Jump to content

Project Settings > lowest temperature for "warm" countries


Recommended Posts

Good morning,

I have re-created some existing PV plants around Australia (near Canberra or near Port Augusta) and each time the Voc at -10°C is way bigger than the inverter absolute max input voltage. I usually have to adjust the "Lower temperature for Absolute Voltage limit" in the Project Settings, so that it would accept the sizing of the system as it is in reality (Voc temp. adjusted at -6°C for Canberra, and at +5°C for Port Augusta!).

Being mostly involved in projects in Africa or Australia or on tropical islands, these are warm places where, I assume, the -10°C best practice wouldn't apply.

My question is: is there a rule of thumb to know what's the Lower temperature for Absolute Voltage limit acceptable for a particular place? (e.g. take the average temperature of the coldest month and apply -10°C, or related to the latitude, or to the yearly avg. temp. or something like this)



Example with Voc at -1°C:


Edited by julmou
Adding illustration
Link to comment
Share on other sites

This is often covered by national standards or national annexes to international standards. From memory (a few years ago so might be wrong), the AS/NZS for PV system design specifically references design temperatures. Beyond that is a matter of contract and engineering judgement.

Link to comment
Share on other sites

Here is one reference. Select from here temperatures under the "extreme annual design conditions". From there, it is up to the designer to select. They generally go from least to most conservative starting with the mean on down towards n=50 years. I personally wouldn't select the mean or n=50, however. I usually go with the 10 or 20 year minimum depending on other factors. 



Edited by dtarin
Link to comment
Share on other sites

Thanks a lot, I wasn't aware of the Australian standard AS/NZS 5033, that's really helpful! I couldn't get my hands on the 2021 version, but I got the 2005 version and the amendments of the 2014 version in a separate document.

Well, for the case I'm studying, the inverter specs (Sunny Central 2500-EV) are "Maximum MPP Voltage: 1425 V" and "Absolute max. PV Voltage: 1500 PV", so I really think getting a good estimate of the max. voltage of your array is crucial for safety aspect and for not wearing off the inverters too soon after already a couple of years.

I am using two types of modules in this study: JINKO SOLAR 325 Wp PV MODULE (JKM325PP-72-J4V)    and    JINKO SOLAR 330 Wp PV MODULE (JKM330PP-72-J4V), with respectively Voc = 46.7 and 46.9 V, datasheet is attached.

The first version of the standard was pretty conservative (1.2 * Voc) and I would get an array max. voltage of 1681 and 1688 V, too high for our inverter.

Since the 2014 revision, if I use the method 1, I get 1514 and 1520 V. And if use method 2, I get 1569 and 1576 V. With PVsyst, I get an array max. voltage of 1521 and 1528 V. Now, with these 3 methods though, I used -1°C as the Tmin ("Lower temperature for Absolute Voltage limit", "lowest expected cell temperature in °C"). In my case, I inspected the ambient temp. values in the MET file (hourly values) and found the lowest one to be -1°C, at 6am, so just as the sun is about to rise.

Obviously, the engineers designing the plant opted for a higher Tmin value (which feels dangerously close to me, but let's move on). It really seems to me, the key here is how do you choose your Tmin?

  • do you rely on site measurements? (at least a year? more than a year?), and if you don't have site data, do you take the nearest weather station?
  • can you base your Tmin on the hourly values of the .MET synthetic file? I mean, even if it's synthetic, it is still based on 10+ years of recorded data, so you can expect the generated hourly values to be reliable, can't you?
  • do you really have to take the lowest recorded temperature from hourly values (synthetic or measured), or would it be more reasonable to take like an average of the 10 or 5 lowest values, or the lowest monthly value?

-> I am genuinely curious how to decide the Tmin for a site. Thanks in advance for your help.



Link to comment
Share on other sites

If, for some reason, you're not using the values set out in a local standard, then selecting the Tmin is indeed a task in itself, and would in my view need to be justified considering all available data plus a reasonable (there's that engineering judgement again) headroom depending on the quality and quantity of that data (thirty years of hourly data from a meteo station a mile away in uniform terrain is rather different to a synthesised TMY monthly from a satellite). You could as part of this go back to basics and model the actual array voltage over the time series of data that you have, your specific design etc with an allowance for manufacturer tolerance.

I'm assuming you're a designer; While you might be happy to keep that internally during initial sizing, at some point I would advise that it is specifically agreed with the client and also the equipment suppliers as warranties / performance guarantees will likely be at stake.

Link to comment
Share on other sites

One year of data is not sufficient to establish what an outlier condition is (what if it is warm that year?), nor would TMY data be sufficient. A TMY (synthetic or based on time-series) contains "typical" values, and not what you are looking for when designing a system to protect against damage in the rare situation where you have both low temperatures and sufficient irradiance. Ideally you want a few decades of temperature data. I've provided one resource which gives you several different timespans containing the low temperatures. I am not sure if NREL or NOAA has data for AUS, but there might be some public source for weather data similar to what we have in the US. 

Edited by dtarin
Link to comment
Share on other sites

Thanks again for the link, this is a great resource. I'll save it in my bookmarks. From the two projects I studied in Australia, at Royalla Solar Farm (near Canberra), they opted for a Tmin of -6°C, which correlates with the min temperature at 5 years from your screenshot. With Bungala Solar Farm (Port Augusta), they opted for a Tmin of +5°C (which, I already mentioned, seems dangerously high to me), and the ASHRAE data gives extreme min temp at 5 years -1.5°C, and at 10 years -2.3°C (see screenshot below).

It does feel to me though, that most design decisions on this aspect are not really conservative (often matching the min temp at 5years rather than 10 or 20 years)... ?

Would we be right in thinking that, if a manufacturer gives a max. input voltage Vdc-max at 1500 V in the datasheet for an inverter, that there is some tolerance margin and that the equipment could handle, say 1520 V, once or twice or year?

The PVsyst Help (Project Design > Grid-connected system definition > Array voltage sizing) mentions:


- Absolute Cell lower temperature for determining the Maximum possible voltage of the array. The default is set to -10°C for most European countries (best practice rule).

 For this limit, the cell temperature is considered as the ambient temperature (worst case when the sun suddenly appears on the field).

 You should define the lowest temperature ever observed during the day for this site.

-10°C being the default value, if we observe the Geneva data from ashrae-meteo.info/v2.0, here again it matches more the min value at 5 years rather than at 10 or 20 years, so not that conservative.

I find it interesting/strange that this is accepted as such a broad "go-to" value for most of Europe, where the lowest temperature observed during the day at Geneva or in Spain or in Hamburg, is obviously going to be very different (screenshot below for Hamburg, where -10°C really doesn't seem appropriate)... ?

It would be interesting to know the thoughts of a PVsyst moderator too on this topic of choosing the right "Lower temperature for Absolute Voltage limit"  ☺️


Bungala SF, Port Augusta:


Geneva Cointrin, Switzerland:


Hamburg, Germany:


Link to comment
Share on other sites

Hi ! For us -10° is "just" the default value, and indeed I agree that in several places in Europe you may want to choose something lower than that.
Of course in the end it is the responsability of the system designer to ensure that there is no issue, so being more conservative seems best (i.e. 20 or 10 years rather than 5). It is all a matter of how much risk you consider acceptable.
Of course, economic considerations however will tend to pull the value upwards... but at the cost of increased risk.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Create New...