Design Optimization and Sizing for Fly-Gen Airborne Wind Energy Systems
Round 1
Reviewer 1 Report
The article presents an optimised design approach for Fly-Gen AWE systems. I found it very interesting paper and valuable to the community of Airborne Wind Energy researchers. I think the manuscript is publishable after minor revision as follows.
1- Please add a reference(s) for what is claimed in lines 15-17, "Airborne wind energy (AWE) is a technology with the potential to harvest wind energy less expensively over deeper water, and therefore to be more widely deployable than current wind energy technologies."
2- Some parameters are not defined in the context. Please review all the equations and define every parameter at the first time it is used.
3- Lines 53 to 64, the authors compare Fy-Gen and Ground-Gen technologies. One can understand that the authors try to promote their favourite technology i.e. Fly-Gen. However, the comparison looks unfair without enough supporting data; Please consider to modify this part. For instance, it is mentioned that "Ground-Gen systems frequently use flexible aircraft,...". What do you mean "Frequently"?. Many Ground-Gen developers like AmpixPowr are using rigid wing. Flexible aircraft cannot be attributed to the Groud-Gen systems as an inherent feature since Ground-Gen system are able to work with both rigid and flexible aircraft.
4- Reference number 8, Please add more details to the bibliography(web address, etc.)
Regards
Author Response
Thank you for your work reviewing this paper. The authors have produced a revised paper which addresses your comments.
1. The revised paper rephrases lines 15-17 to be clearer about what is being claimed "Airborne wind energy (AWE) is a technology with the potential to harvest abundant wind resources located over deep water less expensively than current wind energy technologies" and adds a citation.
2. The revised paper more carefully defines variables where they are introduced.
3. This revision attempts to be fairer to ground-gen technology (lines 50-66), separating the discussions about ground-gen cycles and flexible vehicles, and mentioning the higher mass and tether drag of fly-gen systems.
4. This revision also adds URLs or DOIs to several of the bibliography items, including number 8 (now 9).
Reviewer 2 Report
The topic of the paper, which really centers around multidisciplinary design optimization for airborne wind energy (AWE) systems, is of great interest to both the AWE community and the optimization community. As the authors point out, and as is made clear as one reads through the paper, the level of coupling between various design parameters in AWE systems is extremely significant and worthy of deep investigation. The authors also provide an excellent background regarding the basic power equation formulated by Miles Loyd and then adapted by others to account for things like cosine losses due to non-zero elevation and azimuth angles. The various modeling sections, demonstrating the impact of aspect ratio on performance, structural scaling, etc., are independently useful. Having said all that, the core contribution of the paper, namely the optimization itself, is quite unclear and needs to undergo a significant revision for this work to be useful from an optimization standpoint. Hopefully, my comments below will help the authors to reshape the paper in that regard:
1) In order to formulate an optimization problem, one must define decision variables, an objective function, and some set of equations that relate the decision variables to the objective function value. The decision variables for this optimization are not laid out clearly. In Table 4, I see 13 rows...do these all represent decision variables? Or is one of these variables supposed to be an independent variable, and the other 12 rows represent the optimal values of the decision variables (for different values of the independent variable, given by the columns)? Is the objective function simply LCOE?
2) What algorithm is actually used to perform the optimization? How is that algorithm justified, given the complexity of the mathematical properties of the optimization problem (e.g., convexity)?
3) The authors appear to claim their system has a power curve that is linear with respect to the wind speed (bottom of page 5 and in Figure 4), up to the rated wind speed. Am I missing something here? Why is the relationship not cubic, as would be predicted by equations (4) and (8) (along with the relationship between wind speed and power for any other wind turbine)?
4) Given that reference [25] is still in review, some detail regarding the trajectory optimization algorithm would be valid. Of particular concern is the statement that "No controller design or tuning is required." First off, the trajectory optimization itself represents a controller design. Secondly, what are the outputs of that trajectory optimization? Looks like path, velocity, and lift coefficient. If so, then there needs to be some closed-loop controller that tracks the path, which should be acknowledged in the paper.
Author Response
Thank you for your work reviewing this paper. The authors have produced a revised paper which addresses your comments.
1. & 2. The revised paper clarifies the optimization problem including decision variables, objective function, and optimization algorithm in section 2.2, reiterates the decision variables in section 3, and discusses potential future work trying global optimization algorithms in section 4.
3. The revised paper clarifies the use of and reasoning for the linear fit on power curves in section 2.3: "Power vs. wind speed is cubic for unconstrained wind energy systems, however tension and power constraints cause the power curve to taper off and reach a maximum at the rated wind speed\cite{VanderLind2013}. In practice, fitting power curves with a linear approximation up to the rated speed and a constant above the rated speed produces a good estimate with reduced computation." and "For the final iteration of each optimization, full power curves are evaluated at each integer wind speed between $5m/s$ and $25m/s$ in order to validate the use of the trapezoidal power curve estimates for the optimization."
4. The revised paper attempts to clarify section 2.1. The point that line was attempting to underscore is that this analysis method is not using a simulator or anything else that needs a controller. Yes, the trajectory optimization can be used as a controller design, and yes, used for control, it would require an additional feedback controller, and that is discussed in "A Non-linear Inverse Model for Airborne Wind Energy System Analysis, Control, and Design Optimization", but that is not how the method is applied in this paper. The language reads: "A non-linear model inversion based on simplifying assumptions appropriate for fly-gen systems has been validated against a high-fidelity simulation and is used for this design optimization. Rather than running a simulation and tuning a closed-loop controller to track a desired trajectory, the inverse model calculates power produced from operation on a given trajectory in a given wind environment directly."
Round 2
Reviewer 2 Report
The authors addressed my comments from the earlier review. They did a good job clarifying the decision variables, the optimization technique, and exactly what the optimization accomplishes (e.g., noting that a closed-loop controller will in fact be necessary at some point).
The justification for the linear power curve up to rated wind speed still seems somewhat weak.
Overall, as noted in my previous review, the topic of this paper is very well-suited to the airborne wind energy systems community, and I believe it should be published.
