[RE-wrenches] Demand Charge Reduction by PV

Matt Lafferty gilligan06 at gmail.com
Fri Mar 19 12:26:00 PDT 2010


Wrenches,
 
I agree with Joel.
 
This is my policy on the matter: Consider any Demand Charge reductions to be
bonuses in favor of the customer. I think it is OK to say that there MAY be
some savings, but DO NOT try to guarantee or insinuate that there WILL be
any Demand related savings. 
 
Here's why: UNLESS you have multi-year interval data for the site... AND the
ability to accurately interpret it.... AND the facility has a favorable
tariff... AND a site with a very predictable load (think in terms of being
able to predict the 15-minute interval which will set the Demand Charge
within +/- 1 hour)... AND the 15-minute period which would otherwise set the
Demand Charge coincides with a period when the PV system is operating at a
predictable output.... AND your overlay of predicted generation on top of
predicted load indicates a FANTASTIC Demand reduction... AND your proposed
system has multiple inverters (the more the merrier here)... AND the weather
is reliably predictable around the 1-hour period you predict the "new" Peak
Demand to occur... Don't bother having the conversation or spending
otherwise productive time analyzing the matter... 
 
If all the the ANDs above are true and you really want to go thru the
exercise, for whatever reasons you may have, be sure to consider the
following: In order to "prove" whether or not there was an effect on the
Demand Charges due to the PV system after the fact, you will need to have
interval data for both the generation and the facility. You will need to be
able to evaluate whether or not the customer's load profile, independent of
the PV, changed from the predicted values and, if it did, what the reasons
for that were. For example, if they increase or decrease their loads
independent of the PV, the net change is due to the customer's actions and
the PV. Once you have determined these factors, then you can begin to
calculate the effective "value" of the PV was on a Demand Charge basis.
 
It is possible and, in fact, likely that the PV will have some positive
effect on the Demand Charges for a given facility. It is, however, very
difficult to predict what that will be unless all of the ANDs noted above
are true. One of the keys driving the final calculation is the ratio of the
PV system's power rating to the customer's load coincident to the 15-minute
interval when the Demand Charges are set. The larger the system is, compared
to the facility load, the more likely and greater the savings will be.
Embedded in this relationship are the seasonal and hourly load profiles of
the facility, the reliability of the weather during this period, and how
closely they line up with the production profile. If you are intending to
"guarantee" some number of kW reduction, be sure you have multiple inverters
on the project. The more inverters you have, the more confidence you can
have in your predictions. 
 
Over the years, I've spent plenty of brain-damaging hours (weeks & weeks)
working on this. For utility companies, large integrators, and small
integrators. Before and after installation. With and without complex load
and generation profile data. On facilities with loads of all magnitudes and
hourly/seasonal profiles. The end result is the same... The effective Demand
Charge savings due strictly to the PV is relatively small and terribly
difficult to predict from year to year. I've had to analyze and report on
"why" the mucky-muck MBA & Engineers' predictions were so far off from the
actual experience after the projects went in. In the minds of these folks,
it seems like a no-brainer and just another column in their spreadsheets
during the sales cycle. It's a lot more complicated than that. I know of
several institutional customers who bought into paying the integrator for
these Demand savings on projects and have walked away from doing it on
future projects. The ones who have bought into paying for projected future
Demand reductions in the up-front cost of the project have regretted it. The
ones who have agreed to pay for it as a line-item on a PPA bill, aren't
doing it in the future. It's simply too much brain-damage. Too complex. If
they are going to have to pay somebody for it, it's a lot easier to pay the
utility company.
 
I've had to analyze and report on cases where the facility Demand Charges
have increased after installation of PV systems. After you isolate the load
and generation factors, and demonstrate that the customer's load profile has
changed from the historical (generally larger loads later in the day when
the PV can't help as much), you can do the dance of trying to show them that
they avoided something and trying to quantify exactly what that is/was. A
lot of Demand tariffs actually punish lower kWh consumption, so you have to
factor that negative impact into the value analysis. A LOT of time goes into
this when you consider all the communication, data gathering, analyisis,
presentation, and negotiation. Remember, it only takes one 15-minute period
that occurs outside of your effective generation curve to blow the whole
theoretical thing up.
 
IF you are set on going thru the exercise with the intent of
predicting/promising something, you need a facility with a very stable load
profile that occurs during peak PV periods, stable utility voltage, a tariff
that doesn't punish lower consumption, reliably predictable & favorable
weather coincident with facility peak demand, a large PV to Load ratio
(>50%), and multiple inverters. The combination of requisite weather,
PV:Load ratio, and favorable tariff pretty much makes the number of
facilites, which you can model accurately, very small.
 
I recommend going to the beach or mountains instead.
 
$0.02001
Solar Janitor

  _____  

From: Joel Davidson
Sent: Thursday, March 18, 2010 7:47 PM
To: RE-wrenches
Subject: Re: [RE-wrenches] Demand Charge Reduction by PV


Hello Peter,
 
I have seen 40% to 70% monthly demand charge reduction for some southern
California PV projects for some months, but it is still a crap shoot. 15
minutes and 1 second of clouds during the peak demand period will trump a
client's energy management efforts unless they are willing and able to shed
loads during cloudy periods. I tell clients that they cannot rely on the
weather to cooperate, to monitor and control their demand, and to think of
any PV savings on their monthly demand charge as a windfall.
 
Joel Davidson
 
----- Original Message ----- 
From: Peter Parrish <mailto:peter.parrish at calsolareng.com>  
To: 'RE-wrenches' <mailto:re-wrenches at lists.re-wrenches.org>  
Sent: Thursday, March 18, 2010 9:38 AM
Subject: [RE-wrenches] Demand Charge Reduction by PV

I failed to clean up the subject line on this post a few minutes ago. Please
respond to this post so that we can keep track of the topic properly.
 
Esteemed wrenches,
 
I have been wrestling with this concept about as long as we have been in
business. How to estimate how much a pv system will reduce the demand charge
for a customer.
 
I know the "worst case" goes as follows: 
 
(1)     Demand is based on measuring the consumption every 15 minutes and
keeping track of those numbers for the entire billing period.
(2)     The customer gets socked with a demand charge that is based on the
highest 15 minute consumption for the entire billing period.
(3)     The customer also gets soaked with a "facilities charge" that is
equal to the greatest monthly demand number for the trailing 12 months.
(4)     Now you have a solar system pumping out Wac varying over the
familiar bell-shaped curve during the day.
(5)     In the southwest US, peak demand typically occurs early in the
afternoon in the summer, during the week. Our LADWP has a mantra that goes
something like this, "Peak demand occurs at 3pm PDT on the third Thursday in
August!" I believe them.
(6)     So one would expect something like 40% of the peak Wac to offset the
peak demand, but what happened if the sun goes behind a cloud for those 15
minutes? Answer, "Bad luck. Your demand is back to what it was before you
bought your solar system.
(7)     It is actually worse than that. Peak demand recurs with
approximately with the same value with some regularity for an extended
period of time, so the sun will have to shine with full intensity every day
when peak demand is expected to occur, which in LA could be every day (M-F)
of the 30 day billing period.
 
I have always taken the position that we can't guarantee that any of the
demand charge will be reduced with a solar system. But what do other PV
integrators tell there customers? Better yet is there any actual data on
demand reduction with PV systems? It seems to me that occasionally the
monthly peak demand will in fact be shaved by PV production, the question is
how often in practice?
 
I once thought of taking actual insolation data and comparing it with actual
demand data and doing a Monte Carlo simulation (throwing the dice = randomly
matching up demand data with solar production data) - but I haven't retired
yet.
 
I would love to hear what others are doing about this.
 
- Peter
Peter T. Parrish, Ph.D., President
California Solar Engineering, Inc.
820 Cynthia Ave., Los Angeles, CA 90065
CA Lic. 854779, NABCEP Cert. 031806-26
peter.parrish at calsolareng.com  
Ph 323-258-8883, Mobile 323-839-6108, Fax 323-258-8885

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20100319/8bce0bff/attachment-0003.html>


More information about the RE-wrenches mailing list