Accelerate Your Product Design

Design products faster by running HPC workloads like FEA on the Cloud.

AWS has submitted this post.

Written by: AWS 

(Image courtesy of AWS.)

(Image courtesy of AWS.)

Utilizing modern technology to enhance product design through finite element analysis (FEA) is nothing new, as seen in the automotive industry. FEA is a dependable testing method that helps engineers analyze the vehicle’s structural components to check how they relate to the overall design and the outer environment. 

The benefits of running FEA simulations can be found across many industries, and include shortening the product design process, reducing the amount of physical testing and saving costs.

The challenge of running FEA is that it requires quite a few moving pieces to fall into the right place. To elaborate, running the vibration, acoustics and stress analyses requires high-performance computing (HPC), purpose-built tools and time to produce accurate and valuable insights. 

Organizations can choose to run their HPC simulations on-premises. However, the modern climate has made us eager to get results as fast as possible, and on-premises infrastructure comes with inherent restrictions, resulting in longer wait times for results.

While having the ability to do larger and more simulations is a shared goal among engineers and scientists, the cost of HPC in the cloud can get in the way of achieving them.

Roadblocks in the Automotive Production Process

Building infrastructure that can support FEA workloads can feel like an uphill battle. Large simulations, such as FEA, use multiple compute instances as they must work in parallel, which means scalability and high-speed interconnectivity are essential.

Meanwhile, scientists and engineers face extensive pressure to complete their most intensive FEA workloads faster. Any reduction in the time needed to test the product equals faster time to market.

The stakes are high, especially with the rise of electric vehicles. Now, engineers are remodeling how modern cars are being built, so the analysis requirements are growing due to the different positioning of components. Innovation with new electric vehicle body styles is on the rise, while testing for driver’s safety remains imperative. 

(Image courtesy of AWS.)

(Image courtesy of AWS.)

For example, the manufacturing process of a car requires vibration, acoustics and stress analysis to ensure a good consumer experience. During the product design process, numerous car body simulations are done with this type of linear dynamics vibration analysis to eliminate unpleasant vibrations for a better driving experience. More complex simulation requirements include testing the vehicle’s ability to absorb crash impact or its performance in the event of multiple car crashes.

To produce a sustainable amount of analysis, FEA workloads need complex infrastructure support to deliver accurate insight. The infrastructure needs to be purpose-built to satisfy the needs for scalability, accuracy, performance and cost-effectiveness.

Luckily, there is a way of reducing the time and cost of running FEA analysis while increasing accuracy. 

Speed, Scale, Accuracy and Cost-effectiveness are Possible

Cloud computational services, such as the Amazon EC2 Hpc6id instances from AWS powered by 3rd Gen Intel Xeon Scalable processors, deliver the infrastructure capacity needed for scalability. This is especially true when it includes a high-performance file system and high-throughput networking for faster inter-node communication, ensuring FEA jobs can run in parallel while communicating effectively. 

Additionally, if the infrastructure is easily scalable, it enables users to procure capacity when needed, and turn it off when they’re done, optimizing their HPC costs through pay-as-you-go services.

With the ability to do more and more extensive simulations, engineers can complete their workloads in a shorter amount of time. The result will lead to fewer physical prototypes and experiments, a shorter design development phase and increased product quality at a lower production cost.

Learn more and get started at Amazon EC2 Hpc6id Instances – Compute – Amazon Web Services.