Using trapz for integrating with nonuniform spacing

Hello,

I am trying to compute the loss of an inductor by using data points for time and flux data from simulation. The simulation result for time has non-uniform spacing so I am using trapz to integrate instead of using sum since dtime is varying. However, the result I get when using trapz is significantly small and not close to the expected value.

Case 1: sum - take the average value of the time steps and do the integration using sum - result is close to expected value

Ploss = (dtime/T) * ki * Bptp^(b-a)*sum( (abs(dBdt)).^a );

Case2: dtime in this case is the time step between each data point so I am trying to integrate using each time step

Ploss=(ki * Bptp^(b-a))/T*(trapz(dtime,(abs(dBdt)).^a,1))

What could I be doing wrong in case 2? I believe that case 2 is a better approach since it uses the actual time step data instead of taking the average.

Thank you.

Answer 1

[Torsten]

You mustn't use "dtime" in the call to "trapz", but the actual time vector t for which the simulation gave you the corresponding vector dBdt as result.

[Firehiwot Gurara]

Thank you! This solved it.