Vox Mechanica roughly translates to Voice of the Machine and that counts double for what I'm planning to do with it and how it works. I'm developing an application that generates components for internal combustion engines, making use of physics, mathematics, FEA, CFD and a whole bunch more of fancy calculators and algorithms.

It takes user input, like cylinder capacity, compression, fuel type and engine aspiration (forced induction or natural), among others. It then takes these values and runs them through a Computational Engineering Model, specifically designed for engine design, but made free of heuristics, assumptions, work-arounds and guesses. No rule-of-thumb physics, but verified calculations to stay "closest to the truth". 

What it puts out is a tale of millions of calculations, generating, shaping and finetuning a part that is the epitome of mechanical efficiency and a prime example of biomimicry. A component designed by an algorithm that knows everything about every single little feature, and has a reason for every surface and angle you can see. 

Engineering without CAD

I love CAD. It's an easy tool for most engineers to put idea's on paper/screen and have a visual of what you're working on. However it is often not the most friendly environment for complex geometry generation. Programming is the almost direct opposite; it is abstract and does not require the engineer to have a sense of spatial awareness. It does however provide ample room for very complex calculations, performed at lighting speed.
During my hobby of engine design and mechanics I decided I needed something more than CAD if I truly wanted to create something special. Programming has provided me the solution; engineering within an IDE with the help of AI and some basic programming courses.

Inspiration and Research

The single largest trigger for me to actually start developing the application came from Leap71 with their PicoGK library. One of their founders took the time and effort to write a course specifically for engineers on how to bend your mind from geometry to abstract function and help each engineer start a coding journey of its own. I'll link it below this post. This forced two things; me dusting of several books for actual physics calculations and learning how to read code. 

Leap71 and their CEM 'Noyron' is definitely a prime example where programmatic engineering outperforms CAD. 

I mainly use AI to write the code for me now. Development is much faster, but I always check its input and output. In the past few months I've learned how to properly prompt AI, build skill sets and harnesses, and much more. Without the AI I would've never build this much understanding of the physics involved, nor would I have found them this easily.

In future posts I will keep track of the progress, challenges and more, with some in between results. However, the first few posts will be a log of what I've already done so far.

The link to the course:
https://picogk.org/coding-for-engineers/