Matching and Control Optimisation of Variable-Geometry Turbochargers for Hydrogen Fuel Cell Systems

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. It is recommended that the title is revised as “Matching and Control Optimization of Variable Geometry Turbochargers with Fuel Cell System”.
2. The abstract and conclusion should be rewritten, which should include both quantitative and qualitative analysis results.
3. For the equations applied in the modelling and the model of fuel cells, some relevant references should be cited, such as https://doi.org/10.1002/fuce.202200121.
4. The reason why two different models are proposed should be clearly stated.
5. To establish the models, the detailed parameters of FGT, VGT, FC stack, Turbine and other important components should be listed in a Table.
6. In Fig.6, the curve of FGT is not clear. Besides, the four comparative curves in Fig.11(a) are not clear. All the Figures should be revised.

7. The limitations and future work should be added at the end of the conclusions.

Comments on the Quality of English Language

 Reconsider after major revisions

Author Response

Dear Reviewer,

Thank you for reviewing our work and for the well-considered comments and recommendations, we have found them to be extremely useful in refining the manuscript. Please find below detailed responses to the individual comments.

“It is recommended that the title is revised as “Matching and Control Optimization of Variable Geometry Turbochargers with Fuel Cell System”.”

Thank you for your comment, we agree that the application to electric vehicles in the title is not required and the work is appropriate to a wide range of fuel cell applications. We updated the title as recommended to:

“Matching and Control Optimization of Variable Geometry Turbochargers for Hydrogen Fuel Cell Systems’”

“The abstract and conclusion should be rewritten, which should include both quantitative and qualitative analysis results.”

Thank you for indicating this. We have added important quantification to the more descriptive results in the abstract, conclusions, and generally throughout the manuscript. These include, for example: the 10.5% peak difference in results of the two modelling approaches (abstract, results), less than 0.25% increase in fuel cell system efficiency when using the VGT vs. FGT (abstract, discussion), and <1% flow through the wastegate (results, discussion).

“For the equations applied in the modelling and the model of fuel cells, some relevant references should be cited, such as

Thank you. For further information about the fuel cell including fundamental equations, parametrisation and validation of the fuel cell model, please refer to our previous work in reference [24].

“The reason why two different models are proposed should be clearly stated.”

Thank you. We have added more clarity on this to our introductory paragraphs of section ‘2. Materials and Methods’. In particular, we describe how the 1D modelling approach compliments the reduced order model with greater detail and more accurate results, but is less computationally efficient, so the different modelling approaches are leveraged for their respective advantages.

“To establish the models, the detailed parameters of FGT, VGT, FC stack, Turbine and other important components should be listed in a Table.”

Unfortunately, the components used in this work are covered under a confidentiality agreement and we are unable to publish detailed parameters beyond what is already present in the figures provided. For further information about the system including fundamental equations, parametrisation and validation of the fuel cell model, please refer to our previous work in reference [24].

“In Fig.6, the curve of FGT is not clear. Besides, the four comparative curves in Fig.11(a) are not clear. All the Figures should be revised.”

Thank you for your comment. We agree and have reviewed all the figures in the manuscript, replotting figures 6, 8, 9, and 11. A key improvement is the use of distinct symbols (x and o) for the two turbo architectures, which more clearly shows results that overlap, as is common in our data – as you will see in Figure 6, for example. Additionally, we have combined the content of Figures 11 and 12 and divided the two turbo architectures so each plot only compares the two modelling methods. Now, the plots in figure 11 compare only two datasets compared to the previous four. Throughout the manuscript, we have also adjusted the axes’ limits where appropriate, to more clearly show the minor fluctuations. Overall, the comparative curves present our data more clearly.

“The limitations and future work should be added at the end of the conclusions.”

Thank you for this comment, in revising this section, we found that the conclusions, limitations and future work are closely related to one another so it makes sense to discuss them together. We have renamed the final section “Conclusions, Limitations & Future Work”, and added a summary of the key findings to the beginning of this section.

Thank you again for your comments and suggestions, we hope that the explanations above, along with the revisions to the manuscript are satisfactory.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript describes the computational study of the hydrogen FCEV systems with FGT and VGT. The topic discussed in this manuscript is quite interesting. However, the obtained results are disappointing. I fully understand that the conclusion of "no difference" is one form of the scientifically meaningful finding. But the results presented here don't seem very useful. With this observation I believe this manuscript is not suitable to publish on this journal.

Some major issues to be addressed/revised are:

1) I strongly recommend the authors to add "Nomenclature".

2) The basic rule about the abbreviation (write down in full at the first appearance and use the abbreviation thereafter) should be followed. For example, EGR in page 2 or FCMT in page 3.

3) Equation 3. M_{air}, not M_{O2}?

4) Equation 6 should be (p_{out}/p_{in}), not Δp (=p_{out}-p_{in})? Or maybe Δp defined as p_{out}/p_{in}?  

5) Is Equation 7 correct? I believe it should be T_{out}-T_{in}.

6) In Equation 11, the efficiency η must be multiplied, not division.

7) What are the specifications of the fuel cell stack? What are the operating conditions of the fuel cell stack at design point. In Figure 6 the fuel cell power is up to 350kW. Is this heavy duty vehicle?

8) I wanted to see the performance map of the turbines (FGT and VGT) only, in particular how much performance changes in VGT when the geometry is changed. Without such information it is hard to understand why nozzle gap actuation changes only in a small range.

9) In most operating conditions of a fuel cell, in general, the cathode outflow is saturated (or close to saturation), and therefore it is very likely to have water condensation. This causes large performance loss of the turbine. I would strongly suggest the authors to include the impact of the condensation (maybe in the future studies).

Author Response

Dear Reviewer,

Thank you for reviewing our work and for the well-considered comments and recommendations, we have found them to be extremely useful in refining the manuscript. Please find below detailed responses to the individual comments.

“The manuscript describes the computational study of the hydrogen FCEV systems with FGT and VGT. The topic discussed in this manuscript is quite interesting. However, the obtained results are disappointing. I fully understand that the conclusion of "no difference" is one form of the scientifically meaningful finding. But the results presented here don't seem very useful. With this observation I believe this manuscript is not suitable to publish on this journal.”

Thank you for your comment, this is a very valid point and also was our first impression of the initial results. However, after deeper investigation, particularly the differences observed in the sensitivity study, we do not feel that these results are simply a ‘null’ finding (although as you mention, there are benefits in publishing ‘null’ findings).

Firstly, there is currently considerable debate in the literature about whether variable geometry offers a tangible benefit over fixed geometry turbines for fuel cell applications. In our literature review (lines 68-94), we have discovered that some authors (e.g., Schoedel et al. [35]) have demonstrated large potential efficiency gains, some find marginal differences (e.g., Filsinger et al. [27]), while yet others have found that the efficiency of VGTs is in fact lower than that of FGTs (e.g., Zhang et al. [15]).

Our approach differs from previous work in that we attempt to eliminate potential causes of these discrepancies by;

1. Optimising the mass air flow and pressure to ensure that both the FGT and VGT were able to operate at their maximum system efficiency across the range of current densities.
2. Ensuring that both the FGT and VGT are appropriately sized for the application.
3. Using an identical map, representing the VGT ‘flush gap’, for the FGT. Essentially, isolating the ‘variable’ aspect of the VGT.

We argue that it is, in fact, a very interesting result that when these considerations are taken into account, not only do we get almost identical performance, but the optimisation of both turbines’ operating points converge on almost identical flow rates and pressures across the vast majority of the operating range. In our opinion, this conclusively answers the question that varying the geometry does not provide an efficiency benefit, at least for steady state operation.

In addition, we do identify a potential benefit of VGT in our sensitivity study results (Figure 10). Here, VGT shows a significantly reduced sensitivity to the operating point and is able to maintain high efficiency over a larger region than FGT. This explains the slight difference observed in operating point at the extreme ends of the current density range, and also has potential for improved transient efficiency, for example under real-world conditions. This has been identified as key opportunity for further research. It is also good to see that, despite our differences in approach, these results correlate well with the findings of Filsinger et al. [21], giving us confidence in the overall conclusions.

In response to your comment, we have highlighted these key findings at the beginning of the conclusions section, please see lines 419-433.

“I strongly recommend the authors add "Nomenclature"."

Thank you, we agree, and we have added Nomenclature and list of Abbreviations at the end of the paper as per the journal’s template.

“The basic rule about the abbreviation (write down in full at the first appearance and use the abbreviation thereafter) should be followed. For example, EGR in page 2 or FCMT in page 3.”

Thank you, we agree and have proof-read and rectified this issue.

“Equation 3. M_{air}, not M_{O2}?”

Thank you for highlighting this. For clarity, we have clarified and corrected Equations 2 & 3 as follows:

Equation 3 now describes the mass flow rate of oxygen required for the stoichiometric reaction in the fuel cell. Equation 2 mass flow rate of oxygen along with the mass fraction of oxygen in air and the excess oxygen ratio to calculate the mass air flow required for the compressor.

“Equation 6 should be (p_{out}/p_{in}), not Δp (=p_{out}-p_{in})? Or maybe Δp defined as p_{out}/p_{in}?  

“Is Equation 7 correct? I believe it should be T_{out}-T_{in}.

“In Equation 11, the efficiency η must be multiplied, not division.”

Thank you, in answer to the three errors in the equations listed above, all three corrections are accurate, and the manuscript has been updated accordingly. In addition, we have checked for errors in the other equations throughout the manuscript.

“What are the specifications of the fuel cell stack? What are the operating conditions of the fuel cell stack at design point. In Figure 6 the fuel cell power is up to 350kW. Is this heavy duty vehicle?”

The fuel cell is a nominal 300kW which, as you mention, would be suitable for heavy duty applications. However, in answer to a comment by another reviewer, we have removed references to electric vehicles because we agree that the work is not constrained to a single fuel cell application. In terms of the operating conditions, during the matching exercise, the operating points are weighted as described in equation 15 and the subsequent paragraph (lines 213-218). During the operating point optimisation, each operating point, defined by current densities in the range of 0.05A/cm² to 1.7A/cm², is individually optimised for cathode mass flow and fuel cell inlet pressure, as described in Equation 16.

For further information about the fuel cell including parametrisation and validation of the fuel cell model, please refer to our previous work in reference [24].

“I wanted to see the performance map of the turbines (FGT and VGT) only, in particular how much performance changes in VGT when the geometry is changed. Without such information it is hard to understand why nozzle gap actuation changes only in a small range.”

This is an important point and one which we have considered. Unfortunately, the VGT used in this work is covered under a confidentiality agreement and we are unable to publish details of the maps. The VGT in question was originally designed for internal combustion engine applications. We can share, however, that peak efficiency of the VGT is just over 80% at ‘flush gap’, with an actuation of 51% as shown in Figure 7. Efficiencies of over 80% can be achieved for a wide range of nozzle gap actuations (from approximately 30% - 75%).

The question of VGT range was originally intended to be part of this investigation. The VGT strategy, shown in Figure 4, was derived from preliminary work where it was found that maximum efficiency was achieved when wastegate and back-pressure valve actuation was avoided unless necessary. We wished to understand whether the range of control authority of the nozzle gap (for a VGT designed originally for internal combustion engines) would be sufficient across the full range of current densities of the fuel cell, and subsequently whether the other actuators would still be required. As you can see from the results, the nozzle gap only varied between 48-59% across the full range of testing, and the other actuators were not required for the VGT.

Retrospectively, we have noticed that this broadly correlates with the results of Zhang et al. [15], Menze et al. [23], and Schoedel et al. [35]. However, these works involve bespoke VGT designs for fuel cells where the focus has been on such things as blade solidity and reduced gaps to the housing. As a result, the results are not directly comparable, and we have refrained from drawing detailed comparisons on this aspect within the manuscript, although we do discuss the range of VGT actuation on lines 368-377 in the discussion.

“In most operating conditions of a fuel cell, in general, the cathode outflow is saturated (or close to saturation), and therefore it is very likely to have water condensation. This causes large performance loss of the turbine. I would strongly suggest the authors to include the impact of the condensation (maybe in the future studies).”

Yes, certainly, this is a key area of interest for understanding the absolute performance of turbines when working with fuel cells and the effects on turbine work recovery are described in the work of Wittmann et al. [43]. However, further experimental studies are required before these effects can be fully quantified, and hence we focus on the comparative performance of VGT and FGT in this manuscript. As explained in Section 5 - Further work (lines 448-461), we provide suggestions of how our work might be combined with that of other authors to provide a deeper insight into this issue.

Thank you again for your comments and suggestions, we hope that the explanations above, along with the revisions to the manuscript are satisfactory.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

accept

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for providing thoughtful responses to my review comments.

The clarification and revision from the authors look sufficient, and now I think the manuscript is ready to publish.