BESS Reserve Optimisation in Energy Communities

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper investigates optimising Battery Energy Storage Systems (BESS) to enhance the business models of Local Energy Markets (LEMs). LEMs are decentralised energy ecosystems facilitating peer-to-peer energy trading among consumers, producers, and prosumers. By incentivising local energy exchange and balancing supply and demand, LEMs contribute to grid resilience and sustainability. This study proposes a novel approach to BESS optimisation, utilising advanced artificial intelligence techniques, such as Multilayer Perceptron Neural Networks and Extreme Gradient Boosting Regressors. These models accurately forecast energy consumption and optimize BESS reser've allocation within the LEM framework. The findings demonstrate the potential of these AI-driven strategies to improve BESS performance and overall LEM efficiency. Here are some detailed issues.

1. The principles and application scenarios of the advanced artificial intelligence technologies adopted, such as multi-layer perceptron neural networks and extreme gradient enhanced regressors, need to be elaborated in more detail in this paper.

2. The data acquisition method and processing process should be explained to verify the reliability of the model prediction results.

3. The specific effectiveness of AI-driven strategies in improving battery storage system performance and local energy market efficiency should be elaborated.

4. It is recommended to fully verify and compare the proposed optimization methods for battery energy storage systems. Other traditional methods or different AI techniques can be considered as control groups to demonstrate the effectiveness and superiority of the proposed methods.

5. In the conclusion part, it is suggested to further explore the significance and limitations of the research results, and look forward to possible future research directions and expanded application fields. This helps readers to better understand the far-reaching impact and development potential of the research.

Comments on the Quality of English Language

 Moderate editing of English language required.

Author Response

The author's reply to the Review Report is attached as the MS Word file BESS Reserves Optimisation in Energy Communities- Reviewer Report #1.docx

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors,

First of all, thank you for your work. While the paper seems well written and organized on the verbal bits, it did not look as meticulous on the display of its results. 

Starting from figure 1, the battery bit takes up more space than the graph that describes the excursions which are more important. If the graph was given more space, it would be easier to read.

The section highlighting of figure 4 under the figure 3 does not look consistent or necessary.

Every figure since the 5th is only readable at 200% zoomed in. I am aware that the selected time window is huge with lots of samples but if there is a way to make these figures more readable by changing the sizes/resolutions it would help the paper become easier to read.

The layout of tables mixing with bar graphs is misaligned and breaks the flow.

In figures 10 and 11, despite the titles saying comparisons, why does there seem to be only one piece of variable (KPI2)?

In addition, these excursions look nothing like the ones described in figure 1. In figure 1, the excursions look more like the region in between. Can you explain this? 

At the start of the discussion, it is written "as presented" can you indicated the presented?

Going back to the introduction what are the DNO/DSO? It would be good to expand and introduce these elements. The excursions, even if they are as minimal as possible thanks to predictions are inevitable by nature so what makes an excursion too much for the DNO/DSO? What are these boundaries? It would help to explain.

The paper should elaborate a little bit on the XGR versus MLP-NN. While the results speak for itself, it is still helpful to comment a bit. For example, the performance difference looks small but would it mean a huge difference for excursion penalty by the terms of contract? What are the possible trade offs between these approaches? 

It is a technical paper but since the purpose is to help the BESS operators and investors, it would be a great contribution to mention about the economic aspect a bit more. How much money could machine learning forecasting save for the BESS? Are there any figures that can be spoken of?

Author Response

The author's reply to the Review Report is attached as the MS Word file BESS Reserves Optimisation in Energy Communities- Reviewer Report #2.docx

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The methods proposed in this paper have certain value and significance. However there are still several areas that need to be further improved.

1．Please check the format of the article ，such as in” Model assessment will use the RMSE, MSE, MAE, and R2.”, is it R2 or R² ?

2. Please check the serial numbers “1. and 2.” used in lines 144 and 145 and other places. They are the same as that part “1. Introduction” “2. Materials and Methods”, so it is hard to distinguish them.

3. Please provide clearer pictures in the article to facilitate readers' reading, such as Figure 1.

4. Figure 8 and Figure 9, table 6 and table 7 seem to be the same. Please describe their differences.

 

5. Do those indicators have units in table 2-7 and figure 8-9? 6. Line 195: please explain the differences between eXtreme Gradient Tree and XGB. 7. Please explain the meaning of the icon ‘B03 Consumption kW_Mean, $N-.B03 Consumption kW Mean, $XGT-.803 Consumption kW Mean’ in figure 5-9, especially ‘XGT’ . Do they correspond to the algorithms (MLP-NN and XGB)? 8. Line: 199, please further describe or explain the calculation of p-values. Why are the results of the two algorithms (MLP-NN model: 0.315 (stable), and XGB model: 0.814 (stable)) significantly different? How do they relate to the data in Table 2, Table 3, Table 6 and Table 7? The article proposes two algorithms, and most of the XGB’s residuals are smaller than MLP-NN’s residuals. Which algorithm is better? Why? Which algorithm is ultimately recommended in this article? Comments on the Quality of English Language

Minor editing of English language required.

Author Response

The author's reply to the Review Report is attached as the MS Word file BESS Reserves Optimisation in Energy Communities- Reviewer Report #3.docx

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for addressing all of my comments in your revised paper. I have but one more correction, that is, the last formula you provided for tariffs. It looks wrong and must be corrected which is the reason i recommended minor revisions. However, no further comments from me is needed and the paper is 99% good to go. 

Author Response

Answet is included in attached file BESS Reserves Optimisation in Energy Communities- Reviewer Report #2 020924.docx

Author Response File: Author Response.pdf