suavecode / suave

An Aircraft Design Toolbox

SUAVE: An Aerospace Vehicle Environment for Designing Future Aircraft

SUAVE is a multi-fidelity conceptual design environment. Its purpose is to credibly produce conceptual-level design conclusions for future aircraft incorporating advanced technologies.

License: LGPL-2.1

Guides and Forum available at suave.stanford.edu.

Contributing Developers

• Andrew Wendorff
• Anil Variyar
• Carlos Ilario
• Emilio Botero
• Francisco Capristan
• Jordan Smart
• Juan Alonso
• Luke Kulik
• Matthew Clarke
• Michael Colonno
• Michael Kruger
• Michael Vegh
• Pedro Goncalves
• Racheal Erhard
• Rick Fenrich
• Tarik Orra
• Theo St. Francis
• Tim MacDonald
• Tim Momose
• Tom Economon
• Trent Lukaczyk
• Walter Maier

Contributing Institutions

• Stanford University Aerospace Design Lab (adl.stanford.edu)
• Embraer (www.embraer.com)
• NASA (www.nasa.gov)

Simple Setup

git clone https://github.com/suavecode/SUAVE.git
cd SUAVE/trunk
python setup.py install

More information available at download.

Requirements

numpy, scipy, matplotlib, pip, scikit-learn, plotly

Developer Install

See develop.

Citing SUAVE

This respository may be cited via BibTex as:

@software{SUAVEGit,
  author = {
    Wendorff, A. and
    Variyar, A. and
    Ilario, C. and
    Botero, E. and
    Capristan, F. and
    Smart, J. and 
    Alonso, J. and
    Kulik, L. and
    Clarke, M. and
    Colonno, M. and 
    Kruger, M. and
    Vegh, J. M. and 
    Goncalves, P. and
    Erhard, R. and
    Fenrich, R. and
    Orra, T. and 
    St. Francis, T. and
    MacDonald, T. and
    Momose, T. and
    Economon, T. and
    Lukaczyk, T. and
    Maier, W.
},
  title = {SUAVE: An Aerospace Vehicle Environment for Designing Future Aircraft},
  url = {https://github.com/suavecode/SUAVE},
  version = {2.1},
  year = {2020},
}

The most recent publication covering the general capabilities of SUAVE was presented at the 18th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference and may be cited via BibTex as:

@inbook{SUAVE2017,
author = {Timothy MacDonald and Matthew Clarke and Emilio M. Botero and Julius M. Vegh and Juan J. Alonso},
title = {SUAVE: An Open-Source Environment Enabling Multi-Fidelity Vehicle Optimization},
booktitle = {18th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference},
chapter = {},
pages = {},
doi = {10.2514/6.2017-4437},
URL = {https://arc.aiaa.org/doi/abs/10.2514/6.2017-4437},
eprint = {https://arc.aiaa.org/doi/pdf/10.2514/6.2017-4437}
}

suave's Issues

regression script - propulsion

vary throttle, maybe Ma, Re
check thrust force

Need to add sweeps data class

We need to break out sweep into the sweeps data class with fields leading_edge, quarter_chord, and half_chord.

Add Engine Parasite Drag

Parasite drag functions do not include skin friction drag on the propulsion system.

Write Sizing Loop Functions

Summer 2014 Priority: 4
Use the aircraft function and simple drivers
Defer the issue of speed to a later task
Prototype a generic sizing interface
Sizing functions:
Mission by Fixed Range
input: aircraft geometry, mission profile(range, cruise alt)
variable: range of cruise segment, max togw
constrain: total range
Mission by fixed fuel burn
Mission by fixed mission time
Vertical Tail Geometry
Horizontal Tail Geometry
Main Wing Geometry
Engine Characteristics

# TODO: elaborate on inputs, variables, and constraints for each method

Solar network torque convergence.

Change solar network convergence to torque based rather than Cp based. This will preserve generality for the propeller model when used in conjunction with a variety of engine types.

Write Optimization Wrappers

Summer 2014 Priority: 7
Write the wrappers and drivers
Test whether suave behaves with GBO
An important demonstration of the code
Defer speed considerations to later task

vehicle has max_lift_coefficient_factor

should move this to the aerodynamics model

External Interfaces

Summer 2014 Priority: 14
Geometry (CAD, OpenVSP, EGADS)
Analysis tools (avl, xfoil, su2, nastran)

strange behavior for climb segment speed variation for E190

mission doesn't converge properly for different climb speeds of the 3rd climb segment (code line 443)
example: converges for speed = 389., 391., but not for 390.

see file in:
https://github.com/suavecode/SUAVE/blob/f098efed859812223645a2d4568bd073061388f3/scripts/regression/test_mission_Embraer_E190_constThr.py

weights - ignore missing components in empty.py

also consider adding warnings

Converge Naming Convention

Summer 2014 Priority: 3
Make sure we’re happy with the names of variables and outputs
Identify and clean out dead variables

reference this document:
https://docs.google.com/document/d/1L_T05NWoyAKlwvofUGF6MZHFxmtGtLuSxU-w5v3l2C8

regression script - aerodynamics

vary aoa, ma, re; test cl, cd, and components

regression script - static stability

update the cmalpha, cnbeta test scripts to use the new naming convention

Explore Gradients

Summer 2014 Priority: 9
What Method?
Complex Step
Automatic Differentiation
Analytic Differentiation

check payload-range diagram script

just making sure this works with the new class structure

Write "The Aircraft Function"

Summer 2014 Priority: 1
The forward analysis function
It should be ok with infeasibility

Input (Python script for this)
    Geometric Parameters
    Max Take Off Weight
    Mission Profile
Function (Python function for this)
    Build up empty weight (Andrew)
    Evaluate mission
        evaluate lift and drag
        evaluate thrust (evaluate propulsor interface - Anil, Emilio)
        evaluate stability (get this in to mission - Andrew)
    Evaluate Landing and Takeoff Field (Tarik)
    Evaluate Noise Correlations
    # Evaluate Cost
    # Evaluate Structure
Outputs (Data Structure for this - Trent)
    Component Weights
    Total Weights
    Range, Endurance
    Lift, Drag
    Stability Moment Derivatives
    Thrust, Fuel Burn, SFC, Efficiencies
    Landing Field Length
    Takeoff Field Length
    Second Segment Climb Gradient
    # Direct Operating Cost
    # Structure loads (Dive, Gust, Cruise)
    Magical fuel
    Magical thrust

landing segment initializes on landing weight

could initialize on weight of previous segment

https://github.com/suavecode/SUAVE/blob/master/trunk/SUAVE/Attributes/Missions/Segments/Ground/Landing.py#L45

Check the consistency between position and origin for vehicle inputs. Especially in the weights section.

explore documentation options

regression script - takeoff and landing field lengths

update fidelity zero tofl and lfl estimate tests, make sure they have expected values

Calculate the sea level static thrust for a given engine. Add this information to the vehicle data structure.

document typical development pipeline

might be something like experimental -> regression -> suave method -> suave attribute
the_aircraft_function.py is a good example here

scrub passt legacy code

Operational Empty

Replace operational empty with the correct operating_empty, mostly within the aircraft function.

Code Cleanup

Summer 2014 Prioirty: 10
Check for dead code
Check for good comments and style

document bare minimum segment functionality

need a name regularization method for appending vehicle components

for example: "Main Wing" -> "main_wing"
to match style and allow attribute acces

Input/Output Functions

Summer 2014 Priority: 13
Input: python script
Output: Text table in the Embraer Format

make stability model it's own model

currently - it's part of aerodynamics_model, and doesnt need to be

update mission regression tests to use the_aircraft_function

left over of issue #8

2.0 Release

Summer 2014 Priority: 12
Target: January SciTech

regression script - missions

update test missions to use the_aircraft_function

Address the Robustness of "The Aircraft Function"

Summer 2014 Priority: 5
Implement automatic_regression protocol for unit tests
Evaluate smoothness for continuous variables
For sub modules of the aircraft function too (ie Aero, Propulsor)
Test whether it accepts crazy values, and returns useful results

airfunc: integrate high lift correlations

Finish Continuous Integration

Summer 2014 Priority: 2
Fill out the test suite with functions that support
“The Aircraft Function”, so that we detect broken code
Spool up cruise control for nightly checks

Scripts:
    aircraft at one design point, single analyses
    start at low level (unit tests)
examples
tests
benchmarks

Implement offdesign engine performance

Address Code Speed

Summer 2014 Priority: 8
Profile the code to find bottle necks
Goal: 100 evals / min for “The Aircraft Function” at low fidelity

Check the units for the thrust for the 1-D propulsion module.

Component values difficult to update across entire setup for optimization

Variables such as wing area are copied into several different structures for the full setup. Different copies are used in different methods and determining which one to change in the optimizer is difficult. Changing all at once is also difficult.

Wing eta

Need an explicit name for eta to match conventions. (I assume eta is a ratio of dynamic pressures for S&C). Tim, let me know if you have any questions.

Need to implement improved weight estimation methods (TASOPT like and from Embraer )for engines.

Add/check ability to run single point segments

Current segments are only tested to run across multiple condition points. Add functionality if necessary and test segments at a single point.

Negative value for Cn_Beta

Fuselage contribution to the Cn_beta appears to be too high. Check for the K_Rel (#K_Rel: Correction for fuselage Reynolds number. Roskam VI, page 400.)

Line 152 on the Tube_Wing/taw_cnbeta.py

compatibility issue with scipy 0.10 - root

root was added as a consolidated interface for scipy's root finding functions in 0.11

Setting the Second Segment Climb

We need to include in the mission for the Boeing 737 a point that we can evaluate the second segment climb gradient.

Check aircraft induced drag for proper efficiency usage

Check that proper efficiency is being used for calculation. Current method uses full aircraft efficiency with aspect ratio of main wing.

https://github.com/suavecode/SUAVE/blob/master/trunk/SUAVE/Methods/Aerodynamics/Drag/Correlations/induced_drag_aircraft.py#L42

Write or Update Post Processing Functions

Summer 2014 Priority: 6
Placard Diagram
VN Diagram
Mission Profiles
Drag polar
Payload-Range

clean up Aerodynamics_Model()

list the things that are unclear

Code Documentation

Summer 2014 Prioirty: 11
Like eating broccoli
Tutorials

mission convergence issue: climb with constant throttle

When climbing with constant throttle and fixed speed, if the vehicle is not able to climb (too much drag, or lack of thrust) the solver will never converge since the final condition for altitude is higher than initial.

We must find a way to obtain some result different than error.

See file:
https://github.com/suavecode/SUAVE/blob/9b5885aa6b046d70286d575357903c7ecb47402b/scripts/experimental/test_mission_Embraer_E190_constThr.py

Try changing additional_drag_coefficient variable (line 35) from 0.0000 to 0.0075.