Simple Mass Component #810
Conversation
Replace Aviary example Jupyter notebooks
Removed jupyter output cells
Add files via upload
…nal_flight_phases_2.ipynb
…aft_mission_optimization.ipynb
| desc='optional data file of fuselage geometry') | ||
|
|
||
| self.options.declare('custom_fuselage_function', | ||
| types=FunctionType, |
There was a problem hiding this comment.
FunctionType is not defined? I don't know what kind of object is is supposed to be
There was a problem hiding this comment.
This component is super broken, the only input that actually does anything is thickness distribution. Span, chord, thickness, twist, all unused
| n_points = num_sections | ||
| jnp.linspace(0, 1, n_points) | ||
| 1 / (n_points - 1) |
There was a problem hiding this comment.
Not sure if the last 2 lines are supposed to be modifying n_points? Maybe missing a n_points = ?
| num_sections = self.options['num_sections'] | ||
|
|
||
| # Wing spanwise distribution | ||
| jnp.linspace(0, aircraft__wing__span, num_sections) |
There was a problem hiding this comment.
Numpy functions do not modify arrays, and instead create a fresh array and return that - need to capture the output of this function, probably intended to be num_sections = jnp.linspace(...)?
|
|
||
| if airfoil_file is not None: | ||
| aircraft__horizontal_tail__mass, _ = quadgk( | ||
| density * 2 * thickness * jnp.sqrt(1 + jnp.gradient(camber_line) ** 2), |
There was a problem hiding this comment.
wing.py uses thickness_dist here, I think that is the intended calculation? Missing here
| 'twist', val=jnp.zeros(self.options['num_sections']), units='deg' | ||
| ) # Twist angles -- no aviary input | ||
|
|
||
| self.add_input( |
There was a problem hiding this comment.
thickness_dist is a calculated value for the tail components, but an input here?
|
There are major discrepancies between the wing and tail mass computations that need to be resolved before I'll approve merge. Tests could use some additional cases, there are currently no tests happening for airfoil data files. If a data file for a NACA airfoil is given, the results should match the calculated value |
Summary
This is the "simple mass" contribution. It is fairly bare bones right now, accounting for a relatively simple wing, fuselage, and tail components, with mass calculated using basic physics. There is also a materials database dictionary inside to allow for various densities to be used. The test folder contains all the unittests for each component as well as the summation.
The components were built using OpenMDAO JAX -- with jax and jax.numpy.
Integration was done using quadgk, a package created to mimic scipy.quad but with JAX compatibility. The JAX developers do not plan to add scipy.quad to jax.scipy, and scipy.quad was favored for the integration compared to jax.trapz.
The basics for the wing and tail involve a 4-digit NACA airfoil and the corresponding equations that describe the thickness, chord, and camber. I have basic functions implemented inside it to allow for .dat files (assuming the airfoil data file follows the convention that UIUC's airfoil database follows) and interpolators to pick out these.
Related Issues
Backwards incompatibilities
None
New Dependencies
JAX, jax.numpy, quadgk, quadax