Fire in photovoltaic panel 2

[Ciaci]

Hi guys,
I am making a fire investigation in Italy. In order to consider also that risk, I have to evaluating the volts and ampere in a photovoltaic plants.
First of all I have to specify that this fire started in the morining of the 28 december, so it is possible to assume that the solar irradiance was very low.
I tried to use ecotect to evaluated but then I found this web site

https://www.nrel.gov/midc/solpos/solpos.html

and I thought that to find the Extraterrestrial Global Irradiance on a tilted surface could be a good approximation. I assumed that if it is an "extraterrestrial" irradiance then this does not take in to account the absorption and reflection of the atmosphere so this could be a good safety factor.

But I am not sure about this output value. Does someone think that these could be realistic? Do you think that there is a good alternative to this website?

Do you think that is proper use the I-V curve of the photovoltaic system datasheet in order to evaluate the Volt and Ampere?

Thanks guys everything can help

 

[Skogsgurra]

I did a similar investigation in Southern Germany a few years ago and used a rough estimate that there was 1 kW/m2*cos(alpha), where alpha is sun's deviation from panel's perpendicular. I was not looking for power delivered from the solar panels, even if that was an important factor, but was looking for other root causes. So I didn't need exact numbers.

One thing that you should keep in mind is the nature of the tie inverter (or battery system or rotating machinery). In my (rather limited) experience, a solar panel does not set itself on fire, because of the current limiting characteristic, but it can be killed if there is a damage in the panel and the receiving battery or rotating DC motor (that drives the grid tie generator) feeds back into the panel without any fuses.

Some installations do not use fuses because of the current limiting V/I curve and then, a defect in the panels can cause a high current to flow back into the panel.

Systems with grid tie inverters (using a PWM inverter) cannot feed back and therefore doesn't have this problem.

Out of curiosity: is there a PWM inverter, a battery or a rotating DC motor in the system?



Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Ciaci]

It is an istallation were is used a pwm inverter. Which do you think are the risk in this installation?

 

[Skogsgurra]

How large is the installation? Are there many high-power solar panels in parallel?

That could be a risk in itself if there are no fuses or overcurrent protection (or diodes) for each panel. Then, a failure in one panel will be subject to current from all panels. Otherwise, a PWM inverter shouldn't be a risk per se.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Ciaci]

It is a 29 m^2 array composed of 18 panel. 9+9.

If it is possible I want to ask you another thing, the panel were SLA 245P (

http://www.silfab.eu/core/media/downloads/master/sla-p-225-245prot_eng_.1347524346,2999.pdf).

Know that for example the irradiance was 180 W/m^2, with 10°C. How can I know Volt and Ampere? I ask you that because in every solar panel datasheet there are I-V curve with at minimum 200 W/m^2, how can I extimate my data if I have a different irradiance?

You are very helpful Gunnar...thank you so much

 

[Skogsgurra]

Since current is proportional to irradiance, your I curve shifts down if irradiance goes down. So, at 180 W/m2, draw the I curve just below the 200 curve (at around 1.60 A instead of 1.75 A).

The idling voltage is not very irradiation dependent. Instead, it is temperature dependent and increasing when temperature decreases. That may be surprising, but is a result of the diode forward drop decreasing with temperature.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Ciaci]

Oh I see, thank you very much. But does these curves are empirical or they follow a specific equation that I can use to evaluate volt and ampere.
How can I see how much the voltage increases depending on temperature... Sorry for all thes questions

 

[irmalcol]

I'm not sure if you are asking specifically about this, but it is definitely possible for an individual solar module to light on fire. If there is a defect or damage in the module, a DC arc can form between two components within the module. Because the module is inherently current-limiting, a DC arc is often not caught by circuit protection devices (circuit breakers or fuses). A DC arc fault usually makes the damage worse, further widening the arc gap, creating more arcing, and spreading the fire (possibly to adjacent modules). When left undetected, DC arcs can continue almost indefinitely, since they do not self-extinguish when the arc gap becomes large (an AC fault self extinguishes when the gap is sufficiently large because the voltage is not sufficient to sustain dielectric breakdown of air after the voltage crosses zero (which it does many times per second with alternating current)).