Implement continuous version of SignalExtrema

[AHaumer]

Triggered by a discussion with @GallLeo and looking into #3762, I developed a continuous block to detect signal extrema, based on derivative. It works for non-differentiable inputs, too.

[christiankral]

@dietmarw @AHaumer We had a discussion in #3762 (comment) regarding whether or not the name shall contain Continuous or Sampled not. I think we are back to this point. One simple way out is to move the models:

• Modelica.Blocks.Math.ContinuousSignalExtrema => Modelica.Blocks.Math.Continuous.SignalExtrema
• Modelica.Blocks.Math.SignalExtrema => Modelica.Blocks.Math.Discrete.SignalExtrema

[dietmarw]

I think the proposal of @christiankral is the way to go.

[christiankral]

I think the proposal of @christiankral is the way to go.

@AHaumer @HansOlsson @MartinOtter do you have any objections

[HansOlsson]

It seems to work.

My only concern is that it may cause unnecessary events in some cases - e.g., using an expsine-signal it seems we will continue to get events every cycle (due to u>=x and u<=x) even though we are far from an extremum.

I don't know exactly how to safely fix that (I can think of several possibilities - the problem is to ensure that they work), so I propose we delay such improvements until needed.

[beutlich]

As already mentioned by @dietmarw in #3762 (comment), it might be worth to point again to the LastLib proof of concept how to calculate the signal extrema without events or derivatives.