Orientation to optimize for EV-A plan in San Francisco Bay area

[bcroe]

It's not just the mounting cost and complexity for additional rail and steel to orient them differently than flat against the roof. Yes, from a geometry standpoint you can put panels on any roof and point them in any direction with enough support. Say if you first laid them all side-by-side against the south roof first as in a typical install, then tilt each one to face southwest, you'd find that during your optimal hours, they would be shadowing each other. Even shadowing a tiny corner of a panel nearly kills all the panel output.

So you'd have to space them farther apart to avoid shadowing each other, esp during your peak production times. And there I think most residential roofs wouldn't have enough area for a decent-sized array with the spacing needed.

Hurray for ground mounts!

[J.P.M.]

I started this thread because i read this article.

http://www.nytimes.com/2014/12/02/up...abt=0002&abg=0

IMO, that article is a lot of simplistic B.S., the author needed a story and, knowingly or not, acted as a shill for the CA POCOs.

Some simplistic thoughts:

What follows is mostly for SDG & E cust., but probably applicable to at least some others and perhaps useful as illustrative to other situations.

This is a complicated subject but more manageable to my way of looking at it if taken piece by piece. I try to remember that things are changing fast in CA at least, and the future is never certain anyway, so the best I can likely hope for is to stick a wet finger in the wind and SWAG it. Thus:

One interesting factoid before I go farther, if I read the testimony correctly, an SDG & E employee named Leslie Willoughby, in testimony before the CPUC last Feb., stated that only about 1% (one percent) of SDG & E's residential customers are on T.O.U. and that T.O.U has been around since 1983. I'm checking on that. Seems surprisingly light.

As usual, there are, for starters, at least 2 ways to view things, the POCO's and the customer's.

The POCO's version may go something like this: Among other things, they'd like to get some way to maximize revenue, help balance/meet their peak loads, and as long as they're stuck with net metering, use it to their advantage. Incentivizing shifts in use patterns ( i.e. draconian pricing during peak use times) is one way to do this. Note that concern for the solar customer's generation revenue is probably not part of the thinking, nor should it be in any big way.

I'd assume the T.O.U. customer, solar owner or not, wants to minimize or at least reduce their bill. Note that the profoundly simple and most cost effective way to do this is to use less electricity. After that, the next best way (for the T.O.U. customer) is to avoid using electricity during peak hours, and confine use as much as possible to off peak or super off peak hours. Vampire laundry at 2 A.M. for example ? Maybe so, but in any case the amount of electrical use or use patterns have nothing to do with the capacity of a solar generating system to produce revenue under a T.O.U. tariff (or any other tariff for that matter).

For the T.O.U. customer who also has a solar energy system, one additional and perhaps helpful way to think about this business is to separate solar generation revenue from electricity expense, and think of the solar electric system solely as an income generator. To do this, the system probably can't be sized for more than 100% of the load.

Any properly operating grid tied solar energy system will generate power or not, independent from and irrespective of how much or when the residence is drawing power. Software and spreadsheets are commonly available to estimate the solar system's power output on an hourly basis (or smaller time increments) for periods of one year. Usage is not needed to determine this # (PVWatts with hourly output option seems the easiest, but least accurate, SAM, others). The key to this (and the PITA) is loading 8,760 hourly tariff $$ values for 1 kWh of electricity onto a PVWatts spreadsheet, but not as bad as it sounds - more like loading 168 hourly T.O.U. tariff values (for 7 days) and copying 52 times plus 1 day, then adjusting for holidays, at least as far as SDG & E is concerned.

Then, multiply the hourly system output by the hourly T.O.U. rate, sum 8,760 of those and voila - yearly system revenue - again, provided the estimate of system output is less than 100% of the annual load.

Do that for several or as many array orientations as you please and you'll likely zero in on an optimum (max. annual) revenue. The best system orientation for max. revenue is independent of customer use or use pattern. Provided that the system produces less than 100% of the household use, a system that produces, say 9,500 kWh/yr. in some particular orientation and, because of the T.O.U. tariff in effect produces, say, $2,500 in income, will do so regardless of the household use or time/pattern of that use given TMY or a "Solaranywhere" year data to estimate output.

The annual residual bill is the difference between revenue and expense. At least for SDG & E T.O.U. customers who are not "excess generators". Those folks currently get about $.05/kWh or so for the excess on a yearly basis.

I've done something similar to the above for my system. It's sort of an academic exercise since my and most other systems are fixed. However, it does seem to shoot down the idea that west facing arrays (270 az.) are more cost effective.

In my case, I took a 20 deg. tilt (common) and varied it in 10 deg. az. increments. I used some spreadsheet stuff I did years ago, before SAM, which produces output very similar to SAM's, but easier to work with, and quickly zeroed in on an op. revenue az. for DR-SES tariff ( T.O.U. for users with solar energy systems) of about 205 deg. That seems similar to data Sensij provided in a prior post. I'd not be surprised if 20 to 30 deg. west of south az. at most commonly used tilts produces the most income for the greater part of CA under most T.O.U. schemes. That's most of the reason why I'm of the opinion that the Times article is bogus road apples.

More on my data and revenue as f(az.), and gory details in a later post if anyone is interested. As I wrote, it appears similar to what Sensij found.

For customer electrical use and annual bill - the use is often available in hourly or smaller increments from the POCO - that's one benefit of smart meters. Do the same for electrical cost as done for revenue = hourly use times hourly per kWh cost, sum 8,760 times and you have a yearly bill.

A nice little bennie to all this: To zero out a bill, or to any other amount and/or to optimize system size as you choose, depending on how you do the process economics, the procedure usually becomes the following: setting the tilt and az. to the roof where the system will be located, figuring out how much revenue $$ 1 kW of nameplate capacity will produce as described above, then divide your bill by that 1kW revenue $$ amount. The result will be the required nameplate capacity in kW.

SWAG, Q.E.D.

[subdriver97]

Look at the data provided by sensij. Even though generating more of your own electricity during the higher rates in the afternoon sounds like a good idea it may not really be. In cases where there is no shading you will generate more power in a 24hour period with the panels facing South then in any other orientation. It is not when you generate, it comes down to how much you can generate and South gets you more than West or even SW.

When considering TOU, it very much does matter when you generate and not just how much you generate. We all know that South facing is ideal to maximize how much you generate. The original question on this thread was "what is the optimal azimuth" when considering TOU. I agree that it doesn't make much sense to build a special rack to face 225. However, a more plausible situation it a home that has roofing surfaces facing in SE and SW directions. In that case it would make more sense to choose SE over SW. In my personal situation my roof faces 125 and 305... so it was an interesting question to consider if 305 might actually be better for a TOU rate plan, turns out it doesn't because too much production is lost going that far away from south.

[SunEagle]

Are you familiar with Time Of Use(TOU) metering? In California it is all about WHEN you generate. Peak prices are 4x higher than off peak, with net metering you are banking $ not kWh. From the spreadsheet looks like 225 is ideal and south vs. west is a wash.

I thought that the spreadsheet that sensij showed was based on TOU metering.

He also pointed out that the difference between the best and worst was just $100 per year per 3kw. I don't see that savings as a motivator to add special racking to point the panels more toward 225 then S or W.

[prhamilton]

It is not when you generate, it comes down to how much you can generate and South gets you more than West or even SW.

Are you familiar with Time Of Use(TOU) metering? In California it is all about WHEN you generate. Peak prices are 4x higher than off peak, with net metering you are banking $ not kWh. From the spreadsheet looks like 225 is ideal and south vs. west is a wash.

[SunEagle]

I started this thread because i read this article.

http://www.nytimes.com/2014/12/02/up...abt=0002&abg=0

If you look deeper into that article you may find that the Electric Utilities may have had some input into the data collected and presented.

First off most POCO's want to make a profit. If they can convince their customers to do something that puts more money into the POCO's pockets they will make it sound like it's the best thing since slice bread.

Look at the data provided by sensij. Even though generating more of your own electricity during the higher rates in the afternoon sounds like a good idea it may not really be. In cases where there is no shading you will generate more power in a 24hour period with the panels facing South then in any other orientation. It is not when you generate, it comes down to how much you can generate and South gets you more than West or even SW.

[araghava]

You can see in my post above that if you are comparing due south vs due west for any particular tilt, south wins every time under the terms of the EV-A plan discussed. The improved $ / kWh you get for west is more than offset by the increase in total kWh by facing south.

This will not be true for all installations... local shading or other factors could change the result. However, in general, due south is still better than due west, even for TOU.

I started this thread because i read this article.

http://www.nytimes.com/2014/12/02/upshot/why-more-solar-panels-should-be-pointing-west-not-south.html?module=Search&mabReward=relbias%3Ar%2C%7B%222%22%3A%22RI%3A15%22%7D&abt=0002&abg=0

[sensij]

So I think your choices are really only how many panels facing south vs how many facing west. I don't think most folks on this forum live in a region where they can appreciate paying 40 cents a KW, or 4X the price of off-peak power. I think south vs west will make only a minor difference in energy production, but it will make a huge impact in dollar production. Especially with "peak" stretching to 9 pm, if you get home between 6-9 pm from work, offsetting that evening energy use with more solar production between 2 and 6 pm will make a big difference. It doesn't matter if you use more than you're producing, any amount of peak you can offset is huge.

You can see in my post above that if you are comparing due south vs due west for any particular tilt, south wins every time under the terms of the EV-A plan discussed. The improved $ / kWh you get for west is more than offset by the increase in total kWh by facing south.

This will not be true for all installations... local shading or other factors could change the result. However, in general, due south is still better than due west, even for TOU.

[wwu123]

So I think your choices are really only how many panels facing south vs how many facing west. I don't think most folks on this forum live in a region where they can appreciate paying 40 cents a KW, or 4X the price of off-peak power. I think south vs west will make only a minor difference in energy production, but it will make a huge impact in dollar production. Especially with "peak" stretching to 9 pm, if you get home between 6-9 pm from work, offsetting that evening energy use with more solar production between 2 and 6 pm will make a big difference. It doesn't matter if you use more than you're producing, any amount of peak you can offset is huge.

[wwu123]

Why can't the panels be mounted at an angle. i.e not aligned with the roof.

It's not just the mounting cost and complexity for additional rail and steel to orient them differently than flat against the roof. Yes, from a geometry standpoint you can put panels on any roof and point them in any direction with enough support. Say if you first laid them all side-by-side against the south roof first as in a typical install, then tilt each one to face southwest, you'd find that during your optimal hours, they would be shadowing each other. Even shadowing a tiny corner of a panel nearly kills all the panel output.

So you'd have to space them farther apart to avoid shadowing each other, esp during your peak production times. And there I think most residential roofs wouldn't have enough area for a decent-sized array with the spacing needed.

[control4userguy]

Charge before 2 or after 9

Read the response he gave me. He is not interested in charging during peak.

[russ]

1) Do you have grid tie?

2) East facing would produce slightly more possibly - depending on location and micro climate

[cf400]

Hi,

I'm going to have panels installed next week. I have an electric car and am on the EV-A plan. Peak rates are from 2:00 pm to 9:00 pm. Has anyone tried to optimize for this by orientating their panels somewhat west of south. If so could you share your orientation and tilt angles.

Thanks.

Ananth

Charge before 2 or after 9

[sensij]

I have the raw data but don't know how to sum up the different time blocks. i.e. add up all the blocks for 7am to 2pm, 7pm to 9pm (partial peak) and 2pm to 7pm (peak).
However i can see that production for peak period increases as you orient westward while partial comes down.

This is an academic exercise more than anything, because I don't really know how you would design your roof mounting to take advantage of the results. I had created a spreadsheet a few weeks ago to evaluate a couple specific orientations on TOU, and modified it to take data from a SAM parametric simulation of tilts and azimuths. Better software could do more, but this is what I could get from excel:

PGE-EVA.PNG

The rows are azimuths, the columns are tilts, and the values in the table represent the annual dollar value of the energy produced by a 3 kW array, using the San Francisco airport TMY3 weather file. The EV-A rate plan used was this one.

Since no particular rate plan is ever guaranteed, I would be cautious about optimizing too tightly on these assumptions. However, as a thought experiment, it is interesting to me, and since this TOU plan has a definition of peak that could be reasonably expected to continue in the future, it has a chance of being useful.

Note that the difference between the absolute worst of the modeled conditions and the absolute best is less than $100 annually per 3 kW of array. If the custom tilt and azimuth mounting is expensive, I think it would be hard to justify.

[subdriver97]

I have the raw data but don't know how to sum up the different time blocks. i.e. add up all the blocks for 7am to 2pm, 7pm to 9pm (partial peak) and 2pm to 7pm (peak).
However i can see that production for peak period increases as you orient westward while partial comes down.

Just sum up the power generation numbers for each time block and multiple by the associated rate for the time block... then sum those products. Do this for each azimuth and choose the azimuth that is the greatest.