The nature of engineering is complicated, and engineers need every bit of help they can get to make sense of their decisions. In the end, if bad decisions are made, everything is on the line. As a result, the engineering community has worked to standardise engineering processes.

Traditionally, the engineering processes include design, prototyping, testing, and iterating. The development hierarchy still dominates the engineering community, yet we see rapid advancement in the engineering ecosystem. Tools that were formerly an addition to the engineering process have now become the norm. This includes the integration of computer-aided design (CAD) in the conceptualisation stage, the introduction of additive manufacturing for rapid prototyping, and the adoption of computer-aided engineering (CAE) for design tests and iteration. Together, these innovations have revolutionised how we engineer the world around us. In this blog, we discussed the concepts that are expected to become the standard for the future engineering ecosystem.

3 Layers of Knowledge in the Engineering Ecosystem πŸš€

To make engineering better and easier, and get rid of all unnecessary restrictions and limitations within the development process, we’re focusing on developing expertise in three key areas, says Pierre Sabrowski, co-founder and CEO of Dive a particle-based simulation platform, in an interview with Jousef Murad on the latest engineering mind podcast episode. The three expertise areas that Pierre talked about are first, understanding the numerical basics. Second, validating these, like a proof of concept. Finally, it’s understanding the specific needs of industrial applications and if our model can solve real industrial problems.

In the first area, Dive invests a lot into the R&D efforts, which leads into the second, a robust and reliable platform. But they also ensure that the platform is put to the best possible use and solves their customer’s concrete needs and pains. That’s why they offer customer success engineers as a continuous service, who ensure their customers maximize their success with the platform.

To see the future of the engineering ecosystem, we need to be capable of participating in all three stages. The rapid evolution of the ecosystem necessitates parallelising the creation of knowledge and value in all knowledge stages. And it requires figuring out how this can be interconnected. However, the industrial stage players, which primarily include the ecosystem's user or customer, are not expected to play across the entire pyramid. In this case, it falls on the shoulders of the developer and the providers to help out with the knowledge through customer success management and leadership.

The Role of the Knowledge Pyramid in the Democratisation of Simulation πŸ”Ί

Engineering simulation is known to be difficult and that it should only be performed by experts in specific areas of simulation. This has become a big obstacle to the wide adoption of CAE in the engineering community. To overcome the challenge, Pierre suggests focusing on making sure people are capable of dealing with the industrial application side of the ecosystem and making sure the user doesn’t overestimate what they can do. Still, the need for knowledge on other layers is inevitable, and it is essential to provide that information in structural conditions.

The structure of the provided information has a significant effect on knowledge dissemination. Instead of provisioning the knowledge, including the physics and the numerical, in bulk, providing it in digestible pieces is preferable. Also, tailoring the information to the particular aim of the customer project and providing it promptly as needed in a seamless collaboration environment can have a significant positive effect. In the end, this effort should lead to the development of a workflow and template that are validated by simulation experts for designers to utilise. At the same time, the required software and hardware infrastructure need to be democratised.

The Role of Cloud Infrastructure πŸ’»

I can't work on my workstation while running a simulation; I need to upgrade the hardware; I need to install the latest update; I can't handle bigger problems, etc. are familiar phrases in the simulation ecosystem. The introduction of cloud-native and even browser-based simulation platforms is making these problems irrelevant. We see a cloud infrastructure that makes both software and hardware with high-performance computing capacity accessible on demand. The cloud infrastructure evolves at a faster rate, allowing for large numbers of parallel numerical computations and more benefits. The penetration of cloud-native CAE platforms has enabled a real-time collection of experts across all layers of the simulation knowledge ecosystem. Such features will play a big role in the democratisation of knowledge.

With the current trend, it is inevitable to see the engineering ecosystem shift to the cloud in the coming years. It is essential to realise that "if you are still trying to work on premises, nobody is developing for you," says Pierre. But this argument should not be used to push customers into a discomfort zone. Customers should get the value of their investment in their on-premises HPC system.

The other most promising subject, along with the development of the cloud engineering environment, is data management. What we see is that the cloud ecosystem of engineering tools is helping to standardise data and avoid individualisation. This will ease data accessibility, data handling, and post-processing of the data. This will open the door for a standardised data structure for deep learning purposes. Due to legitimate fear factors, the engineering community initially thought it was unwise to put a project in cloud-based systems. However, the due diligence taken by providers on their cloud infrastructure, such as transparently outlining their data security policies, has helped accelerate its adoption.

The Role of AI in the Engineering Ecosystem ☁️

In the wake of ChatGPT, everyone wonders what the role of AI is in their respective fields. Generally, in engineering, we have seen AI assist in smart production lines and complex manufacturing tasks. Major tasks in the engineering workflow, such as design and simulation, can also benefit from artificial intelligence assistance.

AI requires substantial data in the point cloud dimension for training, and each simulation creates a point cloud that can be input for the wider dimension. By filling the mighty size cloud dimension with all available simulation data, whether it is a 2D or 3D simulation, experimental data, observation, etc., we can create a substantial cloud of knowledge. Clustering this point for analysis and applying machine learning will help create input and output machines for future reference. However, the data structuring efforts have been hampered by the non-standardised and individualised simulation data management systems of the conventional on-premise CAE tools.

As mentioned earlier, the data handling problem is shown to be improved with the adoption of cloud HPC infrastructure. Hence, the idea of applying data analytics and, from there, also machine learning, is becoming very tangible. Yet, how far we will actually come to have a whole design built with machine learning is not known. But once enough data is collected structurally, Pierre is optimistic about seeing the role of AI in the engineering ecosystem advance in the near future.

You Pay Innovation with Money, not With Time ⏰

In conventional simulation solutions, time is a major factor. Faster solutions mean more material is completed. However, with cloud computing solutions, large amounts of work can be done in a shorter time, and nobody wants to do things in series nowadays. Hence, cloud computing has made time irrelevant, as engineers can run hundreds of simulations in parallel. Because of that, the cost has become the only sticking point when deciding to move to the cloud. The cloud will help engineers drive innovation based on money because time is not a factor anymore. As the ecosystem's efficiency grows, so does its use. Thus, the cost can be offset by the higher value it provides to drive growth. What the user should focus on is the return on investment, and if the ROI is good, you are fine.

If you’d like to see more podcasts and blogs on simulation, consider subscribing to my blog or YouTube - and please feel free to leave a comment down below! πŸ™‚

Keep engineering your mind! ❀️