I'm planning to make a proper simulation of a circuit. Before the simulation, I need information about the load (resistance, inductance, and capacitance).
Example: the load is a CPU. range of wattage 30 - 120 W. I can calculate the resistance:
P=IV, I=P/V, I=120W/240V thus I=0.5 A.
V=IR, R=V/I, R=240V/0.5A thus R=480 Ω.
or
P=V²/R, R=V²/P, R=240²/120 thus R=480 Ω.
How about inductance and capacitance? Which formula should I use? Or can I use an LCR meter?
Identifying and characterization of a an LRC configuration requires time-dependent information. In general:
$$R = \frac{v_R}{i_R}$$ $$L = \frac{v_L}{\frac{di_L}{dt}}$$ $$C = \frac{\frac{dv_C}{dt}}{i_C}$$
Only for the resistance you can get away with a constant voltage. For the other two you at least need a changing voltage and/or current, since their expressions contain derivatives w.r.t. time.
The cheaper LCR meters will simply apply a sine wave to the circuit and then try to measure the complex impedance for that frequency (eg. sine wave amplitude and phase). In that case (I'm using phasor notation here)
$$Z = \frac{\underline{V_R}}{\underline{I_R}}$$
The impedance of a capacitor and inductor is:
$$L = \frac{\underline{V_L}}{j\omega\underline{I_L}}$$ $$C = \frac{j\omega \underline{V_C}}{\underline{I_C}}$$
Depending on how you are combining them in your model, you can usually calculate it pretty easily from the phasors.
For example, if you have an \$R\$ and a \$C\$ in parallel, then the combined impedance is:
$$Z_{RC} = \frac{\underline{V}}{\underline{I}} = \left(\frac{1}{R} + j\omega C\right)^{-1}$$
You calculate the complex impedance, and match the real and imaginary parts to find both \$R\$ and \$C\$. If you find a negative \$C\$ then you most likely have an inductive load instead and you'll have to use this formula instead:
$$Z_{RL} = \frac{\underline{V}}{\underline{I}} = R + j\omega L$$
You can also try to derive the capacitance/inductance using transient waveforms of known cases. For example, if you have an almost ideal step response you can perhaps derive a model from that (if it's not too complex). For example, a parallel RC network will exponentially charge up, or an RLC network might show some ringing. Depending on the results, this can be pretty hard to do though.
With one frequency you cannot characterize more than 2 components. If you however really want to go all the way, you have to up the number of frequencies measured. Techniques for such characterization are for example applying random phase multisines and measuring the fft's of both voltage and current across the load. Then you come up with a model taking the whole frequency spectrum into account. I wouldn't really recommend the latter if all you want is a "simple" model, as it's more something you might see in more advanced research-driven projects.
