Solar Street Lighting

Re: Solar Street Lighting

As a start, I would suggest looking at CandlePowerForums and their LED section.

They have lots of discussions about LED, electronic ballasts, and even companies that market built and kit ballasts for LEDs (CFP is a "flash light" centric forum--So these are going to be low power/high efficiency devices--at least last I looked a few years ago).

Don't know if you are looking for a few DIY type solutions, or want to look at mass production--But both are doable.

White LEDs typically need around 3-6 volts to operate, and they need a "ballast" to limit current to specified level (LEDs have a very non-linear current knee where slight changes in voltage and temperature can cause dramatic swings in current).

A ballast can run anywhere from a simple resistor (3-6 volt drop one or two LEDs, balance of drop across power resistor) to some sort of buck (voltage dropping) or buck/boost (low, middle or higher battery voltage) DC switching converter setup in current regulation mode. I don't see any reason for going with AC inverter based system here unless you have a central "power hub" connecting to a string of street lamps.

This is done all the time with (better) LED flash lights... You can put in the ballast; automatic lighting control, battery voltage limits, even adjustable current modes (vary light according to needs).

I would agree that 25 watts is "old school" for lighting. Using 25-100 Watt metal halide or high pressure sodium lights (white or orange) very efficient lightning and replacing with equivalent amount of Lumens of lighting from LED.

I would suggest getting some modern Luxeon type LEDs and experiment with reflectors and various current (power levels) and see what is needed.

The more you can focus light to where it is needed (older street light used to put 1/2 their energy into the sky), the less power you will need.

Also, the human eye observes light "logarithmically"...

In day to day work, we are used to measuring something as 3' or 6' long and saying that is a big difference.

In lighting (and sound, and many engineering fields) we use logarithmic functions to describe/measure things because that is the way to manage "large differences" in the math (a radio transmitter outputs watts to 100,000's of watts, the receiver receives millionth's of a watt).

With lighting, it is similar. Our eyes can barely perceive a 2:1 difference in light levels (i.e., a street light running at 25 watts will look virtually identical to one running 12.5 watts). It is when the power levels approach 10:1 or greater that we see a "day/night" difference.

So--You should be able to reduce your LED output by 1/2 or even to 1/4 and still have "reasonably" equal lighting. Or, if dimmer lighting is OK, reduce to 1/10th the power level--The savings of battery/solar panel energy is huge.

Another idea is to add a motion detector to the lighting. One company in Europe (at least) has been advertising such street lamps... Run at 1/4 power or less until motion detector trips from person walking, then ramp up light to full power (ramping is probably more acceptable than 10% to 100% switch to pedestrians and people in homes nearby). And back down when the person walks out of range.for

Regarding 70 watt panel and 7 hours of sun--Probably closer to:

70 watts * 0.77 panel+controller derating * 0.80 battery efficiency * 0.85 inverter efficiency * 5 hours average sun = 183 Watt*Hours per day for "average" 9+ months of year
183 watts / 25 watt lamp = 7.3 hours per night

So, I would guess that your setup is right on the raged edge... If the AC inverter is inefficient at 25 Watts, panels are dirty/not tilted to latitude, etc.--Then you are looking at some darkness out there.

Anyway, interesting subject.

-Bill

Re: Solar Street Lighting

I am not sure, but I believe Bangalore (zero degree tilt?) is in the region of Nila:

Solar Radiation
  Annual Average: 5.26 (kWh/m2/day)
  Monthly Average
  Jan     5.36
  Feb     6.06
  Mar     6.56
  Apr     6.38
  May     6.03
  Jun     4.84
  Jul     4.50
  Aug     4.47
  Sep     5.03
  Oct     4.63
  Nov     4.50
  Dec     4.74

Goa on the east coast (from PV Watts) with 13 degree tilt:

Month      Solar Radiation (kWh/m2/day)
1      6.31     
2      7.01     
3      7.16     
4      6.96     
5      6.36     
6      4.65     
7      3.85     
8      4.59     
9      5.74     
10      6.08     
11      6.40     
12      6.25     
Year      5.94      

And Mysore India (zero degree tilt?):

Mysore, Karnataka, India

  Solar Radiation
  Annual Average: 5.10 (kWh/m2/day)
  Monthly Average
  Jan     5.44
  Feb     5.95
  Mar     6.39
  Apr     6.11
  May     5.65
  Jun     4.24
  Jul     3.88
  Aug     4.08
  Sep     4.89
  Oct     4.76
  Nov     4.81
  Dec     4.99

Generally, you do not design street lighting for the best--but the worse amount of sun in the area...

And due to local weather patterns, it appears it is hard to generalize the amount of sun available in the region.

-Bill