This is the first in a multi-part series documenting what we hope will be the successful design, planning, installation, and evaluation of a residential solar photovoltaic array. There are two main types of solar power installations: battery systems and grid-tie systems. I chose the latter since we will not have to purchase or maintain a large number of batteries. Instead, the electrical grid and our local utility will act as our battery. This also gives me the ability to produce only a fraction of our power while purchasing what we need in excess of our own generation capacity in the usual way. Our choice does not affect much of what you’ll see and learn in this series, however.
I’ve been watching the performance of photovoltaic (PV) panels improve and their price drop for a few years now. Commercially available PV panels with efficiencies of 15% are on the market at reasonable cost. Recent developments in power inverter technology allow for an economical and efficient inverter to be coupled directly to each individual panel. I’ll address each of these shortly. The recent federal rebates covering 30% of the cost finally pushed us over the edge and I will be placing a modular, expandable PV system on our home. This first installment will cover how to set realistic goals for system performance, how to assess your site, and on-line tools to help with some aspects of the system design.
How much power do I need?
We originally had an electric water heater, oil furnaces and average efficiency air conditioning in our home. I began charting our energy usage month by month beginning in 2007. That year we used 22,000 kWhr of electricity. Over the next two years we replaced the electric water heater and oil furnaces with high efficiency natural gas units. We also replaced the Air Conditioning units with more efficient 14 SEER ones. Our incandescent lamps were upgraded to either compact fluorescent or LED lamps. Doing those upgrades reduced our annual electricity usage to 15,800 kWhr or by nearly 30%. I suspect there are other things we can do to lower our usage, like getting rid of the old refrigerator in the garage. The point I’m trying to make here is that you need to estimate how much power you use in a typical year and that there are a number of things you can do to reduce your usage. Installing solar power is not cheap and you need to do some homework before you decide to spend money on a new system. Our power usage has been pretty stable since the upgrades and I feel that I have a good estimate for what we typically use in a year.
The state we live in, Virginia, has fairly good regulation of the utilities and has put in place reasonable laws governing how utilities have to work with residential customers who wish to generate their own power. This will certainly vary from state to state. In VA, we cannot generate power in excess of 10kW of what we use. This means we cannot push more than 10kW of power back onto the grid. Practically speaking, this means we probably should not plan to generate more than 10kW of peak power. We are allowed to generate more than we use in any given month with the excess being carried over from month to month. We are never allowed to carry over more than we use in a year. Anything beyond that we provide to the utility for free. You will need to read up on the regulations applicable to your specific location as this information will be used later. Given the fact that fluctuations in solar energy due to the weather can cause year to year changes of 10% or sometimes 20% in certain locations, I wanted to limit our solar PV systems size so that we produced no more than 80% of what we expect to use in a typical year. For our system, that meant we want to produce about 12,500 kWhr per year in an average year.
Where to place the array
The next thing you need to determine is just where are you going to put the solar panels. If you have a heavily wooded lot, you will be severely limited as to where you can locate the solar panels and how big your array can be. We are fortunate to have a large open lot that gets very little shade, so finding a place to put a large number of solar panels should not be a problem. What I will call ‘full size panels’ are typically ~36″ wide and ~64″ tall. These produce 180 to 230 Watts of power in full sun, depending upon make and model of the PV panel. There are a lot of different ways to mount solar panels. The two most common are roof mount and pole mount. Which method you choose will be dictated by your site, structures, and aesthetics. We want to mount ours on a roof, but none of our roof slopes have a good southern exposure. Our best sections of roof are oriented with their exposure to the south west. I thought this would be huge problem because common sense indicates that you want your solar panels to face south. I even went so far as to design a “solar shed” that I could place on our property with the proper southerly orientation. Talking with a number of local folks who have solar panels and doing some simple searches on the web for information on how much energy you can get from the sun, it appeared that our area has an average of 4.6 hours of sun per day on average over the course of a year. Doing the math (4.6 hours/day x 200W/panel x 365 days/year = 336kWhr/panel/year) told me that I’d need at least 37 optimally placed panels to produce the 12.5MWhr/year that I wanted. That is roughly 600 square feet of solar panels or a 20′ by 30′ rectangle. After spending some time with a tape measure in my backyard, I concluded that I had the room for an array of that size. With those estimates, I wanted to build a new “solar shed” and put the solar panels its roof with the building oriented perfectly to the south for maximum illumination.
Expected performance – PVWatts
Once I got an estimate on what a new building would cost I began to think about how many more solar panels we could buy instead. It turned out to be quite a few and I wondered if we had other more cost effective options. Our house is a rancher with a lot of roof area. The problem is that the house has the corners aligned with the cardinal directions N, S, E, & W, so the slope of the roof is mostly aligned towards the SW direction. I was going to write a program to calculate how the annual power output varied when you changed the tilt of the panel (slope of the roof it is mounted to) or the orientation of the roof with respect to south, but I thought that someone had certainly already done the calculation. Indeed they had. I found a couple of journal articles which concluded that the orientation relative to true south was not too important. After thinking a bit about the ecliptic, the length of the day as function of season, and how the power output varied with the cosine of the angle between the position of the sun and the vector normal the plane of the solar panel, I could see why this might be true. A little more searching led me to a really neat online tool called PVWatts that calculates the power output of a solar panel. PVWatts is extremely useful and pretty easy to use. By clicking on a map of the US (a version covering Europe also exists) and filling in some values on how big your solar array is and how it is oriented, PVWatts will calculate the average month by month output an total annual energy production at your location. It even includes the average local weather for cloud cover and the variations in solar cell efficiency as a function of temperature. This is exactly what I needed to size my solar array and evaluate the several possible installation locations. You can save a lot of time by skipping the map clicking if you save the URL of the form that follows.
I ran a series of simulations for my location. The first was to see the effect of varying the tilt of the solar panels. From my reading on the subject, the consensus was that you wanted to tilt the solar panel to match the latitude of your location, in my case 37 degrees. Other reading indicated that a tilt of less than the latitude would produce the maximum annual output. My roof has a 10/12 slope or a tilt of 40 degrees. I ran PVWatts several times changing only the tilt of the array from 0 (flat on the ground) to 90 degrees (standing vertically as though mounted to a wall). The first figure shows the relative monthly energy and annual energy output relative to a solar panel tilted to match the latitude. What you can see in the figure (click the figure to see the full sized version) is that as the panel tilt is reduced, the power output increases in the summer months and decreases in the winter months. The opposite it true as the tilt increases. At my latitude, the sun goes almost directly overhead in summer and the power output of a panel tilted to 90 degree drops close to zero. The last point on the graph is the total annual power generated. It shows that over the course of a year, the total energy output varies only slightly for tilts within 10 degrees of the local latitude. The tilt alters the amplitude of the seasonal variation. You can use this to your advantage if you wish to change the amount of power generated in summer vs winter. Based on my power use history, we use 2.5 times more power in the summer than we do in winter.
My second set of simulations explored what happens when you do not orient the panels to face due south. The results of this came as a complete surprise. Summer power production changed very little. The reason for this is, during the summer, the sun travels over more than 180 degrees of azimuth. It rises north of east and sets north of west. The solar panel can only see 180 degrees of sky, so there is a range of orientations where the solar array will receive the maximum illumination. What happens is that panels oriented slightly to the west produce their power later in the afternoon than those oriented due south. At my location, orienting them to the west of south actually produces more power presumably due to a diurnal asymmetry in cloud cover and temperature. My house roof is oriented almost exactly to the SW (221 degrees, green line in the figure). With this orientation, the power output is reduced by only 20% in winter, but is still more than sufficient to supply almost all of our electrical needs. The real important result is that putting the solar array on my existing house roof reduces the total annual energy generated by only 7%! That means we have to add only another 2.5 solar panels to make up for the difference (40 instead of 37). The extra cost is less than $3000 as compared to 10 times that for the cost of build a new “solar shed”. This was a huge revelation. PVWatts also includes some real world parameters for efficiency of the inverter, losses in the power wiring, and light loss due to dirty solar panels. Based on the results of the PVWatts calculations, we are going to need to size our solar PV array at 9.2kW if we want to produce 80% of our electrical needs. PVWatts will also calculate the annual savings based on your cost of electricity.
Panels, inverters, & mounts
PV panels come in a wide variety of sizes, voltages, and total power. Depending upon which direction you plan to orient the panels, the panel length is a critical dimension to consider when deciding which panel to buy. I plan to have one small array of 12 panels mounted on the garage and another 30 mounted on the house. The total is 42 panels. I put together a little spread sheet to help evaluate the panels and how many I could fit into the available space. On the detached garage I can mount the panels upright (Tall side vertical) and have the mounting rails run horizontally. With that orientation I could fit twelve 230W panels in 2 rows of 6. Over on the house, I hit a little snag. Many panels are just a little over 64″ long, not including the required spacing between panels for the mounting hardware. Standard roof framing has the rafters on 16″ centers. The mounting rails are to be placed at the 25% and 75% positions along the length and have to align with the rafter (or you have to add stringers between them) for the lag bolts. Do the math and you’ll soon realize that these long panels can only be mounted vertically unless you can tolerate having a 15″ gap between the columns of panels. The available area on my house’s roof dictates that my panels have to be oriented horizontally into 6 rows of 5. Otherwise, if I mount them vertically I cannot get 30 panels on that roof, only two rows of 12 panels. You would think that the US manufacturers would have figured this out and made all of their panels shorter than 64″ long. But no, the Asian manufacturers are the ones that make them the proper size. For the garage, I decided to go with 230W panels made by Solon. These are made in the US. On the house, I’ll have 215W panels made by Sanyo. Total rated power is 9210 watts.
Inverters are what turn the DC power from the solar panels into 240V AC power that you need to power your house. Until recently, a typical installation had one or two of these in the system. You would daisy chain you PV panels serially to produce a rather high DC voltage. 600V is not unusual. There is nothing wrong with this, however in this arrangement, one under-performing panel (perhaps partially shaded) limits the current output of the entire string of panels. The ideal way to convert the DC to AC would be to have an inverter for each panel. This is now a viable option with the new micro inverters on the market. Enphase makes a micro inverter that can be tied directly to the home’s AC power system. This saves money since the system does not require a separate Automatic Transfer Switch to isolate the solar array from the grid when the power goes out. These Enphase micro inverters turn off when the grid voltage is interrupted. The micro inverters can also be daisy chained to simplify the system wiring.
You can find information on all of the hardware mentioned here in Part 1 over at Wholesale Solar. Next time, we’ll get into the process of producing a system plan, obtaining the permits, and getting some of the preparatory work done.