Generative Design at PIX
In April 2018, Autodesk released the ‘Autodesk Generative Design’ Software. It is now part of Fusion 360. Frustum is another company that has Generative Design software available as an online tool. Generative Design is an optimization technique based on Genetic Algorithm. Users define design variables, constraints and objective functions to setup an optimization problem.
Design optimization is a way to obtain best possible design given constraints and design goals.
These software can be used for structural topology optimization. The user sets up the problem on his desktop and the computations are performed on the cloud. The final obtained geometries can be so complex (and beautiful!) that they just cannot be manufactured even by most advanced machining methods. Additive manufacturing (3D printing) is the way.
Generative Design is now gaining attention widely in the industry. NASA Jet Propulsion Lab, General Motors and many other companies have realized the potential offered by the combination of generative design and additive manufacturing.
The challenge now is to bring together generative design and additive manufacturing technology. While powder-bed based 3D printing is capable of making such complex structures, it is by no means economical, thereby limiting its usage for prototypes or high-end manufactured goods. The challenge remains.
A true revolution in design will happen when additive manufacturing evolves to a point where these structures can be built quickly, cheaply and with reasonable accuracy. PIX Moving is working to solve this problem.
Here are a few parts designed at PIX Moving through Generative Design
The above two parts represent the upper and lower A-Arms of PIX car respectively. The design below has a chassis that was designed using Frustum at PIX, to explore the possibilities of applying generative design to automotive structures.
Let’s now take a part as a case, the upright or spindle of a car’s suspension system. The upright of a car holds the wheel hub and attaches to upper and lower control arms. It also serves as an attachment for the brake caliper and the steering arm. The primary loads acting on the upright are due to braking and steering. Additionally, the uprights on the driven wheels experience propulsive loads.
Here is the design of an upright made at at PIX. The material selected for this design is 7075 -T6, a high specific strength aluminum alloy with a yield strength of 450 MPa and Ultimate Tensile Strength of about 520 MPa. This upright weighs about 2.9 kg.
Autodesk Generative Design (GD) offers two ways to optimize this upright- one way is to use this upright as the starting geometry and then perform a topology optimization. Autodesk offers another novel way — Autodesk argues that a starting geometry is not necessary. If the loads, constraints and design goals of this upright are fully defined, then it’s possible to set up the optimization problem without a starting geometry. It is as if the computer is building a structure in air! This capability by Autodesk is truly generative and sets it apart from its competitors.
The figure below shows the problem setup using this novel method
Red parts represent obstacle geometries that prevent the computer from putting any material at these places. For instance, the left most cylindrical obstacle is actually space occupied by the brake caliper and the middle cylinder is space for the bearing and hub. Green parts represent preserve geometries that will be left untouched and preserved till the end of optimization. They also serve as the points for load application, represented by blue arrows. The accuracy of load magnitude and direction play vital role in topology optimization and care is taken to define them as accurately as possible.
Two separate manufacturing constraints are considered, additive manufacturing and CNC milling. An overhang angle of 55 degree and a minimum allowable thickness of 10mm are used to describe additive manufacturing constraint. Milling constraints include those needed for a 5-axis milling machine. The objective function is set as mass (minimization) with a factor of safety constraint of 1.7. Autodesk GD allows material selection of up to ten materials in a single optimization project. For the current problem AISI 1045 (500 QT), Aluminum alloy 7075 T6, Titanium alloy Ti6Al-4V and Steel 310 are considered. The optimization problem is now well defined and ready for computation.
The computation time for this problem was about 4 hours. Autodesk employs Amazon Cloud Services for computation. Results for every material and manufacturing constraint are displayed and the user can select the final solution based on his/her needs. Here are two results for the upright for material 7075 T6
The right upright is much lighter and can handle loads just as well as the left upright. However, it is not practical as the lower A-Arm cannot be attached at bottom of this upright, interference is observed. After studying all the possible solutions generated by the software, the left solution is chosen. Even though the weight savings are not as large as the design on right, it is highly feasible and does not involve interference with other car parts. The left upright offers close to 10 percent weight saving.
Here is the final assembly of the upright with our suspension system
Generative Design is a technique and it can be used for problems other than structural optimization. It offers Engineers, Industrial Designers and artists the freedom to design structures that are so often restricted by traditional design methods. At PIX we seek to exploit this freedom and change the way cars are designed and manufactured.
This article is contributed by Mechanical Engineer at PIX, Siddharth Suhas Pawar, who majored in Aerospace Engineering, Structures and Materials, graduated from University of Michigan.
