Guru
Joined: 09/08/2007
Location: AustraliaPosts: 4406
| Posted: 07:54am 31 Mar 2020 |
You need a variable voltage dc bench power supply and a series resistor of your choice.
Couple up your buck converter to that along with some type of convenient low impedance load to dissipate the anticipated power.
You will be drawing maximum power from the source when you pull the voltage down to exactly half of what it is unloaded.
By changing the value of your series resistor over a wide range, if the software "brain" always loads it down to half the open circuit source voltage, with a combination of different source voltages and series test resistor values, your algorithm works fine at finding max power transfer.
The idea is that max power transfer always occurs when a load impedance is perfectly matched to the source impedance. That will always pull the voltage down to exactly half, because its really just two equal resistors in series, source and load.
The source impedance will be your series test resistor, plus whatever internal impedance the dc power supply has. The load impedance will be your buck converter and actual final dissipative load in series with it.
Turn on the power and watch your software creep towards loading down the source voltage to EXACTLY half the open circuit voltage. That operating point equates with max power transfer.
Now solar panels are odd things, they are very non linear source, a charging battery makes up a complex non linear load as well.
If your algorithm can truly seek out and lock onto the power peak, it will aso work on something much more predictable such as a voltage source of known impedance.
So you really don't need to duplicate crazy solar panels and batteries for a valid test of your software.
Edited 2020-03-31 18:11 by Warpspeed
Cheers, Tony.