Mehta et al., 2023 investigated the drivers for sizing offshore wind turbines when minimising the LCoE. In this follow-up paper, the authors explore how different day-ahead market price scenarios affect the turbine sizing and compare it to the LCoE-optimised design. 

Both papers use the same MDAO modelling framework. It consists of a chain of low-fidelity models aiming to capture the combined effect of two wind turbine (WT) design variables: rated power and rotor diameter. Finding the optimal design requires the definition of an optimisation objective and wind farm constraints. The new paper defines objective metrics that include the day-ahead market price but applies the same wind farm constraints to all investigations: rated power (P_farm=1 GW) and area (A_farm=150 km²). While the paper provides valuable insights into market value objectives, keeping the wind farm constraints fixed raises questions about the study's validity. Denmark's new 6 GW offshore tender has a per GW area of about 450 km², or about three times that of the current work. 

The authors conclude that the benefits of making market-optimised WTs are minimal— the LCoE-optimised WT performs similarly to the market-optimised. However, I am unsure if that conclusion is not due to the wind farm (WF) constraints. The purpose of redesigning the WT, e.g., by increasing its rotor area, is to increase production at lower wind speeds where electricity prices are higher. Similarly, decreasing the WT's rated power reduces costs while only decreasing low-value output at high wind speeds. Conceptually, the authors optimise the WT's power curve to new market conditions. However, with the fixed constraints on WF area and rated power, the resulting WF power curve is locked irrespectively of the chosen WT design. The placement of the wind turbines further enforces this fixation. The turbines are aligned towards the main wind direction, i.e. they are placed to maximise production, not to optimise market value. Like the WF area, the regular rectangular layout does not represent current commercial wind farms.

The paper provides a good overview of objective metrics, including the day-ahead market value of wind energy. It also shows how changing the WT size influences these metrics for a single wind farm case. However, with the new market incentives, the authors should adapt both the design objectives and the constraints. The paper should provide other examples with WF constraints and layouts representing modern wind farms. The authors can do this by removing the wind farm power constraints, changing the wind farm area and optimising the WF layout for market value instead of AEP. As it is, only a single case study is presented— with very restrictive constraints making it hard to draw the general conclusions that they are.

In the paper “Designing wind turbines for profitability in the day-ahead markets”, the authors present an interesting and timely study of how variable electricity prices and especially negative correlation between prices and wind speed could impact the optimal design of wind turbines. However, I have some questions about the assumptions, and some comments related to the interpretation of the results and the conclusions. I also suggest that some additional case studies/sensitivity studies could help to generalize the findings.

• It is mentioned that the coefficient of variation (CV) is kept constant. What is this constant value? (maybe it’s mentioned somewhere, sorry if I missed it. But it would be nice to write it in Table 1 so it’s easy to find)
• The authors say: “It is known that the value for CV also differs, but it is expected to have the smallest range of variability of all the three market parameters”. I would slightly disagree. E.g., in the reference Swisher, et al. (2022), CV is expected to roughly double from 2025 to 2045, and similar growth is seen also in other studies (reflecting the increasing share of wind and solar generation with near-zero operating cost, increasing CO2 tax and decommissioning of base load thermal generation, which all tend to drive price variability up). Also, while the mean price can vary a lot between years (e.g., due to weather and gas prices), the increase in CV is usually seen as a more permanent change as we go towards a highly weather dependent system with only a few peaker plants. Thus, I find a bit surprising that CV is not varied (especially to higher values than seen historically). It would be great to hear more justification on this selection, or potentially to see a sensitivity study where also CV is varied (especially a high CV and strong negative correlation between wind speed and prices would be an interesting case to study).
• The study is performed for a location with high mean wind speed. However, many of the very low specific power turbines in the literature are suggested for lower mean wind speed sites (usually onshore). Do you think that the results could be different for a low/medium wind resource onshore site? (even nicer would be to add such sensitivity study to the paper)
• The paper presents the optimal designs over a design space of rotor diameter and rated power for a “standard turbine”. E.g., the LowWind concept presented Swisher, et al. (2022) is somewhat more radical, with the objective of reducing CAPEX by completely shutting down the turbine at a very low cut-off wind speed. Would you see such more dramatic design choices becoming potentially interesting in the future, or would you still say that the “standard turbine” with LCOE-driven optimization is the way to go?
• A comment which you may consider related to the discussion about “designing for the wrong market”. Overall, it’s a very important point, and the robustness of the LCOE optimized design is an interesting result. However, while I agree that the future mean electricity price level is pretty much guesswork, I consider the negative correlation between wind speed and prices to be a more consistent expected future, e.g., in the central and northern Europe (as wind generation is expected to dominate and thus negative correlation seems almost inevitable). Thus, the risk of “designing for the wrong market” in the sense of getting the correlation wrong (that it would be around zero rather than, e.g., around -0.5) would seem to me to be a somewhat lower risk than getting the mean price level wrong.