AI Redesign of Martian Propeller for the Ingenuity Probe
This design targets the operating conditions of the NASA Ingenuity probe. Using the design problem as a challenge, it will test the propeller design AI, to see if can produce a high performance result.
This a thrust targeting, power reduction optimization, requiring 8.9 Newtons of thrust. The flight conditions are at ground level on Mars, 0.017 [kg/m^3] etc.
This video is an unveiling of the new Caladrius AI, brought to you by 3D Printair Aerospace.
The system is comprised of a new and groundbreaking Aeronautical Engineering process developed completely in house, to be able to allow high bandwidth communication of the the most functionally complex aeronautical designs. Caladrius can easily handle thousands of discrete blueprints at once, handing them effortlessly back and forth between an intelligent optimizer and a test bench proven, simulation workflow.
The progenitors of this AI, designed originally to simply relieve the monotony of meticulous aerospace engineering, were not able to efficiently find solutions in such a large state space. Caladrius adds the function of real time learning, along with the most organized aeronautical data libraries in the world to make the algorithm potent and effective.
The seamless integration of these key features allows for an unprecedented amount of precision, providing things such as real time representative trailing edge thicknesses, section distinct Mach and Reynolds numbers, and exact structural details, just to name a few.
The end result is a system capable of creating a solution based on ALL of its inputs, designing a blade based on its actual lift distribution, instead of its separate pitch and airfoil components. Whereas traditionally, optimizations occurred separately, and final designs where biased towards the most effective lift coefficients of a specific airfoil, Caladrius can weigh the cost benefit of creating more lift less effectively, at a more ideal locations. Caladrius becomes even more effective when designing for multiple flight conditions, especially in blades with no collective pitch.
Compared to CFD, this a true design process, not simply an analysis tool for a single design. The learning ability of the AI presents a window into the future of Aeronautical design, where the design process begins to leave the realm of physics for the first time, to create instant artificial solutions to complex real world design requirements.
When coupled with CFD, this design process can provide an excellent starting point for the design itself. The CFD can be used as a second point of validation, and this design process can supply geometries to be evaluated.
The preferred method of testing is always real life testing. For cases like this, it can be hard to find the right conditions to test without flying to 90,000 feet. But for many other cases, mostly involving UAV technology, the Tyto Robotics series of tests stands is our prefered method. It is an especially easy choice with today’s high tech prototyping tools. The design of this tool would not have been possible with their testing equipment.