Review of SAIP 2021 Proceedings paper
Title: Performance analysis of thin film photovoltaic (PV) technologies in an embedded generation network
Authors: Roodt et al

The authors present the results of a series of measurements stretching over several years where they seek to monitor the performance and possible degradation of solar modules representing three types of thin-film technologies.

The experimental work appears to have been carried out in a solid, systematic and professional manner, and so the results presented are to my mind reliable and worthy of publication. The study fits in well with the broader work of the same group, and as such contributes to the pool of knowledge generated in their niche field.

I have however identified significant shortcomings in the paper itself that will be detailed hereafter. Once these issues have been attended to then the paper should be acceptable for publication.
a) The Abstract needs to summarise the paper with a focus on the findings. Right now it reads like an Introduction and the start of the methodology section.
b) The major problem with the paper as it stands is that there is an almost complete lack of contextualisation of the experimental aims and findings in terms of physical theory and the present research situation. The Introduction is currently little more than a set of definitions, of which a large fraction are not used in this paper anyway (more on that in point (d)). There is no Discussion after the Results chapter, and the Conclusion is basically a repeat of that chapter as well. What the paper should be telling us is: i) what is the historical background of thin-film devices, and why have they not been the preferred option in recent times; ii) what are the advantages and disadvantages of thin-film devices compared to other technologies; iii) the study is about comparing CdTe, CIS and a-Si technologies, so we need a little bit of a background about their development, and again how they compare with each other; iv) after the Results there needs to be an exploration of whether the findings here are in line with previous research, and if not, a discussion on whether that is significant.
c) Partly related to the above, the paper is very thin on references (only 6 listed). Of these, only 3 are journal papers. There is a reference to PVEducation, which, while it is a very helpful resource, is not an appropriate scientific reference (all facts that are given there have been published somewhere in a journal). An MSc thesis is only a suitable reference if it contains finer details not given elsewhere, but that is clearly not the case with the material referring to it here. What is needed for this paper are more references to cutting edge articles in thin-film research, and these should tie in with the points made in (b) above.
d) The bold letter introductions to reference, array specific & final specific yield in 1.1 are not only never used later, but are also defined in a very confusing manner ("time in hours for which the PV array must operate ..."? It does not need to operate for hours to generate power). The performance ratio is simply (dividing equations 3 into 1) E_F/E_irr, which is perfectly clear, so why not just start with that.
e) Given that the performance ratio ultimately requires E_irr, it is not clear to me how this value was obtained. Was there a measuring device, and, if so, we need a line or two giving the details. If it was a from a solar irradiance estimate based on solar position and some model, then we again need a sentence or two explaining that.
f) What exactly does the last sentence of section 2.2 mean? This requires some elaboration (and again preferably not a reference that amounts to just a website).
g) I am a bit puzzled by the interpretation in the last paragraphs of both page 3 and page 5. If the drop in performance ratio is due to degradation, then why is there no more downward trend visible in Fig. 2? Why should degradation only be happening soon after 2017? Is this not perhaps due to something else? In the second line of page 6, was this effect as seen by Rawat et al for the same three technologies, or for what?
h) The plots. These appear to have been prepared with Excel with its default formating, which can easily lead to badly inappropriate graphs in a scientific context. This paper is a good example of how things can go wrong when one does that. The result is that the actual plots are sometimes so small that they make up less than 20% of the figure area. In particular:
   i) The ".00" in Figs. 2, 6, 7 and 8 are completely unnecessary, and just take up needed space.
   ii) The x-axis scale in Fig. 2 is odd. If these are indeed monthly averages then indicate which set of points corresponds to each month (impossible to tell right now).
   iii) Is it really necessary to put plots for every date on Figs. 3, 4 and 5? Many are invisible because they are effectively identical to those from other days (sometimes even the colour looks the same). I would just show two dates in Figs. 3 and 4 (2017 and one other) and three in Fig. 5 (2018, one from the middle loop and one from the lower loop)
   iv) There are currently 14(!!) curve labels in Figs. 3-5. Even when cutting out plots as suggested in (iii), there are still too many labels in my opinion. The dates also take up about 1/3 of the space in Figs. 6-8. There has to be a better solution, e.g. giving the dates (which seem to mostly be the same set in Figs 3-8) labels "a", "b", "c", etc. and identifying the date in the first caption where they come up. Or write them between the error bars, or any other practical solution the authors can come up with that displays their graphs to an appropriate size.
i) Excessive capitalisation. There are a long list of words in the text that have been written with an initial capital letter even though they should be in smalls. These include chemical elements (cadmium, etc.), power, current, voltage, array, ratio, photovoltaic (in title), standard, performance, meteorological, aerial and maybe some more that escaped me.
j) Excessive repetition of some terms, sometimes even in the same sentence: "technology" (appears 4 times in 1st 4 lines of the abstract - it is not even the right word as in some instances "modules" seems more appropriate) and "important" (appears 4 times in section 1.2 - it is in any case a term that should be used very sparingly and be reserved for the most crucial points).
k) Miscellaneous language points:
   i) 1st sentence in Introduction: "Thin-film PV module performance has ... materials. These don't offer ... is therefore important ... in an operational setting."
   ii) Line 3 in 1.2: "... one can detect any deviations ...".
   iii) Second-last line on page 2: "... technologies respectively."
   iv) Fig. 1 caption: "peak output power" rather than "size"
   v) Page 3, line 2: "... array were monitored ..."
   vi) Results line 3: I guess you mean "fault" rather than "error" (twice), and "fixed" rather than "corrected" (correcting an error implies that there is data).
   vii) Last line on page 3: "... in both periods" does not sound right. Maybe "... in both earlier and latter measurements"?
   viii) Section 3.2, 1st 3 lines: I think that the second sentence is redundant.
   ix) Page 5, line 4: "... the three arrays. The CIS and CdTe string FFs are ... a-Si string FFs start at 0.60 and decrease to ..."
   x) Page 5, line 7: the "." at the end of the line should be a ":", otherwise the next line is not a proper sentence.