Cf it is recommended to add ITTC reference

Answer: We added the ITTC reference for Cf (Lines 209-210 in the revised manuscript).

1. Table 3: It is unclear how the design speed for the selected ships is unknown. For the “Area opposed to wind” it is more reasonable that the data is missing but also not quite understandable. If one is using ships in the calculation he should take the ships with all known data.

Answer: The vessel data was gathered from sources that used the vessels for resistance calculations in previous studies. These studies did not explicitly publish all related data of the vessels, but rather used the data points that the authors needed in their work. Design speeds were not indicated in any of these studies. However, reference [43] in the manuscript was used to approximate all of the design speeds, which we consider to be a good approximation. It was deemed that including the ship in the analysis with the approximation of design speed was more beneficial to the research than leaving the ship out of consideration.

1. Figure 8: For which ship the speed profile is given? This type of figure will be interested to be given for all considered ships.

Answer: Good suggestion. We decided to compile the speed profiles for all of the vessels into the same figure.

1. It is suggested to couple figures 7 and 8.

Answer: This suggestion makes sense. However, since we modified figure 8 to include the speed profiles of all the vessels, the contents of figures 7 and 8 are no longer strictly linked. Furthermore, Figure 8 became quite large on its own. For these reasons, we decided to keep the figures separate.

 

Round 2

Reviewer 1 Report

The paper is significantly improved from the initial version of the manuscript. Thank the authors for the careful consideration of my comments.