Environmental and Social Life Cycle Assessment of Data Centre Heat Recovery Technologies Combined with Fuel Cells for Energy Generation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper compares a standard scenario, where the DC uses network electricity, and where the industrial gasses are converted into heat in a cogeneration for a district heating system, with a new scenario, where the industrial gasses are converted into biogas and sent to a SOFC that feeds the DC with electricity. If I read well table 6, two situations are actually compared: a baseline DH with limited power production (how is it treated in the calculus?), and a new DH with significant power production treated as avoided burdens.

 

The paper is interesting, as it covers a current preoccupation and has some originality thanks to the use of the less developed s-LCA method. It can be accepted with some additional clarifications and discussions:

 

1.- The system boundaries in the LCA are not sufficiently clear in the text. In figure 1, it is not clear enough (if I understand well) that the DC itself is not modelled, and that the data is not considered as a system output (which explains the absence of related functional unit and of allocation). I suggest to better indicate the system boundaries on figure 1 (dotted line), and to complete it with a figure of the reference scenario, indicating the boundaries as well. The external electricity input of the DC should appear.

2.- It is mentioned that the reference scenario is based on a cogeneration working with a industrial gas mix. I understand that electricity is co-produced. It should appear on the figure requested in point 1. You should also specify in the text how this power production is treated in the baseline scenario. Is it sent to the DC? To the electricity network? When you consider the new scenario, how do you cope with its suppression?

3.- Table 3: “module”: which module do you refer to? Also the SOFC? There are at least two. Maybe the first column title is inadequate (?).

4.- The LCI content is insufficiently described. A more explicit table with the inputs and outputs (and related assumptions, for instance local energy mix – Swedish?) is missing.

5.- More a question than a criticism: is there any solution to reduce the water use impact of the developed solution? 2 m² per kWh is very high: maybe this water can be recovered in a way? If yes, this can appear in the discussion.

6.- L288-289: please check the consistency with the figures/tables: 0.21 kg and -0.15 kg.

7.- The application of the s-LCA  “recipes” to your system is a real originality of your work. However, s-LCA is less developed than classical LCA and the application of s-LCA deserves some critical discussion that is absent here (might appear for instance in the limitation section, or as independent paragraph). To some extent, the selected indicators are adequate to study a macro scale and don’t seem relevant to study the micro changes operated by the module. In addition, I understand that the systems were studied at different times: some differences might be explained by conjunctural factors rather than by the intrinsic features of the systems. Last, studying a technological system social impact based on social indicators at a company scale can lead to uncertainties due to a small number of employees (how many here?). To be more precise about these scale relevance, uncertainty and representativeness, there are some necessary step backs/questions to be raised:

- How much time is there between the “before” and “after” data collections? How does it impact indicators like the number of affiliates to syndicates, number of women, level of salaries, number of feminine managers etc…? Do external events may bias them?

- Lowest paid worker: is this indicator sufficiently robust, as it may depend on the situation of one single person in the company? Maybe other inequalities indicators would better suit? Influence of the inflation during the “experiment”? Are you speaking after correction of the inflation?

- When you speak about proportions of people in the company, how many people do you refer to? How do you explain the considerable improvement of the women employment? (some discussion needed). And the drastic drop of employed disabled people? Real influence of the studied technology, or “random” evolution of the company?

- Salary in the sector: which one? Energy sector? What does it show about the influence of the studied technology, if this sector is influenced by many other factors?

- Table: no arrow should appear after “sick leave” and “accident ratio”, since data is lacking.

- Other factors should be described in the text: environmental certification (withdrawn during the experiment?); relevance for the basic needs (why “yes”, then “no”?); etc.

- The “collection pool” mentioned in the method section should be explained. Questionnaire, sample size? With which stakeholders? How did you assess subjective issues like the access to the information, the relevance of the product or the legal actions? Via qualitative interviews (often appropriate in social science)?

Author Response

The paper compares a standard scenario, where the DC uses network electricity, and where the industrial gasses are converted into heat in a cogeneration for a district heating system, with a new scenario, where the industrial gasses are converted into biogas and sent to a SOFC that feeds the DC with electricity. If I read well table 6, two situations are actually compared: a baseline DH with limited power production (how is it treated in the calculus?), and a new DH with significant power production treated as avoided burdens.

The paper is interesting, as it covers a current preoccupation and has some originality thanks to the use of the less developed s-LCA method. It can be accepted with some additional clarifications and discussions:

Thank you very much for your interesting comments and recommendations, which have improved the manuscript. We also appreciate the time devoted to our paper.

Comment: 1.- The system boundaries in the LCA are not sufficiently clear in the text. In figure 1, it is not clear enough (if I understand well) that the DC itself is not modelled, and that the data is not considered as a system output (which explains the absence of related functional unit and of allocation). I suggest to better indicate the system boundaries on figure 1 (dotted line), and to complete it with a figure of the reference scenario, indicating the boundaries as well. The external electricity input of the DC should appear.

Response: Thank you very much for this observation. We have modified the figure 1 to a better description of the scenarios, including a baseline scenario description, the system boundaries and energy inputs.

Comment: 2.- It is mentioned that the reference scenario is based on a cogeneration working with a industrial gas mix. I understand that electricity is co-produced. It should appear on the figure requested in point 1. You should also specify in the text how this power production is treated in the baseline scenario. Is it sent to the DC? To the electricity network? When you consider the new scenario, how do you cope with its suppression?

Response: Thank you very much for your comment. The electricity produced by the cogeneration plant is sent to the electricity grid. The DH only receives thermal energy from the cogeneration plant, so the electricity has not been considered in the LCA. However, taking into account the reviewer's comment, we have slightly modified the manuscript as follows:

“The thermal energy demand of the science park is currently covered by a DH network, which provides energy from a cogeneration plant using recovered gasses from a steel mill as fuel. The composition of these recovered gasses is: 70% blast furnace gas, 20% converter gas (also known as LD gas), and 10% coke oven gas. It is important to mention that the electricity produced by the cogeneration plant is sent to the electricity grid. The DH only receives thermal energy, so this electricity has not been considered in the LCA. This configuration represents the baseline scenario (pre-WEDISTRICT situation), shown in Figure 1.”

Comment: 3.- Table 3: “module”: which module do you refer to? Also the SOFC? There are at least two. Maybe the first column title is inadequate (?).

Response: Thank you very much. All the information provided by Table 3 (Table 2 in the new version of the manuscript) is related to the SOFC module. However, considering this comment and the following one, we have modified Table 3 (Table 2 now) in the new version of the manuscript to clarify the information shown as well as to include additional information regarding the LCI (in supplementary material).

Table 2. WEDISTRICT scenario energy production

Energy production	Annual production	Unit
Thermal energy	15600	kWht / year
Electricity	78785	kWhe / year

Comment: 4.- The LCI content is insufficiently described. A more explicit table with the inputs and outputs (and related assumptions, for instance local energy mix – Swedish?) is missing.

Response: Thank you for this great observation. We have included a detailed inventory of the WEDISTRICT scenario as supplementary material. Indeed, for the electricity coming from the grid, the Swedish mix has been considered, as shown in the new figure 1.

Comment: 5.- More a question than a criticism: is there any solution to reduce the water use impact of the developed solution? 2 m² per kWh is very high: maybe this water can be recovered in a way? If yes, this can appear in the discussion.

Response: Very good observation, thank you very much. In this case, the impact on the water resource in the WeDISTRICT scenario is determined by the life cycle of the biogas used. Specifically, it is due to the biogas methanization process, which consumes a large amount of water. For this study, data from the EcoInvent database were used, simulated with SimaPRO software, to account for the life cycle of the biogas obtained and all its unit processes. This accounts for 98.9% of the impact on this resource. Therefore, to try to reduce this impact, action would need to be taken on the source of the hydrogen used in the fuel cells. The source and the method of obtaining that hydrogen will completely determine this environmental impact.

We have introduced a paragraph in the results chapter to a better clarification.

Comment: 6.- L288-289: please check the consistency with the figures/tables: 0.21 kg and -0.15 kg.

Response: Thank you very much for your comment. As suggested, we have revised the consistency of the information provided by the Figures and Tables and, apparently, we have not found any mistake. Please, consider that Tables 5 and 6 (Tables 4 and 5 in the new version of the manuscript) and Figures 2 and 3 are representing different things, respectively. The Tables show the values of the impacts of the functional unit estimated for the baseline and the two WEDISTRICT scenarios. On the other hand, Figures are comparing the results of both WEDISTRICT scenarios with those obtained for the baseline just for showing if the impacts produced by the WEDISTRICT scenarios are higher or lower.

As an example, we mentioned the following in the manuscript: “In terms of climate change, including the avoided burden, the LCA results show that 0.21 kg CO₂ eq are avoided for each kWh auto consumed, leading to a global carbon footprint of -0.15 kg CO₂ eq/kWh for the WEDISTRICT scenario.”

In Table 6, it is shown that the Environmental impact for Climate Change (for instance) of the baseline, WEDISTRICT and WEDISTRICT + avoided burden scenarios are, respectively: 0.12, 0.064 and - 0,15 kg CO₂ eq/KWh. Taking these values into account, it can be concluded that the effect of considering the avoided burden is equal to 0.064 – (- 0,15) = 0,21 kg CO₂ eq/KWh.

On the other hand, as previously said, if the results of both WEDISTRICT scenarios are compared with those of the baseline scenario, the information provided by Figure 6 can be obtained. If it is considered that the baseline scenario impacts are equal to 100, then:

(WEDISTRICT/Baseline)*100 = (0.064/0.12)*100 = 52

(WEDISTRICT + avoided burden/Baseline)*100 = (-0.15/0.12)*100 = -118

If the reviewer still has questions about this or if we have not understood the comment correctly, please do not hesitate to let us know.

Comment: 7.- The application of the s-LCA  “recipes” to your system is a real originality of your work. However, s-LCA is less developed than classical LCA and the application of s-LCA deserves some critical discussion that is absent here (might appear for instance in the limitation section, or as independent paragraph). To some extent, the selected indicators are adequate to study a macro scale and don’t seem relevant to study the micro changes operated by the module. In addition, I understand that the systems were studied at different times: some differences might be explained by conjunctural factors rather than by the intrinsic features of the systems. Last, studying a technological system social impact based on social indicators at a company scale can lead to uncertainties due to a small number of employees (how many here?).

Response: Thank you very much for raising this issue. A paragraph regarding the limitations have been added in the corresponding section in the new version of the manuscript.

Comment: To be more precise about these scale relevance, uncertainty and representativeness, there are some necessary step backs/questions to be raised:

- How much time is there between the “before” and “after” data collections? How does it impact indicators like the number of affiliates to syndicates, number of women, level of salaries, number of feminine managers etc…? Do external events may bias them?

Response: Thank you very much for your question. The ‘before’ and ‘after’ are the same as the LCA, which are taken before the implementation of the technology (baseline) and after 24 months. No external events were reported.

Comment: - Lowest paid worker: is this indicator sufficiently robust, as it may depend on the situation of one single person in the company? Maybe other inequalities indicators would better suit? Influence of the inflation during the “experiment”? Are you speaking after correction of the inflation?

Response: Thank you very much for your question. As widely recommended in S-LCA reporting, lowest paid worker is valuable for identifying wage inequality and labor rights issues, although it may not be sufficiently robust on its own due to several limitations. Thus, we considered additional indicators which are not presented in the table 6 as they present no change during the implementation of the technology. These indicators complement the mentioned indicator. We have not considered anything related to inflation as during the analyzed period no significant inflation in Sweden was detected, thus was considered not relevant.

Comment: - When you speak about proportions of people in the company, how many people do you refer to? How do you explain the considerable improvement of the women employment? (some discussion needed). And the drastic drop of employed disabled people? Real influence of the studied technology, or “random” evolution of the company?

Response: Thank you very much for raising this question. We understand that assessing social impacts at a company scale with a small number of employees can introduce uncertainties due to limited data, but this does not undermine the importance of such evaluations. To mitigate uncertainties, the company-level findings were benchmarked against nation-wide social indicators, offering context and reducing variability, as mentioned in the document, For further clarification on this, a clarification sentence has been added in the new version of the manuscript.

Comment: - Salary in the sector: which one? Energy sector? What does it show about the influence of the studied technology, if this sector is influenced by many other factors?

Response: Thank you very much for your question. RISE and the Lulea Science Park belong to the research and education sectors; thus, the salaries were compared to other institutions in the sector. No additional or external factors were considered as none were not reported.

Comment: - Table: no arrow should appear after “sick leave” and “accident ratio”, since data is lacking.

Response: Thank you very much. We totally agree with your comment and these two arrows have been removed in the new version of the manuscript.

Comment: - Other factors should be described in the text: environmental certification (withdrawn during the experiment?); relevance for the basic needs (why “yes”, then “no”?); etc.

Response: Thank you very much for detecting this issue. It has been corrected in the updated version of the manuscript.

Comment: - The “collection tool” mentioned in the method section should be explained. Questionnaire, sample size? With which stakeholders? How did you assess subjective issues like the access to the information, the relevance of the product or the legal actions? Via qualitative interviews (often appropriate in social science)?

Response: Thank you very much for raising this question. It has been included in the new version of the manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

a) The study includes data in Table 1. Some rationale or source of the data should be provided.

b) The term WEDISTRICT has been used extensively used throughout the manuscript. Adding a little description of the author’s interpretation of this term would add more clarity.

c) The manuscript needs improvement in the flow of the discussion. For instance, the reader expects to see Table 1 (not Figure 1) when a reference to the table is cited in the text.

Author Response

Comment: a) The study includes data in Table 1. Some rationale or source of the data should be provided.

Response: Thank you for this comment. We have eliminated this table and we have introduce the data source in the text.

Comment: b) The term WEDISTRICT has been used extensively used throughout the manuscript. Adding a little description of the author’s interpretation of this term would add more clarity.

Response: Thank you very much for this observation.

WEDISTRICT is the acronym of the European project “Smart and local reneWable Energy DISTRICT heating and cooling solutions for sustainable living”. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement N°857801.

In this European project, renewable energy generation technologies are implemented in four locations across Europe, forming four demonstration projects. One of these locations, and therefore one of these demonstration projects, is Luleå (Sweden). For this demonstrator, the situation before the implementation carried out in the WEDISTRICT project (referred to as the baseline scenario in this manuscript is compared with the situation after the implementation of the proposed technologies in the WEDISTRICT project (referred to as the WEDISTRICT scenario in this manuscript).

We have introduced some paragraphs in introduction and methodology sections to clarify it.

Comment: c) The manuscript needs improvement in the flow of the discussion. For instance, the reader expects to see Table 1 (not Figure 1) when a reference to the table is cited in the text.

Response: Thank you very much for your comment. We have slightly changed the manuscript to follow your advice. Among other things, we have removed Table 1 (showing the composition of waste gases used for heat generation in the baseline scenario) because this composition is already informed in the text. In addition, Figure 1 has been changed at the request of another reviewer to include the baseline and the WEDISTRICT scenario flowcharts and system boundaries. According to the authors´ opinion, the changes made in this new version of the manuscript after reviewers' comments have been considered help in understanding the paper.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This paper is about the environment and social life cycle assessment of data center heat recovery technology. However, in the present form, there is a requirement for many improvements:

1. The abstract is too long, the authors should delete some background in the abstract.

 

2. As the manuscript is about fuel cells, some background about fuel cells should be added with the introduction, such as ieeexplore.ieee.org/abstract/document/10526347.

 

3. The Figures 2 and  3 are unclear. Hence, the resolution should be further improved and higher-quality figures should be provided.

 

4. Page 2, 'This research will focus on using LCA and S- LCA as a tool to analyse the environmental and social impacts related to an innovative set-up demonstrator combining waste heat recovery from DC technologies with biogas- fed solid-oxide fuel cells (SOFC) for energy generation.'

    Why is SOFC analyzed？ How about PEM fuel cells? Compared to SOFC, the operating temperature of PEMFCs is closer to the ambient temperature. Hence,  PEMFCs have broad application prospects in fields such as transportation, distributed power generation, and portable power sources (10.3390/en17123050). Some comparisons and discussions should be added between PEMFC and SOFC. 

    Moreover, the hydrogen should be green and generated from renewable energies. "However, most renewable energy sources are intermittent and geographically constrained, resulting in temporal and spatial gaps between the energy availability and the consumption of the end-users." This can be found in Energies 2024, 17(12), 2917. Fuel cells have been extensively employed, and are one of the most promising alternatives for the future power sources for electric propulsion applications and distributed generation systems. 

 

5. Some economic results need to be provided in this manuscript.

 

6. The innovation points of this manuscript need to be further clarified.

 

Author Response

This paper is about the environment and social life cycle assessment of data center heat recovery technology. However, in the present form, there is a requirement for many improvements:

Thank you very much for your interesting comments and recommendations, which have improved the manuscript. We also appreciate the time devoted to our paper.

Comment: The abstract is too long, the authors should delete some background in the abstract.

Response: Thank you for the comment. Following your recommendation, we have reduced the abstract extension.

Comment: As the manuscript is about fuel cells, some background about fuel cells should be added with the introduction, such as ieeexplore.ieee.org/abstract/document/10526347.

Response: Thank you for the comment, but we prefer not to include the background that you mention due the aim of this research and the extension limitations of the paper.

The aim of this research is not to make a comparison between different fuel cell technologies; nor even to make a comparative environmental/social assessment between different types of fuel cells. The aim of the research presented here is to carry out an environmental and social comparison between an existing pre-existing situation and the implementation of a renewable solution that uses waste heat from a data centre and the availability of renewable biogas close to the site.

The specific technologies for using these resources have not been defined in this research and are given by previous decisions taken in the WEDISTRICT project itself.

Obviously, research that carried out an environmental and social comparison between different fuel cell technologies would first require an exhaustive comparison of these types of technologies, highlighting their operating conditions, their main characteristics, as well as their advantages and disadvantages.

Comment: The Figures 2 and 3 are unclear. Hence, the resolution should be further improved and higher-quality figures should be provided.

Response: Thank you for the comment. Following it, we have modified and improved the figures.

Comment: Page 2, 'This research will focus on using LCA and S- LCA as a tool to analyse the environmental and social impacts related to an innovative set-up demonstrator combining waste heat recovery from DC technologies with biogas- fed solid-oxide fuel cells (SOFC) for energy generation.'

    Why is SOFC analyzed？ How about PEM fuel cells? Compared to SOFC, the operating temperature of PEMFCs is closer to the ambient temperature. Hence,  PEMFCs have broad application prospects in fields such as transportation, distributed power generation, and portable power sources (10.3390/en17123050). Some comparisons and discussions should be added between PEMFC and SOFC.

Response: Thank you very much for your appreciation. SOFC were selected as the main raw material is biogas, available in a close site from the treatment of forest wastes. PEMFC is a mature technology for the direct conversion of hydrogen into electricity, and there was not hydrogen source available for the WEDISTRICT project, that is mainly devoted to providing full renewable solutions for district heating and cooling, integrating thermal energy systems. An alternative for the use of PEMFC is the installation of PV panels and electrolysers, but Lulea, at such Nordic latitude, has very reduced solar availability during a significant number of months. On the other hand, a direct use of a local resource (as biogas) was available along the whole year, without any other storage but the biogas tanks. In addition, SOFC has a huge potential for thermal integration, even at higher temperatures, allowing the use of absorption chillers or any other industrial heat.

    Comment: Moreover, the hydrogen should be green and generated from renewable energies. "However, most renewable energy sources are intermittent and geographically constrained, resulting in temporal and spatial gaps between the energy availability and the consumption of the end-users." This can be found in Energies 2024, 17(12), 2917. Fuel cells have been extensively employed, and are one of the most promising alternatives for the future power sources for electric propulsion applications and distributed generation systems.

Response: Thank you for the comment. The energy consumption with biogas+SOFC complies with the requirements for low-emission hydrogen, as the raw material for biogas production is local forest (biomass) wastes in this. In our case, either renewable heat and electricity is produced, adapted to the energy needs for a data centre.

Comment: Some economic results need to be provided in this manuscript.

Response: Thank you for the comment. We fully agree with the relevance and usefulness of including economic results in research. As stated in the LIMITATIONS AND ASSUMPTIONS section, future research will focus on other issues, including "an economic assessment throughout a life cycle costing (LCC) methodology to consider a complete study from a whole sustainable perspective".

Comment: The innovation points of this manuscript need to be further clarified.

Response: Thank you for this observation.

The main innovation of this article lies in the proposed methodology for social life cycle assessment (sLCA) and the development of its indicators, enabling the comparison of two scenarios that meet the same energy demand. However, incorporating both environmental and social analysis allows for a broader evaluation of sustainability, assessing it within the context of a pilot-scale demonstrator that is part of a large European project. It aims to verify the reduction of environmental impact by implementing renewable energy sources and utilizing residual heat, compared to conventional fossil fuels. If a reduction in environmental impact is demonstrated, it must not come at the cost of increasing social impacts. Therefore, addressing these issues jointly is necessary, which constitutes the primary contribution of this research.

We have included a paragraph in the introduction section highlighting this point.

Author Response File: Author Response.pdf