Knowledge Base

Generative CAD Design

Introduction to CAD

Product design is an inherently time consuming process. Let’s take an example from the automobile sector. Car wheels spin at different speeds, especially when turning. A differential gear system allows wheels of a vehicle to rotate at different speeds, even though the torque transmitted to the wheels remains the same. The main parts of an automobile differential comprise the gear system, the housing, the axle shaft and differential pinions. There are a huge number of design possibilities, not to mention the choice of material and manufacturing methods. Traditional designing methodology would take a lot of time to come up with a differential system that works properly, makes best possible use of materials, is lightweight, and is easy to manufacture. CAD (Computer Aided Design) is the use of computer software to design products. Since they were commercially used from more than 50 years back, CAD has changed the way engineering design is done. CAD software such as Creo has proved to be beneficial in diverse industries today, ranging from aerospace, automobile, electrical, mechanical to industrial machinery, medical devices, and to heavy engineering. Today, it is impossible to imagine that commercial engineering drawings were once drawn by hand, and depended upon the skill of the draftsman for accuracy. The growing number of CAD companies that offer CAD software and engineering service companies that offer CAD drawing services is a testimony to their popularity.

Why Generative Design?
Manufacturing companies always seek faster and reliable methods to reduce lead time to market. Customer demands and competition always keep manufacturing companies on the edge. If they don't innovate, they would lose their competitive edge pretty fast. While CAD is helpful, it produces designs only as per inputs given by the designer. If the design concepts developed by designers and engineers are not practical to implement, the manufacturing company stands to lose its competitive edge. Redesigning and tweaking concepts into a design, which meets the functional requirements as well, results in an unnecessary waste of time and money. This is where the concept of generative design comes into play.

Generative design allows software to generate a multitude of different designs using a viable design space and criteria defined by the user. Designers can specify geometry, forces and fixtures as well as maximum costs, manufacturing process and various other parameters. The system then calculates and presents various designs that meet the user’s requirements by applying certain algorithms. Generative design grants designers and engineers the ability to be more exploratory with their designs, and allows generation of multiple designs in a short period of time.

Generative Design – under the Hood
Generative design is a design exploration process that uses artificial intelligence (AI) to create a wide range of solutions and ideas for complex problems like reducing component weight, or optimizing performance, and making designs simpler. AI is a branch of deep learning that uses neural networks, touted to loosely mimic the function of human brain. Deep learning algorithms explore and discover latent patterns and relationships in large amounts of data, something which is not very easily possible for humans. AI and machine learning (ML) technologies incorporate unsupervised learning, where algorithms work toward detecting patterns in unorganized data. Generative design takes this concept a step further by creating algorithms that attempt to generate new samples given some input parameters. Engineers can leverage this autonomous use of AI to create optimal designs from a set of given system design requirements. They can interactively specify their requirements and goals - including preferred materials and manufacturing processes - and the generative engine will automatically produce a manufacture-ready design as a starting point or as a final solution. As a result, engineers can interact with the technology to create superior designs and drive product innovation more quickly.

Under the hood, generative design uses a mathematical optimization model in order to solve an engineering problem. It also utilizes the power of finite element analysis (FEA) in order to create optimal topology.  Generative design makes it possible to explore hitherto unexplored options – something that humans cannot design easily. Complex geometries and artistic shapes are all possible with the help of generative CAD design; all the while maintaining various constraints like weight, type of material, dimensions, and other engineering parameters.

So, in essence, the concept of generative design for CAD is simple – it leverages the capability of technological advances to autonomously generate a set of design alternatives within the constraints defined by the design engineer. Since it is autonomous, the designer or the engineer need not interfere in the process; the CAD system generates a number of alternative options. This reduces the workload of the design team, increasing their output efficiency. Once the generative design process is complete, the engineers can choose among the best option that satisfies both aesthetic as well as functional requirement of the object.

Typical generative design passes through these four stages:

  • Designers enter design goals, constraints and other related parameters in the generative design enabled CAD Software
  • The CAD program generates a range of design options that have been directly validated for manufacturing
  • The design team selects the best possible solution, refines it further and sends it to the production team
  • The product designer / engineer chooses the most suitable manufacturing method for production and proceeds further

Generative design in combination with 3D printing (additive manufacturing) is a very potent combination. The ability to generate thousands of designs adherent to restrictions early on in the design process is a time-saver. Furthermore, algorithms are not biased towards pre-conceived concepts and the untraditional results are easier and cheaper to manufacture with 3D printers.

Generative design has been successfully used in the aerospace, automobile and construction field. Modern CAD software like Creo also allows users to get additional assistance with modules like Generative Topology Optimization and the Generative Design Extension. Used together, these AI-driven generative design tools can help users deliver more innovative, differentiated products, reduce the time to market, and reduce overall product costs.

Summary
CAD is here to stay. Improvement in CAD software and adoption of deep learning technologies has enabled industry leaders to augment their CAD offerings. Generative design is one such facet of modern CAD software. Generative design proactively participates in the design process. This potential of generative design can be exploited by combining it with additive manufacturing (3D printing) to optimize designs faster. Used properly, generative design shortens the design cycles, and coupled with 3D printing can disrupt the whole engineering and digitization product lifecycle.