Discussion Forum

Hi

if you are already in a transient solver case, you can change the radiated power from a constant or step(t) turn on type to a pulse(t) type, the step(9 and pulse(9 operators of COMSOl assures a smooth transition region for turn on and off

--
Good luck
Ivar

thanks for the answer.

I'm Already in a transient solver case, can you help me to set step(t) function? i set 20s in time solver;

then for the step function: i called the function "step"

in parameter i put location: 5. (in the sense that 5 are the seconds, the time in which i want to change the power); is this correct?

in the "from" field i put: 700 (the start power)

in the "to" field i put:0

then i put the name of the function "step(t)" in the port power.

in this way i want to simulate 5s oven ON(with the power of 700W) and 15s oven OFF. is this correct?

thanks

Hi

you can use the Definitions Function Rectangle, lower limit 5, upper limit 15 and smooting 1 then plot it to check the shape

Then define your source power Q0*rect1(t[1/s]) (the [1/s] is to format the time t in seconds without units, you get rid of the orange unit error like that.

Then check your time solver, you need to have it in strickt or intermediate and with time steps that are hitting a few time the up and down stransition ranges, such as {0 2 4 4.5 4.75. 5 5.25 6 8 11 14 14.5 14.75 15 15.25 16 20 25 30 35 40}
and play a little around. The "strict or "intermediate" instead of the "free" mode is adjusted in the Time stepping tab of the lowwer "Time-Dependent Solver" sub node

--
Good luck
Ivar

hi,
i tried to use that function, but there is a problem with the solution. infact there is no convergence.

i'm using the model library microwave oven with te potato; i tried to simply modify the imput power of the port of the microwave oven:

in the model it was 500W, so after 5s i saw the final temperature;

then i change the input power with a time function such as step, or piecewise, or rectangle to see after other 15 s the cooling: 5s input to 500W + 15s input to 0W.

the solution starts, but there is no convergence (after 30 minutes the solution was still running;(without the new function after 3 minutes there was the solution)

i tried also to put not only 0W, but also for exampe 100W, but it is the same.

can someone help me please?can someone open the comsol library model and help me?

thansk a lot!!!!

Hi

what are you time step settings of the time solver ?

that is time steps, and solver type BVDF, ... and free, strict or intermediate time stepping
What does your log say ?

You should look into the data comsol gives ou to learn how to analyse these ;)

and have you turned on the plot while solving , for ALL steps so you see what is happening with your model, or added a few probe plots ? COMSOl provides many tools to monitor the progress of the solver. You should not consider it only as black box ;)

--
Good luck
Ivar

hi,

i'm using a BDF free.

start at 0 and end at 20s;

i put various steps between 0 and 20.

the problem for me is near the 5th second, in which ther is the change of the function from 700W to 0W.

on the screen, viewing the temperature, up tu 4.75s all works fine (as the constant function), then the solution convergence increase, instead of decrease, so it after 30 minutes reaches 10e9.then i stop the analysis.

i put also instead of 0W, another positive value, but the problem is the same.

i tried with rectangle function,step function and piecwise function. but no solution

Hi
Exactly, if you are in free, the solver expect a time response of an assymptote, but you have a rapid change, change the "free" to intermediate or strict and there is a fair chance your model solves without other surprises

--
Good luck
Ivar

hi Ivar,

thanks...now after 25 minutes i have a solution,

i use a piecewise function:

0-5 s 700W
5-20 s 0W

with bdf and intermediate solver.

now i will try again with other setting...

thanks

Dear Francesco,

I tried to do the same that you have done, but I had the same problem with the convergence. Could you check the attached simulation, and try to help me.

Best Regards,

Diana Fidalgo

Hi
I’m using the module of the microwave heating (the same model of the model library microwave oven with the potato) but I modified the power of the port of the microwave oven with a square pulse train (500W on and 0W off). I wrote a local waveform function (wv1) and so the power of the port was Q0*wv1(t[1/s])
I’m using Frequency-Transient and "intermediate" instead of the "free" mode in the "Time-Dependent Solver" sub node. The model is converging! But when the power is 0W, there is a fall in the convergence and increasing the time solution. Is there any solution for that problem?
Can someone help me please?
Thanks