Assessment of Wave Energy Converters Based on Historical Data from a Given Point in the Sea

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article is very important for practical tasks and applications related to wave energy. However, I found significant disadvantages, which does not allow me to accept the work. I propose a major revision, but the authors will have to completely recalculate the results. Otherwise, I will have to reject the article.

Major comments:

1. What are the conceptually new results in this article compared to your previous work? (Prediction of Wave Energy Transformation Capability in Isolated Islands by Using the Monte Carlo Method Deivis Avila et al, 2021; An approach for evaluating the stochastic behaviour of wave energy converters Deivis Avila et al., 2022)

2. Please Specify where you got the technical characteristics of the power generators (Aqua Buoy, Oyster, Pelamis, SSG and Wave Dragoon) which showed on fig 3. 

For the different type of power generators we need to use different wave periods (mean, tm1, peak ...) to get a goog power calculation. Why do you use only peak period? Please add links to technical documentation where we can find this information

3. "significant spectral height" - I do not know this term. You mean significant wave height? if no, please show the formula to describe this term.

4. About "forecast" and "forecasting" terms. We know wave analysis or reanalysis - it is a information about wave parameters in the past. We know the wave forecast - it is a wave parameters in the future (3-7 days), like a weather forecast. I think in your case you calculate "assessments of available wave energy or power"  or "the Potential of Wave Power" I can not see the forecast. 

5. There is no statement of the problem in the introduction. We have archived measurement data, or wave reanalysis data. Why is it impossible to calculate the retrospective wave energy of power for each generator for every moment of time (30 min)? Then you can get assessments to differents seasons or month.  May we will get a big errors in this case? Why the weibul distribution (which initially contains an error, since some storms do not fall into this distribution purely mathematically) is better for this task? Please add clear justification with references.

May be you can find something useful In "Cavaleri et al. (2007) “Wave Modelling - The State of the Art”, 

6. Fig4 and fig 6. "Increments in the wave heights are especially evident in the summer months (July and August)". I am sorry, but it is not true. The storm period for this region is october-march. I see it in wave reanalysis and in your paper "Prediction of Wave Energy Transformation Capability in Isolated Islands by Using the Monte Carlo Method Deivis Avila et al, 2021" where you used the buoy 2442 which in 50 km from 2446.  I think some error exists in your calculations. Please see the attach. Unfortunately, you need to recalculate everything, I'm sorry.

It is the most important comment that you needs to correct.

7. Please add the figure with real SWH from buoy. for 1-2 years enough

 

Minor commetns:

1. Please add to Keywords: wave buoy data,  Canary Islands

2. "The fourth step (Figure 1-iv), consists of fitting polynomial regressions for modelling the five coefficients of the bivariate Weibull distributions as a function of the day of the year (Figure 1-v - check this)

3. "more chaotic  and agitated swell" wind waves and swell. I think it is not correct to say about  chaotic  swell. Mixed Sea - mixed wind waves and swell can be look out as chaotic.

4. Seawatch?

5. Figure 2. Geographic location of the buoy in question ?? check last word.

please add bathymetry isolines lables.

6. Please do not say about the "Gran Canaria Buoy (2442)" which you not used. this only confuses the reader.

7. "significant spectral height, H, were recorded each hour, computed for a period of 24 min, from 1998 to 2017. The measurement accuracy was ±0.05 m for H and ±0.05 s for T." Please add the link to technical description of a SeaWatch buoy where we can read that it measured "significant spectral height" with accuracy was ±0.05 m.

 

 

Comments for author File: Comments.pdf

Author Response

Referee 1

 

Dear reviewer,

Thank you for your useful comments and suggestions about our paper. We have modified the manuscript accordingly, and the detailed corrections are listed below point by point.

 

Comments and Suggestions for Authors

This article is very important for practical tasks and applications related to wave energy. However, I found significant disadvantages, which does not allow me to accept the work. I propose a major revision, but the authors will have to completely recalculate the results. Otherwise, I will have to reject the article.

A complete re-structuration of the paper was carried out, including a major change of the approach and the recalculation of the case of studies. Three different buoy locations were considered.

Major comments:

1. What are the conceptually new results in this article compared to your previous work? (Prediction of Wave Energy Transformation Capability in Isolated Islands by Using the Monte Carlo Method Deivis Avila et al, 2021; An approach for evaluating the stochastic behaviour of wave energy converters Deivis Avila et al., 2022)

The paper was completely rewritten, addressing a new simple straightforward approach for assessing wave energy converters. Gaussian mixed models were introduced for dealing with the complex shapes of the wave parameters distributions. The modelling of the distribution versus the day of the year was removed for simplifying the approach.

2. Please Specify where you got the technical characteristics of the power generators (Aqua Buoy, Oyster, Pelamis, SSG and Wave Dragoon) which showed on fig 3. 

For the different type of power generators we need to use different wave periods (mean, tm1, peak ...) to get a goog power calculation. Why do you use only peak period? Please add links to technical documentation where we can find this information

The source of the information on power converters was included. Conversion between the different wave parameters used for evaluated them, was also included.

3. "significant spectral height" - I do not know this term. You mean significant wave height? if no, please show the formula to describe this term.

The term “significant spectral height” was replaced by “significant wave height” through the entire paper.

4. About "forecast" and "forecasting" terms. We know wave analysis or reanalysis - it is a information about wave parameters in the past. We know the wave forecast - it is a wave parameters in the future (3-7 days), like a weather forecast. I think in your case you calculate "assessments of available wave energy or power" or "the Potential of Wave Power" I can not see the forecast. 

The paper was re-focused to assessment rather than forecasting.

5. There is no statement of the problem in the introduction. We have archived measurement data, or wave reanalysis data. Why is it impossible to calculate the retrospective wave energy of power for each generator for every moment of time (30 min)? Then you can get assessments to differents seasons or month.  May we will get a big errors in this case? Why the weibul distribution (which initially contains an error, since some storms do not fall into this distribution purely mathematically) is better for this task? Please add clear justification with references.

May be you can find something useful In "Cavaleri et al. (2007) “Wave Modelling - The State of the Art”, 

The paper now proposes a new approach based on determining the wave parameters distribution for the assessment period. The use of the bivariate Weibull distribution was eliminated to avoid assuming a certain shape of the probability distribution. Proper references were included.

6. Fig4 and fig 6. "Increments in the wave heights are especially evident in the summer months (July and August)". I am sorry, but it is not true. The storm period for this region is october-march. I see it in wave reanalysis and in your paper "Prediction of Wave Energy Transformation Capability in Isolated Islands by Using the Monte Carlo Method Deivis Avila et al, 2021" where you used the buoy 2442 which in 50 km from 2446.  I think some error exists in your calculations. Please see the attach. Unfortunately, you need to recalculate everything, I'm sorry.

As explained a complete re-organization of the paper was carried out. Three different buoy locations were used for evaluating the proposed approach. The similarities and differences in wave behaviour in the three cases were analysed.

It is the most important comment that you needs to correct.

7. Please add the figure with real SWH from buoy. for 1-2 years enough

A figure with the measured wave data was included.

 

 

 

Minor commetns:

1. Please add to Keywords: wave buoy data,  Canary Islands

They were added.

2. "The fourth step (Figure 1-iv), consists of fitting polynomial regressions for modelling the five coefficients of the bivariate Weibull distributions as a function of the day of the year (Figure 1-v - check this)

Rewritten in the new version.

3. "more chaotic  and agitated swell" wind waves and swell. I think it is not correct to say about  chaotic  swell. Mixed Sea - mixed wind waves and swell can be look out as chaotic.
4. Seawatch?

Sea point 7.

5. Figure 2. Geographic location of the buoy in question ?? check last word.

please add bathymetry isolines lables.

Proper changes were introduced.

6. Please do not say about the "Gran Canaria Buoy (2442)" which you not used. this only confuses the reader.

It was corrected in the new version.

7. "significant spectral height, H, were recorded each hour, computed for a period of 24 min, from 1998 to 2017. The measurement accuracy was ±0.05 m for H and ±0.05 s for T." Please add the link to technical description of a SeaWatch buoy where we can read that it measured "significant spectral height" with accuracy was ±0.05 m.

References to the three used buoys were included.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have employed a combination of Weibull bivariate probability distribution, polynomial regressions, and the Monte Carlo method to predict/forecast wave energy potential at an offshore point based on historical data.

The paper is constructed well. The methodology is well-explained, and simulation results support the concept. Relevant references are provided.

 

Just a MINOR concern that needs to be addressed.

 

1.       The authors are suggested to talk a bit about the Maximum Likelihood Estimate method (perhaps in the Appendix) and provide formulae/expressions to calculate distribution parameters c1, k1, c2, k2, and c12 to best fit the wave data, although references are given.

Comments on the Quality of English Language

 Minor editing of English language required.

Author Response

Referee 2

Dear reviewer,

Thank you for your useful and good comments and suggestions about our paper. We have modified the manuscript accordingly, and the detailed corrections are listed below.

 

Comments and Suggestions for Authors

The authors have employed a combination of Weibull bivariate probability distribution, polynomial regressions, and the Monte Carlo method to predict/forecast wave energy potential at an offshore point based on historical data.

The paper is constructed well. The methodology is well-explained, and simulation results support the concept. Relevant references are provided.

 

Just a MINOR concern that needs to be addressed.

 

1. The authors are suggested to talk a bit about the Maximum Likelihood Estimate method (perhaps in the Appendix) and provide formulae/expressions to calculate distribution parameters c1, k1, c2, k2, and c12 to best fit the wave data, although references are given.

 

In the new version, the MLE method is not used.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors fixed a lot of comments. Good.

Please check the Buoy Data from Tenefire 2446 Peak Peariod on figure 3. I see that a lot of data for peak period more than 20 sec and it is not seems as a real conditions. You can plot this data as time seria and check it, may be sensor has some errors or problems only last years.  Las Palmas and Gran Canaria data seems like good and real.

May be you can correct the energy estimations after you check Tenerifee initial data.

Author Response

Referee 1

Dear reviewer,

Thank you for your useful comments and suggestions about our paper.

Comments and Suggestions for Authors

Please check the Buoy Data from Tenefire 2446 Peak Peariod on figure 3. I see that a lot of data for peak period more than 20 sec and it is not seems as a real conditions. You can plot this data as time seria and check it, may be sensor has some errors or problems only last years.  Las Palmas and Gran Canaria data seems like good and real.

May be you can correct the energy estimations after you check Tenerifee initial data.

Answer

We agree with the observation that there are peak period values exceeding 20 s, being unusually high values. However, these values are those reported in the used database, which come from the official Spanish harbours authority. All the values declared as unreliable were removed from the study.

Observing the minimum, mean and maximum values for each year (see Figure), it can be seen that in the period 2010-2014, there are maximum peak period values above 20 s. However, there is no indication that this is due to sensor failures. It is also worth noting that, of the 148,575 measured data, only 314 are above 20 s, so that, in any case, their impact on the probability distribution, and therefore on the power assessments, is marginal. The fact that the maximum peak height values of 2016 and 2017 are above average and yet the power and converted energy predictions are well within the predicted reliability intervals justifies this point of view.

For all the above reasons, we believe that no records should be removed from the Tenerife South buoy wave database. In subsequent studies, analyses can be carried out to verify the real incidence of these possible outliers in the models.

 

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

I think Accept

Author Response

Dear reviewer,

Thank you very much for taking the time to review this manuscript and for your suggestions. We have modified the manuscript accordingly and detailed corrections are shown below.

 

Answer

Based on the non-seasonal marking of the peak period of the swell at the Tenerife Sur buoy (2446), an analysis of the behaviour of this variable in each year of the period studied was carried out (see Figure 1 below).

As can be seen, the behaviour is similar in all the years, with no significant seasonal behaviour being observed in any of them.

Although, a priori, it cannot be ruled out that this behaviour is real, we agree that it puts the data under suspicion. Taking into account that clarifying the validity of the data would require a consultation with the Spanish State Ports, which could take some time, it was decided to remove the Tenerife Sur buoy from the study. This was done, moreover, taking into account that as it is only a case study, it does not affect the results of the study, leaving the two remaining buoys for validation.

Consequently, the relevant changes were made to the text and figures in the paper, leaving only the Gran Canaria (2442) and Las Palmas East (1414) buoys as case studies.

 

Figure. Measured values of peak period through the years (Tenerife Sur Buoy).

Author Response File: Author Response.pdf