Horrible Maths Q ---- errk!!

[0mega]

Arrgh... i have been ripping my hair out over this one...

I have an equation for a maximum power point tracker someone on the forum kindly donated to me .

The derivative is

dP/dV = Isc + (-BAV*e^(BV) + A(1-e^(BV) ) )

ok... now how do i solve for v that

Isc + (-BAV*e^(BV) + A(1-e^(BV) ) ) = 0

(hell, i can't even transpose for V! erk!)

Thanx,
JB

 

[crust]

It was me that posted those equations. Can I ask what you are trying to do? It seems you are trying to find the maximum power given a voltage. Well, to do that, you would need to know the A/B of your solar cell, which you can measure experimentally. (or I can give you values from my 75W unit).

As for solving the equation, I might be mistaken, and you should definately double check, but I believe that is a transcendental function. There is not (at least that is what a professor told me a few years back) a way to solve it algebraically. You have to just plug in numbers until you get close -or- use a function solver like the one on the HP48 calculators.

 

[0mega]

Dammit~! If that is true, that really sucks....

It is for an MPPT in a solar car... i was hoping there was some way to solve it without using "guess" methods, as that's going to chew up code anyway, thanks for ur reply.

A and B will be solved by regression, btw. Do the a & b values change depending on sunlight, or are they constant for panel ?

 

[tavib]

:lol:
Your transcedental equation is a good example for usefull of a computer other than games or surfing internet etc.
With a little programming knowledge, this ugly equation can be solved with a high precision using a computer(regular PC is perfect).

 

[ljcox]

Arrgh... i have been ripping my hair out over this one...

I have an equation for a maximum power point tracker someone on the forum kindly donated to me .

The derivative is

dP/dV = Isc + (-BAV*e^(BV) + A(1-e^(BV) ) )

ok... now how do i solve for v that

Isc + (-BAV*e^(BV) + A(1-e^(BV) ) ) = 0

(hell, i can't even transpose for V! erk!)

Thanx,
JB

You could do it with Excel.

Let y = Isc + (-BAV*e^(BV) + A(1-e^(BV) ) )

and produce a table of y versus v using steps of say 0.1 Volt intervals. Then once you have an approximate answer, reduce the steps to say 0.001 Volt and do it over a smaller range, ie. the range that includes the approx answer.

PS. I don't see why you included the derative.

Len

 

[StupidDum]

This is not 'guessing', there is something called "numerical method"

 

[crust]

A & B are constant for the solar cell. Once you find the values, you are good to go. Further, using numerical methods to find the solution to that equation is definately not guessing. If it were, I suspect that the menu on my hp calculator would have a choice for "Equation Guesser" rather than "Equation Solver"

 

[carlosthejackle2001]

good program for solving equations is a maths program called derive, download a demo type it in and hey presto sort of....

 

[crust]

Getting off-topic but the HP48 simulator is also very good for this type of thing. Not anywhere near as nice as the actual calculator, but the exact same look, feel, and operation. It is just hard to type in letters and operate the shift keys.

 

[ljcox]

I have attached a zipped Excel file for you that solves the trancendental.

All you need to do is to change the A, B & Isc values.

I manipulated the maths to the form

1 + Isc/A = (1 + BV) e ^ BV

So Y1 = 1 + Isc/A and Y2 = (1 + BV) e ^ BV

The solution is where the line Y1 and the curve Y2 intersect.
If the intersection is not on the graph, change the V range and , if necessary, the increment.

Please check the maths.

Len