What can we learn from a friend's solar home data in Vermont?

[anorlunda]

TL;DR Summary

Data from a Residential Solar PV Project.

I visited a friend who has a very nice solar installation. He also has the software to do data collection and presentation. I thought it would be nice to share some of his data here. Perhaps we can link to this post as a reference in future solar discussions. First, some background.

• The whole installation was done on a power purchase agreement. That means at third party company provided all the up-front investment, engineering, construction, installation, and they monitor performance. I do not know any more details about the financial arrangement.
• Part of the deal was that my friend was required to get his house audited by Energy Vermont, and to carry out any improvements they recommend at his cost. The purpose being to minimize electric energy consumption and thus maximize energy sales to the grid.
• Green Mountain Power, the local utility, also participated. They provided two Tesla Power Wall battery units free to the homeowner. The power walls are controlled by GMP. They charge/discharge as GMP's discretion, to suit GMP's goals. If I were GMP, I would connect those batteries on the utility side of the meter (but they are probably not technically able to do that.)
• The location is at latitude 45 degrees. The climate in Vermont is cloudy much of the time, it is temperate, not desert. Rain plus lack of dust makes the panels self-cleaning. Homes need lots of heat in winter and almost zero AC in summer.
• The installation includes the inverter, utility interface/transfer, GPS, and it uses the home's WIFI. The solar and Tesla units coordinate with each other and communicate with GMP central dispatch, but the homeowner had no info on such coordination.

The first picture shows the physical installation. It is a stalk with about 900 $ft^2 or 100 m^2$ of panels and two-axis motion for sun tracking. I was impressed by the good engineering and structural strength of everything. Sensors detect high wind and feather the panels when needed by making them horizontal. They also feather every evening, and face the opposite direction in the morning. Note the hydraulic actuators for positioning. (This photo was taken near sunset.) Solar position in the sky is done by online lookup using the GPS position. I believe that all data communications to the solar company and the utility went through the WIFI.

The second picture shows 24 hours of history for August 2. The red line shows energy consumption. The green shade shows energy sold to the grid. The red shade shows energy bought from the grid. Note that the green shaded area is not sinusoidal in shape. It approaches a square half cycle. I attribute that to the 2-axis motion system. The difference between the square shape and a sine shape with the same peak is a visual indication of the energy benefit of the 2-axis system.

The third picture shows one year of history. Same colors. In winter, this home's energy consumption increases, while lack of sunlight cuts solar production. The home has a heat pump. AC in the summer is almost never used. That pattern is reflective of the local climate. I like it that the peaks and valleys in the data clearly show the weather pattern of cold fronts passing across the continent about 5 days apart.

Also in pictures 2 and 3 is a green dotted line (hard to see). I believe that is power generated by the two Tesla Powerwall units. They appear to be very underutilized. If they were owned and operated by the homeowner, I believe they would have played a bigger role. There is no way to tell how much skill and attention the utility invested in operating those batteries.

Here are some takeaway lessons.

1. Finance and design: This system is too much for a DIY project, and too much for a contractor. It makes sense to let 3rd party specialty companies provide the finance, engineering, and operation, even for residential installations. The 3rd party can negotiate a deal with the power company aggregating all their solar installations in the whole state.
2. Depending on local latitude and local climate, the details of solar PV installations vary dramatically. In Southern California, rooftop fixed-position installations dominate. In Vermont, relatively little PV is rooftop. Residences tend to use motion controlled stalks like this one. Municipalities, tend to use acres of former hay-fields for a solar farm, but fixed position panels. In desert locations, there will be an added labor cost to keep the panels free of dust. In some northern locations, there will be labor to clear them of snow (or to consider solar a seasonal resource and forget the winter.)
3. The benefits to the homeowner strongly depend on local climate and latitude.
4. We have four parties involved here. The homeowner. The solar company. The power utility. The government agency Energy Vermont. Their self interests may align in most ways, but not all ways.
5. You can never be independent of politics. Net metering, and subsidies should not be considered sustainable.
6. The majority of US people who live in high density housing are unable to have distributed generation projects like this one.

I invite other PF members who have access to similar data to post it in this thread. It would be instructive to see the differences.

 

[phyzguy]

I have a much more modest solar array. I live in Northern California, and have a rooftop array with 16 300W panels, so the whole array is about 25 m^2. I have a nearly flat roof, so the panels lie flat on the two sides of the roof, which face East and West. A couple of the panels get shaded in the late afternoon, so I have micro-inverters, meaning that each panel has its own inverter. Averaged over the year, the array generates more power than I use, although I generate way more than I use in the summer, and less than I use in the winter. The local utility (PG&E) has net metering, so I basically pay nothing for electricity, although I do pay a small fee (about $10/month) to pay for the hookup to the grid and the fact that the utility guarantees me power no matter whether my array is generating or not.

Below are some photos. First is a picture of the array, then the peak production in June, then the production so far this year. The next graphs are the net production (daily production minus use). In June I generate more than I use, in March I about break even, and in January I use more than I generate. The last graph is the annual net use. All in all, I'm very happy with the system.