the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jul 2025
Optimizing the Operation of Energy Islands with Predictive Nonlinear Programming – A case study based on the Princess Elisabeth Energy Island
Abstract. The concepts of Energy Islands or Energy Hubs have gained attention in Europe as a means to enhance offshore wind integration and regional energy systems. These islands can incorporate HVAC and HVDC transmission systems, battery energy storage systems (BESS), and hydrogen production, requiring advanced operational strategies to manage the inherent nonlinearities and time-dependence of their subsystems. To address these challenges, this work proposes a comprehensive framework for the optimal operation of hybrid AC/DC energy islands, addressing: (i) active and reactive power dispatch, incorporating BESS and hydrogen production; (ii) a detailed wind resource characterization based on one year of hourly data obtained using a realistic wind model with local measurements, including wake losses and turbine-level forecasts, used to define representative seasonal and spatial production patterns that inform typical operating conditions; and (iii) operational optimization of a realistic test system based on the Princess Elisabeth energy island, set up using commercial wind power planning tools and advanced forecasting software, and validated with Pyomo/Python.
- Preprint
(1417 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 30 Jul 2025)
-
RC1: 'Comment on wes-2025-102', Anonymous Referee #1, 18 Jul 2025
reply
The article presents a power system modelling approach using non-linear programming applied to a hybrid AC/DC power system, with the Princess Elisabeth Energy Island used as a case study. Based on my review, I believe the manuscript requires major revisions, as outlined below:
Lack of Clear Novelty
The manuscript does not clearly communicate the novelty of the proposed approach. While a non-linear programming method is applied to model a hybrid AC/DC system, it remains unclear what differentiates this work from existing literature. The authors should explicitly highlight the unique contribution and innovation of their methodology.Insufficient Model Validation
The model validation process is not adequately addressed. Using the Princess Elisabeth Energy Island as a case study alone does not constitute validation. The authors should compare their modelling results with real-world data or provide a sensitivity analysis to demonstrate the robustness and reliability of the model.Manuscript Structure
The structure of the article needs refinement. In particular, all modelling outcomes should be clearly presented under a dedicated Results section, separate from other discussions or methodological content.Simplistic Optimisation Objective
The optimisation objective function used in the model is overly simplistic, focusing solely on maximising revenue from offshore wind generation. The authors should justify this choice and consider including additional parameters in the objective function such as operational costs, curtailment, and dispatch down to reflect a more realistic and holistic optimisation scenario.Weak Conclusions
The conclusion section lacks depth. It should be substantially revised to better reflect and interpret the key findings of the study, offering a more comprehensive summary and critical insights into the implications of the results.Citation: https://doi.org/10.5194/wes-2025-102-RC1 -
RC2: 'Comment on wes-2025-102', Anonymous Referee #2, 22 Jul 2025
reply
The authors have created an interesting test system, and they have effectively written a “tutorial” paper on how to create your own test system, along with some simulation outputs to show that their model works. However, the research value of the paper is unclear. A non-linear optimization model is proposed and implemented, but it is not clear whether such an approach is well justified. For example, ac power flow is implemented, and voltage limits are monitored, but it is unclear if these are binding constraints, and hence whether a simpler dc power flow approach could have been implemented to produce “similar” costs, but with a reduced computational burden. (The presented results focus on “dispatch” variables rather than “network” variables.) It also seems that perfect forecasting of electricity prices and wind power is assumed. How do forecast errors impact the methodology and the results and conclusions? Representative days are convenient for showing that a model is working, but otherwise they have limited value, particularly when the BESS start and end state of charge is fixed, despite day to day variations in wind speed and electricity price. The electrolyzer produces and accumulates hydrogen across a representative day, but what happens to the hydrogen, and are there “downstream” constraints associated with the hydrogen production? The term “predictive” is used, but what does it mean? The authors seem to assume perfect knowledge of wind power and electricity price, so what does “predicted” refer to?
Remove the full stop at the end of the paper title
Line 20 – a long list of references is given, but no details are provided on the individual references
Line 220 – forcing the BESS start and end state to be the same is not optimal, noting, for example, wind variability from one day to the next
The paper only considers energy revenues. What about potential revenues from providing system services?
The figures (2 of) spell electrolyzer incorrectly
GB and UK terminology are both used – GB is correct
Table 1 – are the values presented here publicly known, or have they been chosen by the authors? If the latter, how have the values been chosen, and do they lead to revenue maximization?
Line 299 – the word “degradation” normally relates to a reduction in performance over the lifetime of a component, but here it looks as if the term is being used in relation to a change in electrolyzer output. The authors’ definition is unexpected.
Line 299 – in relation to Figure 6b, it would be helpful for the reader to understand the most likely electrolyzer output, and how that might vary across the different seasons, in order to better appreciate whether a “larger” error at low outputs is of minor or major significance
The test results focus on seasonal “normal” days, but “less normal” days are also important, and may well influence equipment sizing.
Page 16 – lots of details are given on modelling individual components, and basic details of the test data, but very little information is given on the nonlinear optimization methodology.
How does pyflow-acdc work, and what are the key differences with the authors’ approach? The paper doesn’t provide sufficient information to judge the comparison, and pyflow-acdc results are not shown in the paper.
Line 350 – it seems that the optimization process assumes perfect (day ahead) knowledge of wind power daily profiles and the day ahead electricity prices, but the validity of this assumption is not justified.
A variety of x-axis scales are used to show a 24 hour day – use one style for all figures, either the style from Fig. 10 or from Fig. 11.
Reference is made to hydrogen targets. What are these, and are they an optimized variable?
Citation: https://doi.org/10.5194/wes-2025-102-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 97 | 21 | 24 | 142 | 7 | 6 |
- HTML: 97
- PDF: 21
- XML: 24
- Total: 142
- BibTeX: 7
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1