For off-grid solar, more panels are better than more batteries

TLDR version:  adding solar panels can drastically reduce the required battery bank size for off-grid solar.  
Adding just two more panels to my 15 panel array could reduce the size battery bank I would need to get me through current monsoon weather from 60kwh to 20kwh.

Long version:
My power coop has proposed a new pricing schedule for solar customers that they are planning on phasing in over the next 4-5 years.  If this goes through it will be cheaper for me to buy batteries and go off-grid than it will be to stay grid-tied.

With this in mind I have been collecting production and consumption data since the beginning of May.  I've been collecting data for years, but previously I wasn't concerned with whether the power I consumed came directly from my solar production or from the grid.  Up until now my power coop allowed me to 'bank' my surplus energy for later use at a 1:1 ratio.

Anyway back in may I started tracking where my power was coming from on a continuous basis.  If I go off grid, anything that I would have pulled form the grid would have to come from batteries and I'll need to know how large of a battery bank I'll need.

One way to estimate battery bank size is to take your average daily consumption and size the bank large enough to handle several days without any solar input.  So if you live somewhere that frequently experiences storms that last 2-3 days then maybe you make sure your bank can power your house for 4 days.

However, since I have hard data available and several years for planning I'm taking the data and feeding it into a spreadsheet.  The spreadsheet takes the energy I'm currently sending out to the grid and adjusts it for estimated charging losses, battery capacity, etc.  With this I can see what size battery bank I'll need taking into account actual usage and production based on current weather patterns.  

My solar production/energy consumption for the last month:

Table showing production, consumption and percentage supplied by the grid

A couple points worth noting:
  While it is convenient to assume a worst case of zero input from the solar array, this doesn't actually happen (short of hardware failure).  In the 7 years my array has been online the lowest production I've ever seen was still approx 10% of maximum production.
Energy use during stormy weather is often lower than during clear days, at least it is where I live.  Most of our cloudy weather happens during the summer and cooling requirements go way down on cloudy days.  Interestingly, during the winter cloudy days tend to be warmer than clear days/nights.  The clouds act like a blanket to trap heat, whereas clear nights and low humidity can result if very cold nights as the heat radiates away.
Lead-Acid batteries last longer if they are kept mostly charged most of the time.  A smaller bank with a larger array will tend to run at a higher State of Charge(SOC) than a larger bank with a smaller array.  In my case a 20kwh battery with a 4kw array would only have been below 50% SOC on three days during the past month and never goes below 20% soc.  A 40kwh battery with a 3.5kwh array would have spent 8 days below 50% SOC and 1 day below 20% and one day completely discharged leaving me 3kwh short on energy.
Panels are cheaper than batteries.  2 panels are a LOT cheaper than 20kwh worth of batteries, plus the panels will last over 30 years, the batteries maybe 6-8 years.

This looks interesting Peter, especially as the market for both solar and traditional sources of electricity continues to change.  Could you possibly elaborate on what you mean with "My power coop has proposed a new pricing schedule for solar customers that they are planning on phasing in over the next 4-5 years."  I've been trying to keep up with development with our own power coop and have been toying with grid-intertie, but also would like to hear what you've been experiencing with your coop.  Thanks.  Also, if anyone with a more northerly location (or southerly if in that hemisphere) has a view on this as well that would be an interesting comparison with respect to seasonal differences of solar incidence.

John Weiland wrote: Could you possibly elaborate on what you mean with "My power coop has proposed a new pricing schedule for solar customers that they are planning on phasing in over the next 4-5 years."

Sure.  Up until now solar customers in my area have been getting a free ride.  Basically, with the current net-metering agreement they keep track of how much energy I push to the grid and how much I pull.  Once a year (September in my case) they 'settle' up, if I've pushed more to the grid than I've pulled they pay me (wholesale price, about 2.6 cents per kwh) for the surplus, if I've pulled more than I push they bill me at the retail price (about 12.6 cents per kwh).
Net-metering (which requires a special meter) costs me an additional $2.70 a month.

Because my array produces much more than I use, I haven't paid an electric bill since it went online 7 years ago.  This is pretty common for solar customers in my area.  
This issue is that the grid has certain 'fixed' costs that the coop has to recover no matter how much power they sell.  The have to pay salaries, maintenance and repair costs, etc.   The number they came up with was $75 per customer per month.
When I'm not paying anything for my connection these costs have to be paid by the other customers. Back when there were only a few solar customers this wasn't a big deal, but now over 2% of the houses in this area have solar.  This means that non-solar customers are paying $5-$10 extra a month right now and this would increase as more people installed solar.

While I think it's not fair that solar customers aren't paying anything, I have an issue with how the coop is calculating their 'fixed costs'. For one thing the coop doesn't have any power plants, so they buy the energy wholesale.  Currently they buy a minimum fixed amount, plus extra energy as needed. They still have to pay for the minimum amount even if they don't use it all.  So they are adding that minimum amount into the 'fixed' costs.  Even with all of their solar customers they have never used less than their minimum, so it doesn't seem fair to me to charge solar customers for energy they are already selling anyway.

I've done the calculations and under the new pricing schedule my monthly bill (remember I produce more power annually than I use) will be about $5 less than if I didn't have any solar array.  Basically I'll be paying about $65 a month just for the privilege of being connected to the grid.
If I disconnect from the grid, maintenance and replacement costs for batteries will be about $25 a month.

Yeah, thanks for this, Peter.   This is going to have to be something to keep an eye on when budgeting for the future.

"Shaffer said the rural cooperatives had much lower fixed charges prior to the legislation. “What has changed?” he asked. “The result of high fixed fees will be to cripple solar in these areas.”

“The rates we’re seeing charged by Minnesota co-ops are the highest in the country,” said Allen Gleckner, a senior policy associate at Fresh Energy, where Midwest Energy News is published. “In Arizona the charge is typically $15 a month, and they have a much greater penetration of net-metered systems than any Minnesota cooperative.”

As an example, Gleckner showed a January bill from a solar customer of Meeker Cooperative in Litchfield that shows a $55 net metering charge on a just under 40 kilowatt (kW) system. Over a year that would cost the customer more than $600, the equivalent of several months worth of electric bills.

In contrast, the state’s largest investor-owned utility, Xcel Energy, has a fixed fee for all customers of $10 a month, but nothing specifically for solar. Minnesota Power has a monthly “co-generation fee” of around $2.55 a month for a 20 kW system."  --  http://midwestenergynews.com/2016/03/29/minnesota-co-ops-rolling-out-high-fixed-charge-for-solar-customers/

Edited to add with respect to a demonstration array at the cooperatives HQ in northern Minnesota:  "Capacity factors (amount of time the array is producing at maximum output over a period of one year) have typically been between 13-16 percent."   How might that compare to an array in southern Arizona or one in Churchill, Manitoba?

John Weiland wrote:"Capacity factors (amount of time the array is producing at maximum output over a period of one year) have typically been between 13-16 percent."   How might that compare to an array in southern Arizona or one in Churchill, Manitoba?

"Capacity factor" is calculated based on the assumption of maximum output 24/7.  Because solar panels don't produce anything at night, even under ideal conditions with a 2 axis tracker the maximum theoretical capacity factor for a solar array is less than 50%, real world values will always be less than ideal.

In Arizona, fixed solar arrays often have a capacity factor over 20%, probably average around 23%.  My array is on a 1 axis tracker with seasonal tilt adjustment, I typically see a capacity factor of around 28-29%.  Two axis trackers probably hit somewhere around 35%.