Two Axis Solar tracking system

[ecaits]

Dear Friends,

I am planning to design two axis solar tracking system using pic controller (like PIC16F877A), LDR and two DC motors.

I will compare voltage of two LDR for one DC motor rotation (Either in reverse or forward) and other two LDR for second DC motor rotation through L293D driver circuit.

I am planning to do programming in MPLABX developing tool.

I want some suggestion regarding reliability, accuracy, performance, hardware requirement etc.

Plz suggest me for that.

 

[udayan92]

If you want to have a sharper and more precise control of the position of the area which will be receiving the sunlight,then use stepper motor(s).They can be interfaced much more easily than the DC motors,as programming them to run,will require much less effort as compared to DC motors which you will be using.Also,use not one,but a group of several LDRs placed together in a row.This will have more inputs and you will be requiring multiplexing,but control and accuracy of the system would be more precise this way.

 

[Sweta25]

Even I am working on a similar project...I am using photodiodes as sensors and two stepper motors and FPGA as controller. Even I would suggest stepper motors

 

[D.A.(Tony)Stewart]

You can use 10 deg RED or yellow LEDs as 10 "photo diodes" , which they do not make in 10 deg and use a hemispherical mount of an array to cover your range every 20 deg. Then read the max voltage using -5V max or reverse bias MAX and 1M//1uF on each sensor to ground and compute if solar energy gain is worth the loss to move motors. Consider clouds diffuse optical power which leads to low variation in sensors and a clear sky gives sharp variation.

If no ESD damage, LED leakage on RED, Yellow is <1uA. White, Blue is <10uA. IR 6 degree may also be considered like those used in remote controllers. To block stray light, use heatshrink over 5mm parts and scan Voltage with ADC up to 150kLux down to 1k-10k Lux.

The mechanical forces of wind on PV array are high and require support from both sides... Not a pivot point. High pulley ratios are needed with a tension control or chain drive.

You might be able to reduce array size with more statistical calculations , but single LDR is very crude.

You could also use cheap camera with wide angle lens and scan pixels for MPT before hunting and wasting power.

Auto- Calibration is require to correct offset between sensors and PV power as both are "LDR's" but may not peak in sync.

smart trackers might have TIme of Day history of target angles and if batteries are fully charged near end of day, use excess solar energy to move to morning position before possible battery depletion over-night.

 

[Easyrider83]

https://www.pc-control.co.uk/howto_tracksun.htm

 

[jiripolivka]

https://www.pc-control.co.uk/howto_tracksun.htm

What a nice waste of time and money!
I know that you can design wonderful electronics to drive solar boards after the Sun! I also know that the angular accuracy needed to keep a solar battery pointed to the Sun is of the order of a dozen degrees. And clouds would drop the output more than any misguiding.

Imagine that since 18th century, astronomers have used simple clock to drive their telescopes after any chosen star (Sun is one if you happen to miss it), with a precision of a fraction of one degree.
With today's clocks it can be done even better, for a cost of one dollar or less.
What a waste of effort of you digital heroes!

 

[D.A.(Tony)Stewart]

I agree but a 1 ppm clock for a few $ is need otherwise solar calibration of clock is required and energy consumed in tracking must be predicted before the gain in wasting energy driving the motors. Which is why I suggested cheap 10 deg LEDs to find optimum vector position has optimum sensitivity.

Solar tracking requirements are affected by latitude but I suspect if it is not cloudy. An optomistic improvement net gain is 25-40% depends highly on number of sunny days.

a pessimistic estimate might include a net loss of energy on some days with an installed higher fixture capital cost and risk of wind damage if underestimated wind gusts.

 

[doraemon]

Hello!

I fully agree with Jiripolivka.

On top of that, most people don't realise that sun movement is almost a single axis
movement if we think about one single day. Over a year, you have a kind of drift of
the axis, +/- 23 degrees north and south (because of earth inclination).
Cos (23 degrees) being 0.92, it means that in the worst case (24th of December or June),
you will loose 8% of the energy for misalignment. Only 8%!!
So the inclination of the axis will depend on your location. Example: here in Japan,
the lattitude is around 35 degrees. So if you spin your solar panel (or whatever) around
an axis at 35 degree (parallel to earth axis), and if the speed is exactly 1 rotation
per day, then you can point exactly to the sun direction.
Let's add that using light sensors is not accurate. They are highly analog, and it will
depend on a calibration. And the calibration may drift in which case the 8% advantage
will vanish. On top of that, if you want to achieve a servo loop, it will cost you quite
some energy.

If you want a clock with no drift at all (not even 1ppm), you can use a radio frequency
clock which exists almost everywhere and can be built cheaply.
By the way, there is no need to drive the motors permanently. Define your maximal loss.
If you want to limit the power consumption of the adjustment, you may tune the position
only every, say, 20 minutes. 20 minutes correspond to 5 degrees. As the sun movement
is quite predictable, you can have a maximal deviation of +/- 2.5 degrees. As cos(2.5
degree) is 0.999, you will loose 0.1%, but spin your motor briefly every 20 minutes to
move by 5 degree.

And of course, if you don't want to loose the above mentioned 8%, then you can add a
daily adjustment of the antenna by reference to the spinning axis, that you can perform at
night for the next day. Note that the spinning axis never changes and is parallel to earth
axis.

Dora.

 

[D.A.(Tony)Stewart]

The biggest loss is longitudinal tracking error but then the solar power does not peak at dusk or dawn.

But then how much more does it cost for 2 axis vs 1 axis. vs capital cost and gain by choosing a bigger array with 1 axis control or a bigger array still with no solar tracking. ( assuming cost parity)

It is a life cycle total cost of capital . maintenance in bucks / pesos / yen per Watt-hour of life time capacity including installation costs and inflating energy costs.

Some claimed as I indicated 25`40% gain but Incremental capital costs per weight or Watt-hr capacity are only known by the few who have already bought.


The tracking design decisions are trivial compared to the optimal decision making of economic recovery of capital. ( unless you specialize in this)

 

[ecaits]

Can you post your circuit for my reference.
Which is better, photo diode or LDR?

Thanks,