Selecting the proper tilt for your panel is one of those subjects which
appears simple on the surface, but as you dig into it, you discover it's
not simple at all. A rough answer is easy to supply, and it might be
the best choice for a fixed rack. Tilt the panel an amount equal to
your latitude and face it towards the equator. For residents of the
northern hemisphere, this can be accomplished without any instruments or
charts by orienting the panel at night, when you can see Polaris, the
star near the north pole. Sight down one edge of your panel, pointing
it stright at Polaris. This is a reasonable first approximation,
appropriate for fixed racks in sunny climates, but it will produce less
power during both summer and winter, and it's not good for cloudy
climates.
The second approximation is to tilt your panel according to season,
setting it to the solar declination at noon . This will optimize
your output at noon on sunny days, and produce far more power in winter
than a fixed rack. But it ignores the path of the sun through the sky.
In summer especially, it's the wrong answer, because in summer the sun
rises away from the equator, north of east in the northern hemisphere
and south of east in the southern. Even if there were no atmosphere and
no scattering of light by clouds, the optimal panel tilt in high summer
isn't the solar declination, it is a perfectly flat panel, a tilt of
zero .
On overcast days, flat panels preform best, because on cloudy days
illumination comes from any point above the horizon. In very cloudy
climates, the optimal panel tilt is again zero, because it exposes the
panel to the maximum view of sky, where light can be scattered onto it
by clouds. But what about most climates? Climates that have cloudy
days interspersed with sunny days have an optimal panel tilt that is a
compromise between zero degrees and the optimal angle for clear sky,
with some weighting related to how much power can be generated on both
kinds of days. If half your days are cloudy, but they offer at most
1/4 the illumination of clear days, and your goal is to produce as much
electricity as possible, the right answer is pretty close to the clear
sky answer.
Yet another issue strongly influencing the optimal panel tilt is the
albedo of the ground or reflector in front of the panel. Especially in
winter, it is very beneficial to use a reflector, and that reflector
exposes the panel to a larger amount of sky, increasing the amount of
power that can be generated on cloudy days. An effective reflector
ought to make the optimal panel tilt greater than it would otherwise be.
It's quite difficult to determine how much of an effect this is,
especially for a home-built reflector like the ones you and I can
reasonably afford. Although a lot of research has been carried out on
solar panels and there are many measurements of horizontal solar flux,
it is quite difficult to determine how much power one can reasonably
expect to produce with a given panel in a given location.
About Tilt Factor
One of the goals of my solar-assisted UPS project is to measure a
unitless quantity I've called the "tilt factor", T. The tilt factor can
be used to convert an observation of daily average horizontal flux, F,
in watts per square meter at a weather station to daily charging C in
amp-hours producted by a tilted PV panel of calibrated aperage A nearby.
C/A = T *F*(24 hours)/(1000W/m2)
With knowledge of the tilt factor, and how it depends on season and
weather, one could use solar flux measurements to predict with
reasonable accuracy how productive any solar power project could be.
In order to measure a tilt factor, one cannot install a solar charge
controller. A solar charge controller throttles back the output of the
panel, to avoid over-charging the battery. Charge controllers
essentially discard some of the solar energy, they make solar systems
less efficient, but they also make them safer. As an alternative to a
solar charge controller, one could always load the solar system heavily
enough that the battery cannot charge fully, one could employ a
diversion load, or one could switch the panel on when measuring its
current output, and turn it off when the voltage exceeds a threshold.
My original plan was to disconnect the panel when the batteries were
overcharged, but turn it on every 10 minutes for measurements. I
couldn't actually do what I had planned, because my panel puts out much
more than its rated capacity of 4.5A. This is due to two effects - the
panel is about 5% more powerful than advertised, it produces about 4.8A
in standard sunlight, not 4.57A. I calibrated my panel by installing it
in a horizontal position for a few days with no reflector, so the tilt
factor was 1. Also, the sunlight in Boulder is frequently brighter than
the 1000W/m^2 standard. I fairly commonly see 6A out of my panel, and
that's too large for the 5A Phidgets relays that I purchased. So
instead I implimented a diversion load, and I consume all power
generated each day with my normal load each night. My interests in
measuring tilt factor and in producing as much power as possible for
each installed watt of solar generating capacity have lead me to size
the solar-assisted UPS differently than one would size an off-grid solar
power system. It should almost always provide less power than the load
consumes, even on the brightest summer day, so no power is discarded.