Model Predictive Energy-Maximising Tracking Control for a Wavestar-Prototype Wave Energy Converter
Round 1
Reviewer 1 Report
This paper deals with the control of a wave energy converter, Wavestar, that is a subject, control, key for wave energy conversion. The problem is mathematically well designed and consistent. The applied tools are already well known for estimation, optimization and control problems, nevertheless their application to the current framework is very interesting. The main issue of the current work is addressed in the conclusions, robustness and practical application of the current proposed method. It should be noted that the Wavestar converter has several floaters that interact with each other at least hydrodynamically by means of the radiation forces. This would introduce a considerable burden on the real time needed computations now with several degrees of freedom.
Some minor editorial issues, on page 4 line 133 it should be “chosen”, on page 12 line 299 it should be “incident”.
After this minor editorial issues (and perhaps others that I didn´t see) I recommend the paper to be published.
Author Response
Please see the response in the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper is very good and well written. I have only two suggestions:
1) I found some typing mistakes, mainly on the first 3 pages of the text. This is easy to correct.
2) Your results seems to refer to one only irregular wave. Would it be useful to evaluate other waves ?
The paper is well written. Only some typing mistakes that can be easily corrected.
Author Response
Please see the response in the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The manuscript deserves a more accurate review. I was given a week and, after asking for more days, an extension of 2 days. This paper needs a much longer review. Also, this page is written hastily, completed in the evening, sorry if it is unclear, try to understand the concepts.
What I can say on the positive side is that:
- the manuscript is promising, interesting, focused on an important topic and sufficiently novel
- the methods are adequate and sound,
- the language is appropriate (only a few typos and some unclear sentences pointed out below).
- The introduction is well done (even if a bit biased in favour of Ringwood investigation)
On the negative side:
- The methods are described too quickly. Too many steps are jumped, it is really difficult to follow the text. I will give below some specific comments
- Description of the experiments is not sufficient (instrumentation used, programme, water depth, how are the measurements taken, logging frequency, filters, how the incident wave is analysed, characteristics of the model,…) you can do better I think.
- The conclusions are drawn on the basis of a single test, which is maybe not sufficient to prove the method (but I know that you are not going to add other analyses...)
- The benefit of the proposed methods is not sufficiently pointed out with comparisons with other approaches. Why not try to point them out?
OTHER COMMENTS
Unclear sentences:
- Reference to high level part in the abstract is unclear. Also, 2 models are, not is.
- You should better explain why you need the EKF to estimate the amplitude of Mex
- Line 74. Besides... the sentence seems incomplete. It is not clear to me
- The concept in line 292, page 11 is not clear/sufficiently supported
- The whole part from line 293 to line 308 should be revised. It also contains some typos: “inclident”; “.. is lack of some robustness”. Etc.
- Conclusions are not fully representative of the work done, and do not clarify the hybrid method used. And some typos (smothess), with reference again to high level/low level (is it important? The casual reader does not understand), The two Gaussian models are presented with no clear purpose. In general, it does not describe well what is done.
Other comments:
Section 2.1 I do not see the 3 "independent" motions and I do not understand what has these 3 "independent" motions. It looks there is only 1 degree of freedom.
The picture shows 2 floating elements. Can you elaborate?
"inconvenient" covolution term? I suggest a different adjective, e.g. problematic. Then add the definition of ra. Actually, if you report the Cummins equation, it will be clearer what you are doing with the state space equation. Explain that ra has 2 variables. Add in correspondence of the Ara matrix the symbol “2x2”, and in Cra “1x2”.
Explain why you call the blocks “High level” and “Low Level”
Kalman filter with random walk. It is difficult for me to follow the explanation: it would be clearer if you could point out the analogy with the typical Kalman problem, explain the prediction step, the adopted gaussian process model, show how the prediction step is derived, point out the transition matrix (Ad), the innovation term, y(k)-C*x(k|k-1), that is important since it is the part that captures the new information in y(k), and Kf(k) explains how much we trust in the innovation. What is k+1|k instead of k|k-1. In short, make the proper analogy with the Kalman Filter update step, without logical jumps. You pretend too much from the reader.
Results. I would appreciate a more comprehensive description of the steps done, the different terms of the equations, etc.
I have other comments on other parts but, unfortunately, I cannot really put them down without clarifying my own ideas, and I have no more time. Next round of review, maybe? this is why I suggest a major review: because I will be able to see again the manuscript and try to give other comments. I think they are needed to allow more people to understand what you are proposing (frankly, I think not too many at the moment).
see above
Author Response
Please see the response in the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
I enjoyed the reading, good job.
