how to select the capacitance of a capacitor?

Thread Starter

almotions

Joined Feb 6, 2009
46
Hi all,
assuming i want a constant voltage of 100V to be supplied from a capacitor,how should i choose the value of the capacitance? Example,in simulink, the capacitor parameters only take in the capacitance value and initial voltage,how do i specify that the capacitor actually supplies 100V?

Thanks.
 

t_n_k

Joined Mar 6, 2009
5,455
Hi all,
assuming i want a constant voltage of 100V to be supplied from a capacitor,how should i choose the value of the capacitance? Example,in simulink, the capacitor parameters only take in the capacitance value and initial voltage,how do i specify that the capacitor actually supplies 100V?

Thanks.
If you had an infinite capacitance charged to 100V you could use it as a constant source. Trouble is, all finite capacitances discharge as you draw current from them, so what you are asking is not possible - unless you are really talking about selecting a capacitive filter value for a rectifier circuit, for instance.
 

John P

Joined Oct 14, 2008
2,026
For most practical purposes, you have to decide how much current you need to draw from the capacitor for how much time, and how much voltage drop you can allow over that time interval. It's like drawing water out of a tank--the pressure will drop as it empties. And the bigger the tank, the longer it will hold up the pressure.

The equation is very simple:
I*T = V*C

So if you had a 1 farad capacitor charged to some voltage, and you let 1 amp flow out of it for 1 second, you'd reduce the voltage on it by 1 volt. It's that simple! But those aren't very likely numbers.

Your "supplying a constant voltage from a capacitor" isn't realistic. You have to think in terms of what variation in voltage you can tolerate.
 

Thread Starter

almotions

Joined Feb 6, 2009
46
hi tnk,
Sry for not making this clear.i'm using the capacitor as a source of an inverter supplying a pure reactive load.it could be two or three level inverter,therefore the current is AC 90 degrees out of phase with the voltage,the charge will remain constant.However,i just have the problem fixing the DC voltage across the capacitor in SIMULINK. Unless my assumptions are wrong, any ideas?
 

MikeML

Joined Oct 2, 2009
5,444
Are you asking how to set the initial conditions (starting voltage)? This is called .IC in Spice; dunno about SIMULINK?
 

t_n_k

Joined Mar 6, 2009
5,455
hi tnk,
Sry for not making this clear.i'm using the capacitor as a source of an inverter supplying a pure reactive load.it could be two or three level inverter,therefore the current is AC 90 degrees out of phase with the voltage,the charge will remain constant.However,i just have the problem fixing the DC voltage across the capacitor in SIMULINK. Unless my assumptions are wrong, any ideas?
I see ...

I imagine Vcap won't remain a constant or fixed DC - rather it will probably cycle at the inverter output frequency about an average DC value.

As MikeML suggests you should be able to set the capacitor initial value. Like him I've not used SIMULINK - but it would seem absurd not to have this capability. Otherwise you might be able to set up an initial capacitor charging sequence through a switch + resistor network which "disconnects" after a preset time in the overall simulation sequence.
 

Thav

Joined Oct 13, 2009
82
tnk is right here, the average capacitor voltage and current will be \(V_{C0}\) and 0 respectively, but there will be ripple currents and voltage quite handily related. If you know the amplitude of the output current you're going for, you can solve this...
\(\Delta V = \frac{1}{C} \int_0 ^{\frac{T}{2}} I_{M} \sin{\left(2 \pi f t\right)} \mathrm{d}t\)
... which will give you the ripple voltage in each cycle, assuming the current is independent of the capacitor voltage. That assumption make this good as an estimation, but if your cap is large enough it is a good estimation.

Also, in any real inverter supplying a purely reactive load, your capacitors will still discharge because of losses in the cap, switches and wiring.
 
Top