Add effects due to area expansion

[rowancllzz]

Fixes #58

This pull request implements changes to the two-fluid EBTEL equations to account for cross-sectional area expansion following the methodology of Cargill et al. (2022). This PR represents a substantial modification to ebtel++ and includes the following changes:

• [x] Add additional input parameters:
  • $L_{TR}/L$ (loop_length_ratio_tr_total)
  • $A_{TR}/A_C$ (area_ratio_tr_coronal)
  • $A_0/A_C$ (area_ratio_0_coronal)
• [x] Modify derivatives for $p_e$, $p_i$, and $n$
• [x] Modify $c_1$ calculation to include area expansion effects
• [x] Modify initial conditions calculation to include area expansion effects
• [x] Modify velocity calculation to include effects due to area expansion
• Check that DEM calculations do not need to be modified to account for area expansion (Move to its own issue)
• [x] Update documentation to include new input parameters
• Update documentation to include EBTEL equations with area expansion and two fluid effects since these are not published anywhere (Move to its own issue)
• [x] Add additional tests for different area expansions
• [x] Confirm that no expansion case here recovers same results as ebtel++ without this PR
• Add HYDRAD comparison tests for area expansion cases (Move to its own issue)

Additionally, this PR also includes several minor modifications:

• [x] Update default compiler options to use C++-14 to satisfy Boost Math requirements
• [x] Move time step control calculations from main() to a separate method on the Loop object

[wtbarnes]

These files are the spatially-averaged HYDRAD results I've been using for benchmarking. I'm preserving them here to remember what we used to test these new additions against.

no_expansion_electron.txt no_expansion_ion.txt no_expansion_single_fluid.txt q1_tanh_electron_trac.txt q1_tanh_ion.txt q1_tanh_single_fluid.txt

[wtbarnes]

See this gist for an example of some of this benchmarking: https://gist.github.com/wtbarnes/ef4b6401d100399503a956be1d038b5c

[wtbarnes]

I verified that the C++ and IDL results match one another to an acceptable degree and that any remaining differences are not due to this PR. The generic differences between this code and the IDL version are being tracked in #40.