Skip to content

Latest commit

 

History

History
68 lines (47 loc) · 4.49 KB

README.md

File metadata and controls

68 lines (47 loc) · 4.49 KB
Raw

Dawn Design Tool

by Peter Sharpe and Annick Dewald

Description

Automatic-differentiation-accelerated multidisciplinary design optimization for high-altitude long-endurance (HALE) aircraft.

Related publication: An Optimization Approach to Mapping the Feasible Mission Space of a High-Altitude Long-Endurance Solar Aircraft

For more documentation, see the memo in this repository.

Example feasibility chart outputs:

Feasible Mission Space

Installation

There are several easy ways to get started with Design_Opt_HALE! (Assuming you already have Python >=3.7 installed, preferably via the Anaconda distribution - if not, do this first.)

  1. First, you need to install AeroSandbox, an open-source Python package I started last year for automatic-differentiation-accelerated multidisciplinary design optimization for aircraft.
    • You can do this by opening up Command Prompt on Windows or Terminal on Mac/Linux and entering pip install aerosandbox.
    • Windows users: if you get any permissions issues, run cmd.exe as administrator and try again.
    • Mac/Linux users: if you get any permissions issues, type sudo pip install aerosandbox and proceed to run with administrator access.
  2. Next, you need to clone this repository (Design_Opt_HALE).
    • Users who are familiar with Git: the clone target is git@github.mit.edu:pds/Design_Opt_HALE.git if you're using SSH or https://github.mit.edu/pds/Design_Opt_HALE.git if you're using HTTPS.
    • Users who are new to Git: I think the easiest interface to Git for most new users is GitHub Desktop. Here is a good introduction video (15 mins). Once you've watched that, congratulations! You now fall into the "familiar with Git" category above and can use the instructions in the previous bullet. Aren't recursive README instructions great?
    • Users who are new to Git and don't want to use Git: Just download a local copy of the current repository from here. Beware that this copy will not stay up-to-date with future improvements, so this may quickly become outdated!
  3. Congratulations! If you've gotten here, you've successfully installed Design_Opt_HALE.
    • To test it out, try running Design_Opt_HALE/design_opt.py in your favorite Python IDE. If you don't have a Python IDE, I'd humbly recommend PyCharm, but this is a matter of personal taste. You can also run it from command line with python <path_to_directory>/Design_Opt_HALE/design_opt.py or similar.

If you run into issues at any point, follow the troubleshooting steps in the eponymous section below.

Usage

All the code that you care about is in one file: design_opt.py. It's nicely sectioned off into the following sections, denoted with the comment format # region SectionName:

  • Setup
  • Trajectory Optimization Variables
  • Design Optimization Variables
  • Atmosphere
  • Aerodynamics
  • Propulsion
  • Solar Power Systems
  • Gas Power Systems
  • Weights
  • Dynamics
  • Finalize Optimization Problem
  • Solve

If you're using the PyCharm IDE, you can instantly hop between these sections with the shortcut ctrl + alt + ..

In the first few dozen lines of design_opt.py, you'll see a bunch of flags that you can change. Feel free to play around with these flags and see how the design changes! (Beware, not all combinations of options work!)

If you want to do some kind of parameter sweep (e.g. how does the span of an optimized airplane change if I change battery_specific_energy), you can use design_opt_sweep.py as a template.

All of the other files in this repository are one-off experiments that I ran at someone's request and can be safely ignored (also, some of them are out of date). As you can probably tell, this isn't a clean, polished repository haha - nor is it meant to be, as we're rapid prototyping!

Troubleshooting

Follow this decision tree:

  • Before doing anything, try updating AeroSandbox by running this in your command line: pip install --upgrade aerosandbox
  • For all other problems, notify Peter.

Contact Info

If you run into issues, please use the following options, starting with the first and proceeding if I don't respond: DM me on the 16.82 Slack, open an issue ticket, email me at pds@mit.edu, send carrier pigeon.

Licence

MIT License, full terms here.