Piecewise function fitting

[mugger]

Please see the attached figure for the question. 

I have this piecewise force signal as the function of time . The piecewise function is made of 2 lines:

One slope line is connected by one flat line. 

From the raw data, or real signal, shown in upper panel, it is quite noisy. I would like to seek a solution to fit the data to the ideal piecewise function as shown in the lower panel. 

1. How can reduce the noise, especially the high frequency components? 

2. How can I get the turning point location and slope by fitting? 

Any help or ideas will be appreciated, thanks!

[altenbach]

How about a nonlinear fit where the VI model has the following four parameters (There are three sections, not two):

slope, offset, turning point1 (where it starts deviating from horizontal), turning point 2 (where it becomes horizontal again)?

(looking at your image, I think your model is an oversimplification of the actual data. What kind of experiment is this? Is there a more comprehensive theory?)

Don't reduce the noise, have the fitting handle the data reduction directly. Filtering noise always has a potential to bias the data in some way or another.

Can you attach a VI containing some typical data in a diagram constant?