# 165

Optimal Location of Solar Photovoltaic Plants Using Geographic Information Systems and Multi-Criteria Analysis

Dear Julio Manuel de Luis-Ruiz, Benito Ramiro Salas-Menocal, Raúl Pereda-García, Rubén Pérez-Álvarez , Javier Sedano-Cibrián, and Carolina Ruiz-Fernández

I read your paper

Optimal Location of Solar Photovoltaic Plants Using Geographic Information Systems and Multi-Criteria Analysis

My comments:

1- In page 2 you say “The electrons are able to transform and become part of a current in an electrical circuit”

In my opinion electrons do not transform in anything. The system works by using the photoelectric effect in some materials, that capture photons from the sun radiation and releasing electrons which constitute the electric current

2 “- A multi-criteria analysis is proposed to analyze large extensions of land with ten duly weighted criteria”

How do you know that they are duly or properly weighted? Certainly, not by using subjective weights

3- in page 3 “This methodology proposes a novel combination and weighting, using a statistical procedure to evaluate the consistency of the weighting

What is ‘consistency of the weighting’? And why it is important according to some MCDM methods? For me they are only useful in compensation, not certainly applicable in MCDM, because it does not make sense to decrease the importance of say criterion ‘Health’, and increase a criterion like ‘Job creation’

4- Page 4 “The difficulty lies in choosing the representative temperature or parameter to use to assess this criterion”

This is a good point and you can, for each site, take a range of two values. That is, for the same ‘Temp. criterion’ you can consider both ends, and let the method decide the best intermediate value

Why and how? Because you are not working in isolation, you must link those intervals with all other criteria intervals, if any. Maybe that for a location you have a daily average value of say 20 oCentigrade, however, it could be that compared to another location that has say 17 o, and because of that offering a better efficiency, the second location may give a best qualification than the first.

You certainly consider this, when said that the analysis must be done simultaneously. In so doing, there is a clear difference with other methods and authors that consider each criterion separately, as AHP does.

Why do we have to take into account all criteria and alternatives simultaneously? Because a decision matrix is a system, and as that, a change in any criteria may influence others.

5- Your concepts on humidity importance are very valuable, because in general, people do not consider it.

6- Page 5 “Selecting the criteria whose analysis is most convenient to find the optimal location is the first step to carrying out a correct study using GIS”

Criteria selection depends on the stakeholders, not on the DM. The financial guy will be looking for maximum return but the engineering guy will be looking for efficiency. Theirs is t he responsibility to select the criteria. It does not matter if there is disagreement between them; they must put what they want and it is not important if it contradicts another area criteria. Leave the M CDM method to find a compromise solution.

7- You speak of weights. They are good to appraise subjectively the relative significance between criteria, especially in simple problems. In reality, they are not weights but trade-offs, and they do not play any role ion selecting alternatives, unless they are objective.

8- Your description of the different weights and explanations are no less than excellent. This could be a very useful aid to readers and practitioners. You established intervals for every criterion, and this is very valuable information. Sincerely, it is the first time that I read a paper of this quality in this aspect

9- Page 9. You use the AHP method that is completely inadequate for this type or work. First because it employs pair-wise comparisons and assigns a preference value, which in my opinion, and in the many researchers, is a fallacy. What it produces is a set of invented weights and on top of that FORCED to respect transitivity. There is not a single mathematical axiom, let alone a theorem, that supports these extravagances

Saaty, the creator of AHP and ANP, said that you cannot use AHP when criteria are related as in this case, and without a doubt, he was right

10- “Thus, each criterion is given its own weight to model the optimal photovoltaic plant location, eliminating subjectivity “

This is inexact for three reasons:

a) The criteria weights do not model anything, and 99% of MCDM methods are very far from a correct modelling. Weighting criteria only produces trade-offs, good for compensation, but they are not even weights.

b) In MCDM there is normally not optimality but compromise, or equilibrium.

c) Eliminating subjectivity? This is a contradiction, since AHP is based on subjectivity

11- Page 10 “Consistency is considered to exist when the consistency ratio does not exceed the percentages shown in Table 12. If it is met, the matrix is consistent, and the criteria can be weighted. If not, the matrix of compared pairs should be reevaluated”

And what mathematical axiom or common sense say thatthere must be consistency in the DM estimates?

In addition, as you say, if there is no consistency the pairs should be revaluated. In other words, the estimates must be consistent, and it is IRRELEVANT what the DM estimated, who must modify his own estimates to conform a formula, like it or not. Not very mathematical indeed!

What happens if the DM negates to do the correction? He cannot continue, how do you call this?

12- Page 10 “. In order to validate the methodology”

Sorry, this is very common misconception since no MCDM method can be validated, because to do that you would need to compare with something that does not exist, liker a yardstick. Think that if you knew the result, the use of MCDM methods would be unnecessary.

13- Page 12 “The next issue requires detailing and classifying by ranges all the criteria raised in the analysis. In this case, it is proposed to score the most suitable areas with a value of 5, while the least favorable areas for installation will be rated 1

But, scoring the most suitable areas is precisely the task of a MCDM method. I agree with your classification for criteria, very good indeed, but it appears that you do not consider that for each location there could be excellent performance values for some criteria and not so good for others.

For instance, you may have very low humidity, which is highly positive, but not good irradiation, compared to other locations, which is not so good, or may be far from the grid, and so on. This is precisely the task of MCDM, and for this reason, a method must compare all data simultaneously.

14- By the way, your maps are noting less than excellent.

These are my comments. I hope they can be of help

Nolberto Munier