Proposals for Updated EMC Standards and Requirements (9–500 kHz) for DC Microgrids and New Compliance Verification Methods
Round 1
Reviewer 1 Report
The paper attempts to develop electromagnetic compatibility standards for DC microgrids, but the focus of the content is not prominent and the research innovation is insufficient. The paper designs an arc detection method and impedance measurement method, but both are common solutions, and the experiment lacks persuasiveness.
I think this manuscript don't provides much useful information to readers.
（1）The arc detection method lacks comparison with traditional methods.
（2）The experimental results lack key indicator parameters.
In summary, the key innovation points of the paper are not clear.
Author Response
We thank the reviewer for the constructive feedback we have received from the thorough review of our work. Please see the attached response letter
Author Response File: 
Author Response.pdf
Reviewer 2 Report
It's a good paper.
I would ask authors to investigate EMI problems to show why authors did this work
Author Response
We thank the reviewer for the constructive feedback from the thorough review of our work. Please see the attached response letter.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper considers a multi-faceted problem of identifying limits for DC grid disturbance suitable for the operation of some connected devices and techniques, and introduces a concept of minimum impedance to limit such disturbance. The biggest problem is that the problem is wide and complex, and the Authors provide just a few elements of discussion, drawing conclusions then on the whole, also imposing limits on impedance that imply an increase of cost, size and weight for all connected loads.
1) It seems that the due distinction between differential- and common-mode emissions is missing. RF conducted emissions are traditionally measured as common mode components; conversely in the supraharmonic interval (2-150 kHz) differential-mode components are getting attention for their amplitude. What causes disturbance to PLC systems, then, is a differential-mode component mainly. Finally, LISN are instead provided for the measurement of common-mode components, or better asymettrical voltages, that are not differential mode. See additional references on common-mode EMI and supraharmonics (usually differential).
2) Nevertheless compatibility of DC grid noise and distortion with the operation of PLC technology is a concern, as commented in sec. 2.1 of "Power Quality in DC grids", where it is also suggested the use of spread spectrum technology for many of the connected power converters.
3) Table 2 and Table 3 report a list of standards with their titles, that could be well read in the references, instead.
What is missing is a critical overview of such standards, deriving for example integrated limit profiles (to comply with applicable standards for one given application) and requirements for impedance (either explicitly given in such standards or calculated from limits and other information --- on this please see references [B] and [C]).
4) Figure 2 is relevant and would deserve some comments. One remark is the indication of power levels in kVA, as we are in a DC grid application. Could you comment as well on quasi-peak and average methods of evaluation, if compared to what represents a disturbance for arc detection techniques and PLC devices?
5) Line 208-209. The sentence on arc measurement falls out of the blue and it's symptomatic of a significant focus on it in the rest of the paper. Arc is a transient exceptional phenomenon that should be soon corrected by maintenance, so it cannot be considered as a main source of emissions, especially if so far you reported on quasi-peak and average methods (or detectors).
6) What are the settings of the results of Figure 3? You say they are not directly comparable to EN 61000-6-3 and -6-4 limits. Why? and why measurements were not carried out to be comparable? what is the purpose
7) Line 260-261. You speak of reserving a frequency interval for arc detection. You say at line 268 that you want to reserve 40-100 kHz for that. And that converters controllers work below 40 kHz. Isn't this limiting the technological progress as it would not be possible to push switching frequencies above 40 Hz? And what about the harmonics of those converters operating below 40 kHz? Take a look at the spectrum you measured in Figure 3(b).
8) Line 265. with "little used" you mean rarely used. Who decides how often? Who monitors that it is really used rarely? then what limit should be assigned to such loads?
9) Line 273. You reserve the band between 120 and 300 kHz for PLCs, whatever the modulation that is used. However, it is known that different types of PLCs have different levels of immunity to noise so that smarter PLC technologies should be promoted, rather than accepting all kind of PLCs reserving a large frequency interval for them.
10) Line 266. If the converter providing the DC supply for an entire microgrid is assigned a higher emission limit, that would disturb all loads significantly.
11) Sec. 3.3. You refer to the LISN values as a reference, but it is known that real grids have lower impedance values.
Now, you say that the minimum impedance should be ensured by inserting filters (additional inductance) at each load.
- this increases the cost, weight and size of such loads; take a look at the Victron converter you have tested, for example (it is 1.3 kg, 130x186x70mm size; the inductors should be sized for maximum input current that for a model of 240 W and minimum input voltage of 8 V is 30 A!!);
- you should justify quantitatively the minimum impedance, based on a model of propagation of disturbance along the grid, not just throwing 2 and 2.5 ohm, simply because you used them for your tests; [B][E]
- you need to demonstrate that the added inductance does not resonate with the rest of the system, including a variable cable length and capacitance in the DC grid (that may be small or large, tens or hundreds metres). [F][G]
12) Line 304. Write formulas in a better way, please.
13) Figure 6. Please, provide results of this criterion using some realistic examples, such as: take a small DC grid (assumed 20-50 m) and then a large one (assumed 100-500 m), calculate the impedance of cables and their capacitance; then think that the main DC converter at its output has additional capacitance. Then possibly with renewable energy sources there is more than 1 main converter. Resonances may occur (almost certainly), depending on location of loads and number/length of branches.
14) Figure 6 and line 334. Z_LI (line impedance) is never really purely inductive. Please, update your schematic, and your rationale, considering what happens due to distributed capacitance and losses (resistive). Consider that you are speaking of a wide frequency range where such inductive approximation rarely (or never) holds.
15) Section 5.1 now proposes a "novel arc detection" method, that is not in line with the scope and objective. By the way no comparison with other published research works is carried out to show novelty and improvement.
16) Figure 9 and 11 are the same. I would remove one.
17) Section 5.2. What does the title mean and what this subsection represents? An alternative method for arc detection? Or a complementary part to what proposed in sec. 5.1.
This section discusses very basic concepts and I do not see any added value. It should be removed.
18) Table 6. What source power means? Is it the delivered power? If so, it should be different for different load impedance values (that are in reality, I think, pure resistances not impedances).
19) Line 452. Picoscope has a FFT receiver?
20) Line 480. Experiments were done on different LISNs. Please, which experiments and what are the results with the various tested LISNs so to conclude that 10 uH is an optimum? See ref [N] for example.
21) References are old. There are some spot references of 2021 and 2022 not on the main topic of the behavior of DC microgrids/grids, for what regards disturbance and impedance.
A search of MDPI publications and others at large hit additional works below, that were sometimes referred to in the comments.
Note. Criteria used in the search: "DC grid" + "supraharmonics" in the Abstract, "DC grid"+"conducted emissions" in the Abstract, "DC grid"+"EMC" in the Abstract, "DC microgrid" in MDPI does not return results.
[A] (2022) Design, Analysis and Implementation of Bidirectional DC-DC Converters for HESS in DC Microgrid Applications, https://doi.org/10.3390/smartcities5020024
[B] (2019) Survey of network impedance in the frequency range 2-9 kHz in public low voltage networks in AT/CH/CZ/GE, https://www.cired-repository.org/handle/20.500.12455/401
[C] (2022) Harmonic and Supraharmonic Emissions of Plug-In Electric Vehicle Chargers, https://doi.org/10.3390/smartcities5020027
[D] (2021) Power Quality Phenomena, Standards, and Proposed Metrics for DC Grids, https://doi.org/10.3390/en14206453 (has collected 44 standards related to DC grids and microgrids)
[E] (2022) Characterization of Supraharmonic Emission from Three Different Electric Vehicle Charging Infrastructures in Time and Frequency Domain, https://doi.org/10.3390/en15020394
[F] (2013) Spread of high frequency current emission, https://doi.org/10.1049/cp.2013.0611
[G] (2022) Supraharmonic Emissions from DC Grid Connected Wireless Power Transfer Converters, https://doi.org/10.3390/en15145229
[H] (2021) Efficiency Optimization Design of L-LLC Resonant Bidirectional DC-DC Converter, https://doi.org/10.3390/en14113123
[I] (2018) Impedance Modelling and Parametric Sensitivity of a VSC-HVDC System: New Insights on Resonances and Interactions, https://doi.org/10.3390/en11040845
[J] (2021) Conducted Emissions Measurements in DC Grids: Issues in Applying Existing LISN Topologies and Possible Solutions, https://doi.org/10.1109/CPE-POWERENG50821.2021.9501169
[L] (2014) DC-link input EMI filter design in a centralized architecture PV inverter: Impedance approach, https://doi.org/10.1109/ECCE.2014.6954055
[M] (2019) Reference Setup for RF Impedance Measurementswith High DC Bias Currents, https://doi.org/10.1109/EMCEurope.2019.8872095
[N] (2021) Comparative Analysis of Conducted Emission of Off-Grid PV Inverter Using Different DC- LISN s, https://doi.org/10.1109/APEMC49932.2021.9596780
[O] (2012) Common-mode EMI noise reduction for grid-interface converter in low-voltage DC distribution system, https://doi.org/10.1109/APEC.2012.6165859
[P] (2018) Conducted Emissions on DC Power Grids, https://doi.org/10.1109/EMCEurope.2018.8485174
Please, note that your ref [2] (the paper by David Thomas) refers to AC networks, so that other references more closely related to DC grid behavior should be added (see list above).
English form has some minor issues, more related to unclear or inelegant expression:
Line202: ... provide the first glance at possible ...
Line 201: exemples
Figure 3: exemplary
Section 2.4: "needs for ..."
Line 258: various electrical devices
Line 265: large and little used loads // large is physical size, then little used ...
Author Response
We thank the reviewer for the constructive feedback from the thorough review of our work. Please see the attached response letter.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The objective of this paper is to propose EMC standards for DC microgrids.
A large amount of tests should be done in different temperatures, humidity, arc gaps to estimate the detection feliability. Arc detection method used in the paper is no novelty.
I suggest this article is not suitable for publication.
Author Response
We thank the reviewer for the second review. Please see the response letter.
Author Response File: 
Author Response.docx
Reviewer 3 Report
Dear Authors, thank you for your replies, argumentation and amendments. I recognize that in some places I misunderstood some of your sentences/figures and thank you for making them clearer.
I summarize the previous comments, and where the necessary I provide some additional information.
1), 2), 3) agreed, thank you.
4) I did not get that the figure was taken "as it is", I thought that it had been prepared for DC grids specifically, from which the remark on "kVA".
Additional consideration: in case of safety-relevant functions in some particular applications a peak detector is used, in order to have a faithful picture of the spectrum, including transients (as far as allowed by the RBW setting).
5), 6) Clear, thank you.
7) Thank you for the details provided. Indeed arc detection and narrowband PLC are quite exposed. What is worrying me is reserving an entire frequency interval, where smaller portions of it could be used instead.
8) ok.
9) Like 7).
10) Agreed on limits based on power for loads, but this is the power source for the entire DC grid, and would be allowed to disturb more than the rest of the loads.
11-i) The impedance of a small/medium grid may be quite variable (resonances, active devices like PV inverters in AC, DC/DC converters for storage in DC), but usually is on low values. Please, see some references:
a) https://core.ac.uk/download/pdf/250164667.pdf (Figures 4.7 and 4.9)
b) https://www.mdpi.com/2079-9292/8/10/1155 (Figure 9)
11-ii) The Victron converter characteristics was taken from the datasheet of the line of products in the order of 100-400 W (link: https://www.victronenergy.com/upload/documents/Datasheet-Orion-Tr-DC-DC-converters-isolated-100-250-400W-EN.pdf): e.g. model Orion-Tr12/24-10 .
11-iii) Weight, size and cost. The comment was general in that such factors were not evaluated or explicitly accounted for. Going to your example:
- Victron converter is : 130 x 186 x 70 mm , weight = 1.3 kg
- Bourns inductor is : 28 x 20 x 16 mm, weight approx 38 g
So 4 Bourns inductors would take 2.4% of volume and 11.7% of weight.
12) The comment was simply on the use of equation editor for example.
13,14) We must be careful using the "standardized LISN" because it is based on old measurements for AC networks, and then there was a reduction (a sharp factor of 10) in the inductance for the automotive. It seems that the interest of committees is on freezing some kind of network, rather than to have a really representative network and approach. Grids may have variable geometry and variable response, as commented before: it should be in some way taken into account in your approach.
15) Understood.
16), 17), 18), 19) ok.
20) understood.
English has improved, but there are still some minor problems.
Author Response
We thank the reviewer for the second review. Please see the response letter.
Author Response File: Author Response.docx
