Modelling friction factors for laminar vs. turbulent flow within the same network #754
Description
Is your feature request related to a problem? Please describe.
As discussed in #740 , we have a mixture of implementations for deriving the friction factor lambda with different methods that all have their drawbacks and are often limited to certain flow regimes. E.g. the colebrook algorithm is only precise for turbulent flows. How can we ensure precise results over the whole range of Reynolds numbers while maintaining integrity with other tools?
Describe the solution you'd like
Ideally, we have a number of different options for the friction factor model, e.g. just choosing one algorithm for the whole flow regime (such as steady Nikuradse) or different models for different flow regimes (i.e. deriving lambda depends on the result of the Reynolds number for one branch. In such a mixed case, we need to ensure that we steadily converge to a solution.
Describe alternatives you've considered
Until now we imitated approaches of other usual tools, such as STANET, but it might make sense to take it a step further. There should be no harm in having a larger number of different approaches, some of which should be physically more feasibly while probably being less performant. It should be up to the users to decide which models fit best for their problems.
Additional context
Related discussion in PR #740 .