Close

MPPT Algorithm work

A project log for Solar Energy Generator

A solar energy system up to 500W in power for use with lithium batteries.

nathaniel-verleeNathaniel VerLee 08/14/2014 at 03:260 Comments

I have been working on getting an MPPT algorithm up and running on my current test setup, which currently looks like a complete mess:

You can see the converter stage, display, microcontroller, a current and voltage sensor board, and a breadboard with supply rails all mounted to a lovely piece of poplar. The big heatsink in the background has a large 4 ohm load resistor for testing the converter stage. 

My current efforts are focused on getting MPPT up and working. A typical MPPT algorithm working something like this: http://low-powerdesign.com/article_CYsolar_071311.html You can see in the Algorithm flowchart that it makes its decisions about how to change the duty cycle of the converter based on both power change and voltage change. Since the voltage of the DC supply voltage remains constant, its kind of a toss up if the voltage increases or decreases - its really just noise at that point. The current algorithm ends up making silly decisions and never finds the max power point of the DC supply. It actually is possible to find the MPP of a DC supply like this, since the supply will go into constant current limiting at some point - I would just need a dramatically different algorithm. I do want to be able to charge batteries from both a panel as well as a general DC voltage source, but instead of making an algorithm that will draw more and more power until it trips the DC supply's overcurrent protection, I will just make an option to input how much power the converter may draw from the source and tell the system what kind of source it is drawing from. 

I have decided that I really need to do two things to move forward: 

1) figure out how well my voltage and current measurements actually are (and if they are very noisy, is it just the ADC or the sensors themselves). This will help in determining how big of a duty cycle step size needs to be made when searching for the MPP. 

2) Test the system out on a sunny day with a real solar panel (since I dont have a solar panel simulator) to see if my algorihm actually works. 

Hoping to get a video done as well this weekend, and also create a short PDF on the specs, design goals, current state of schematics, etc. 

Discussions