Dear Gerro, 

I will share my experience using a SPA with a micro-controller for a solar tracking system project I recently worked on. I originally attempted to run the SPA through the Arduino Uno but the capabilities of the board were exceeded and the script would not run. I upgraded to the BeagleBone Black (BBB) from Texas Instruments and this board was very capable of running the SPA. The BBB operated at 32 bits and has a 1GHz ARM CPU, 512MB RAM and 4GB storage. 

I did not use an external SD card for table storage. I only ran positions per day and stored only the current day's values in the on-board storage. The data was indexed to the times when the positions would occur and the system time was referenced the table to move the solar panels to the proper alignment. The time increments were sized based on motor accuracy and acceptable solar misalignment angles. 

There was no corrective factor included in the design. Instead during night hours the system would reset to a "home" position to ideally reset any mechanical errors that occurred during that day. GPS coordinates were uploaded upon install and no on-board GPS was present. 

I hope this proves to be somewhat useful. Feel free to message me if I can possibly help any further. 

Thomas

Dear Gerro,

My graduation project was on solar tracking system and we designed and carried out this type of tracking. Although we did not carried out the microcontroller programming, as another engineer dealt with this problem, I think I have a little experience about this type of tracking..

You will have to use two stepper motor, one for the altitude and the other for azimuth angles. In addition, two limit switches are required to be used.

When the system starts up, initialization mode is required. In this mode, moving parts for tracking altitude and azimuth angles should move until they press limit switches. At this moment, the microcontroller will realize that the altitude and azimuth angles are zeros.

Then, calculations of these angles will be carried out, as the date and time are saved on the system clock. After determination of the angles values, the microcontroller will send the signals to the stepper motors to take the right position..

Note that:

- angle step of a stepper motor is limited with a known value (I think it is about 1.8 degree, but I do not remember actually), so you have to determine the appropriate time for refreshing the system, i.e. calculate new values of angles and take the new position.

- I think that the main advantage of this type of tracking can be observed when tracking on a partially cloudy day..

Finally, I hope my answer to be helpful for you and I wish you succeed in your mission.. If you have any more questions about this, do not hesitate to e-mail me on [email protected]

Best regards,

Mohamed

Dear Mohamed, Thank you for the response, we are fairly familiar with the research topic of solar tracking (see our book in link attached), but we we more looking for responses on practical challenges experienced in using an SPA ephemeris pre-programmed sun path diagram with solar angle look-up tables in an open-loop PLC processor or micro-controller environment.

Is there an example describing implementation challenges such as memory requirements, lookup table sizes, ROM Eprom memory card SD card sun position tables, almanac date and time synchronization and trajectory indexing, sun trajectory pattern sampling rate, linear interpolation in pilot sun trajectory, mathematical precision, program code parameter handling (integer, floating point, bit-rate, decimal digits of precision), solar tracking error correction, adaptive learning or self learning to correct sun tracking errors due to mechanical misalignment defects, real-time processing speed for memory retrieval, processor speed improvement requirements, industrial automation PLC or processor platforms used, and ftp download tracker configuration setup for an electronic sun tracker controller.

Similar to the way in which a flight-path may be pre-programmed or hard-wired in a quad-copter or drone, I was hoping to receive examples where memory inside a memory or tracker SD card have been indexed according to the date, time and GPS location, and the sun vector memory data sweep data can be retrieved as a look-up function of time values by a micro-controller, reading the corresponding solar angles (azimuth zenith altitude) from an SD card.

Examples of these are be most welcomed.

regards

Gerro

My apology Mohamed this is the link

https://www.researchgate.net/publication/263085113_Solar_Tracking_High_precision_solar_position_algorithms_programs_software_and_source-code_for_computing_the_solar_vector_solar_coordinates__sun_angles_in_Microprocessor_PLC_Arduino_PIC_and_PC-based_sun_tracking_devices_or_dynamic_sun_following_hardware

Book Solar Tracking: High precision solar position algorithms, pr...

Dear Gerro, 

I will share my experience using a SPA with a micro-controller for a solar tracking system project I recently worked on. I originally attempted to run the SPA through the Arduino Uno but the capabilities of the board were exceeded and the script would not run. I upgraded to the BeagleBone Black (BBB) from Texas Instruments and this board was very capable of running the SPA. The BBB operated at 32 bits and has a 1GHz ARM CPU, 512MB RAM and 4GB storage. 

I did not use an external SD card for table storage. I only ran positions per day and stored only the current day's values in the on-board storage. The data was indexed to the times when the positions would occur and the system time was referenced the table to move the solar panels to the proper alignment. The time increments were sized based on motor accuracy and acceptable solar misalignment angles. 

There was no corrective factor included in the design. Instead during night hours the system would reset to a "home" position to ideally reset any mechanical errors that occurred during that day. GPS coordinates were uploaded upon install and no on-board GPS was present. 

I hope this proves to be somewhat useful. Feel free to message me if I can possibly help any further. 

Thomas

Dear Gerro,

Maybe you could contact solar telescope holders (like THEMIS or GREGOR in the Canary Islands). The first telescope uses such table to track the Sun (and even more complete tables since it takes into account the flexion of the telescopes on each axis and the variation of such flexion with time).

http://161.72.34.10/dokuwiki/doku.php

You can also find some papers on the pointing accuracy of the telescope in some of my papers presented in my page.

Carine

Former support astronomer at THEMIS

http://161.72.34.10/dokuwiki/doku.php

I suggest you contact Ibrahim Reda at the National Renewable Energy Lab for an answer to this question. His email is:

[email protected]

I have made some solar tracking devices. For the required precise, the double float data type is needed when use astronomical algorithms in high accuracy sun tracking. So, microcontroller can not satisfy this requirement. Arm7 chip is recommended. I used arm7 to carry out spa algorithm well. For PLC application, the sun position should precalculated using PC with high precise astronomical algorithms. For one GPS location, a 512M flash may store 20 years sunposition data. In practice, it is neccessary to revise the data using a sunlight sensor. Because there are errors in installation of solar tracking devices, errors in deformation lead by gravity and wind load, uniform inground settlement. What  I want to say is that it is not neccessary to use SPA algorithms unless astronomical observation. For solar engineering, low precise algorithms combining with sunlight sensor is more practical in cost and in reliability.

In attach, there are a C file for high precise sunposition and a low precise algoritm in matlab language. 

The link is a paper using  low precise algorithms combining with sunlight sensor

The National Renewable Energy Laboratory (NREL) has written a public domain very precise sun position modeling tool. Tables could be generated from the tool to be used in devices with limited computation power and memory. Contact Ibrahim Reda at NREL.

I appologise for my ignorance or confusion. What "precision" stands for here? Dispersion of computed values (as needed by solar optical observations, THEMIS) or error in computed values (as needed by solar energy collectors, NREL)? And what is the target value of this "precision" to be achieved (absolute or relative to what)? Once this clarified it sets the computational speed needed as the Sun drifts with half a degree (one diameter) in 2 minutes. Have you considered using a small lookup table of a limited number of trigonometric functions values? In this way you can avoid calling trigonometric functions inside the algorithm (as in the early ages of electronic computers).