Time serie (constant curve type) problem with 5.x

I am facing a problematic problem. In the time step editor, the "constant" type is assigned to my curves, however they appears to be interpolated as polylines in the "time relate data", and they are effectively polyline interpolated when looking at the results, which is very annoying in my case (daily recharge with long dry periods and flash rainfalls).

Thanks in advance for your help !

Let me updated my post, here is what is happening:
My input record for recharge is that one:

Day Recharge
3804 0
3812 287
3813 0

And here is what uses Feflow
3804 0
3811 0
3811,19 27,36
3811,41 59,453
3811,68 98,2
3812 143,5
3813 143,5
3813,33 96,7
3813,7 43,16
3814 0

If you plot those on a chart, you will see that the surfaces are correct (I mean the total volume of recharge seems correct), however the input signal has became smoother (half of the amplitude, twice the lenght), which I want to avoid. Any advice on how to proceed ?

The "constant steps" option is not available for time-varying material properties, only Akima or Polylined. If you want to use a "constant steps" type interpolation for material properties, you should set up a polylined time series that looks like a constant steps curve. For your example this would be:

Day Recharge
3804  0
3811.99 0
3812  287
3812.99 287
3813  0

Danke Julia, cheers !
Didier

I am reopening this issue. I had a complete trust on this solution however we just checked our mass balance for each time steps, and we still have anoying linear interpolations. I attached a chart showing the difference between the assigned and simulated recharge time-series. The labels on the chart show which X -(time) values are used.

Any recommendations ?

Cheers,
Didier

Didier,
Have you checked the T-List data (time-varying material data) after transferring the data from the time series to the T-List? It would be important to know if the data input went wrong or if something indeed happened during the simulation.

Peter

Thanks Peter,

Effectively, yes the T-list is the same as the dac signal.

Could you send the time series to support@dhi-wasy.de? If possible, I'd like to check the transfer to the T-List myself to see what went wrong.