A recently published industry study analysed the standards and practises of virtual simulation work for many engineers today. Sponsored by Dassault Systèmes, the survey asked 329 engineers and designers to share their thoughts and experiences regarding the use of simulation throughout the design process, and how it’s applied to their respective industries.
Based on the findings, here are the 3 primary topics we’ll explore in this post:
According to the study, most respondents use their own local hardware and software for simulation and testing, where 50% use local workstations and the other half utilise local networks for data/file storage.
Housing local hardware appears to be a common setup for simulation work among engineers, however it’s making its way out of practise by the minute as alternative high performance cloud computing methods have proven to be more innovative, highly performative and more efficient for end users.
"High performance cloud computing enables users to spend more time designing and less time on IT and local hardware maintenance. Basically, high performance cloud computing allows users to have access to computational power that would be virtually impossible for them to have access to otherwise. The industry is still adapting to security issues and some compliances regarding data management, but HPC is the future, that’s for sure.”
- Niclas Jansson, Head of HPC at Ingrid Cloud
Almost half (49%) of respondents struggle with pre-processing such as the amount of time and effort required to set up the model. 24% of those users found it hard to maintain adequate resources for advanced simulations.
Other challenges were the time required to run the simulation (11%), and not having the right tools (10%).
Regardless of firm size, one of the biggest challenges that remains in simulation work begins at the pre-processing stage when it comes to the time and effort it takes to set up the model.
It is not unheard of that an experienced CFD-engineer can spend one to several weeks preparing a mesh, refining it after iterations and post-processing the result, before even starting to analyse the result of the simulation.
"In order to overcome these pre-processing barriers, automated simulations are a part of the solution. Pre-processing is a big pain for CFD-engineers, we've know that for a long time. Speed and time to result are two of our main goals and that’s why we propose an automated framework for fluid flow simulations."
- Rodrigo Vilela de Abreu, CTO at Ingrid Cloud
The report states: “Respondents identified three main barriers preventing engineers from running simulations more often: a lack of knowledge (47%), the expense (41%) and a lack of trust in the results (37%).”
"A significant competence threshold to performing simulations is still a common roadblock for many in the industry. That's why democratising CFD simulations is so important. Finding new ways of sharing CFD knowledge beyond the CFD experts increases the value output of simulation work and allows data-driven design decisions to be made collaboratively and across multidisciplinary industries. Automated simulations are a huge game changer!"
"When it comes to expense, it's a common misbelief that all accurate CFD technology is expensive. When you compare it to the alternatives of hiring wind consultants and physical wind tunnel testing, virtual simulation can be very cost effective. The reality is that computational advancements and resources have come a long way, and thanks to that, simulation work has become more affordable than ever. In benchmark studies, Ingrid Cloud has repeatedly shown an 84% cost reduction compared to a large competitor."
- Sebastian Desand, CEO at Ingrid Cloud
"The issue with 'lack of trust in the results' is an understandable one when it comes to simulation tools. Many new technologies are emerging to the market and yet, CFD is a complex interdisciplinary task. Not everyone can offer a reliable tool. In order to trust the results, you need a verified code. Ingrid Cloud’s method is constantly submitted to an extensive number of tests, research and benchmark cases."
- Prof. Johan Hoffman, Head of R&D at Ingrid Cloud
On average, respondents went through 19 design/virtual simulation cycle iterations before producing their first prototype. 12% of respondents were not satisfied with that number and reported 12 as the target to aim for in the future.
When asked what specific improvements can be made to enhance the design/simulation cycle:
“The majority (51%) would enable designers to run analytics themselves much earlier in the design phase; they saw this as a strategy to lower the number of cycles.”
Introducing simulations earlier on in the design process enables engineers and designers to bring their best work forward by significantly reducing or even eliminating the need for rework or redesign at later stages- which can be become a more complicated task once a design has gone through multiple rounds of iteration cycles, become more complex and detailed in nature or has received project approval.
"As shown in the report, early-stage simulations and data analytics can decrease the amount of design/iteration cycles, making the design cycle faster and inherently reducing time to market. That's what we see here with most of our clients, high-fidelity simulations performed early in the design stage produce valuable insights."
"24% of respondents would use best practices to encourage more collaboration between designers and analysts.”
"Simulations generate massive amounts of data. As design cycles and large-scale projects often require a high level of collaboration, not only is it evident that the data must be presented in an informative manner, but it must be intuitive, practical and easy to understand. Here at Ingrid Cloud we work constantly with that in mind."
The full report is available here: https://www.engineering.com/