I used to think exactly like Kirk, in that I would design the DC array size to
match the inverter size, figuring why clip potential output. However, when you
have 1000W/m2 shining down on the modules, a crystalline silicone module will
heat up about 30C above ambient, and dropping your output by about 12% to 15%.
As outlined by Chris, adding other loss factors and you will probably never
clip in the summertime even if you are 15% greater with DC module rating vs AC
inverter rating. The benefits of keeping a higher DC power to AC power ratio
allows the inverter to operate closer to peak efficiency during low light
levels, startup and near power down conditions. Also, as the modules degrade
over time, the ratio will drop. So does having the added available power
earlier in the morning and later in the day keep the inverter on longer to
produce more power during those times than the power lost due to noontime
clipping in the winter? Conventional wisdom says yes. (Here I am trusting the
inverter sizing software designers, and my own experience. The only other
drawback is additional heat buildup on the heat sink due to the extra power the
inverter is pushing through, which the inverter should be able to handle.
Overall I think you are best to design residential systems with a ratio of 1.12
+/- .02 to 1 of DC array size to inverter AC rating. For commercial systems go
a bit higher 1.2:1 to 1.25:1.
Mike Kocsmiersky
Principal
Phone: 413 883-3144
Spirit Solar
www.SpiritSolar.net
From: Chris Mason [mailto:cometenergysyst...@gmail.com]
Sent: Thursday, July 14, 2016 8:07 PM
To: RE-wrenches
Subject: Re: [RE-wrenches] array overloading of an inverter
In high temperature locations such as the Caribbean, where we are, the modules
run at about 40C during midday. Combine the temperature derate with losses for
dust, cabling, azimuth and tilt, it makes sense to add ten to twenty percent
more PV.
Given that the string design is never perfect, I would rather go a little over
than way under. When we were installing 240W modules on European Sunny Tripower
20KW units using 1,000V PV designs, we could either go with three strings =
18KW or 4 strings = 24KW. The three strings would have seriously underpowered
the inverter.
On Wed, Jul 13, 2016 at 3:02 PM, Kirk Herander wrote:
Hello folks,
I am composing a response to a question a potential customer asked me. It seems
a competitor is trying to talk him into a 20% larger KW array than the inverter
AC max output rating is. The idea of course is to generate more power on either
side of peak output, but at a cost I don’t feel is justified. My opinion of
this particular big-box installer I will keep to myself. My response to the
customer, trying to keep it simple:
“On the DC array input side, most inverters do allow an overload factor. For
instance, a 10kw AC inverter may allow for 12 kw of DC array as an input.
Whether or not this is a good idea boils down to economics and technical
reasons.
On a sunny day, the inverter generates power as a typical bell curve. Power
output rises in the morning, peaks at noon, declines in the afternoon. In my
example, the inverter can’t output more than 10 kw AC. What overloading the
input will do will widen the bell curve, i.e. generating more power in the
morning and afternoon, BUT clipping the peak at 10kw on either side of noon. So
there is power to be gained in morning and afternoon, but peak power is lost(if
conditions allow the peak output to be reached), since the 12 kw array can
never be converted to more than 10 kw of AC power. Depending upon time of
year(ambient / cell temperature) and weather conditions, the peak may be
clipped at 10 kw for several hours a day. So you are both gaining and losing
power using this method. And typically the inverters are only overloaded in
this manner on large-scale farms where the economics are favorable.
In your case, if you could actually put 200kw of DC array into 150 Kw of
inverter, the economics would never justify it. That extra 50 kw of array would
cost you $100k of more, and the dollar payback for the power that extra 50Kw
would generate will take 2 – 3x the time that the array size does that stays
inside the output limit of the inverters. This is why I’m not a big fan of
dramatically overloading the inverters, if at all, in your case. Any KW
portion of the array which is above the nameplate kw rating of the inverter is
going to have a longer payback for these reasons. “
I feel my reasoning is sound, But I don’t want to be too loose with the facts.
Comments are appreciated. Thanks.
Kirk Herander
Owner|Principal, VT Solar, LLC
Celebrating our 25th Anniversary 1991-2016
www.vermontsolarnow.com
dba Vermont Solar Engineering
NABCEPTM 2003 Inaugural Certificant
VT RE Incentive Program Partner
802.863.1202
___
List sponsored by Redwood Alliance