Simulation-driven design is a process where decisions on a system’s behavior, performance and design are greatly supported by the results of relevant simulations of said system. It is a well-known principle and widely used across all industries, with best-known examples in aerospace and automotive. The types of simulation we associate with those sectors are elaborate finite element analyses and complex model-based system architectures. In this perspective, we would like to shed some light on a less-known implementation of simulation-driven design, which is situated at the very beginning of any development.
Contrary to aerospace & automotive sectors which use conservative waterfall and V-models, Verhaert employs a development methodology called “RICE”. It has been developed and tuned in-house for almost 4 decades now and aims to support projects with many more uncertainties than encountered in previously mentioned sectors. While not going into too much depth, one of the axioms of RICE is “accelerating learning”, which enables us to identify and de-risk critical development items from the very beginning. Many techniques are employed in order to accomplish this target: pretotyping & mock-ups, system design, morph charts & trade-offs, breadboarding and of course simulation.
Simulation can be performed in many ways, across all 4 dimensions and comprehending multiple domains. 1D simulations are a specific subset that is particularly well-suited for the support of system design. They make abstraction of component geometries (as opposed to 2D, 3D and 4D simulations) and focus on the interaction between components (or lumped elements). For these reasons, we love 1D simulations:
- They are easy to set up: depending on the used tools and system complexity, 1D models can be set up within minutes or hours.
- They are computationally inexpensive: this too depends on many factors, though 1D simulations are orders of magnitude faster than e.g. 3D/4D FEAs or CFDs.
- They easily support multi-physics models. Almost any system operates in more than one relevant physical domain: mechanical, thermal, electrical, pneumatic/hydraulic, computing/control, … Incorporating additional domains in your model is easy with the availability of adequate libraries in the simulation environment.
- They’re inherently easy to understand for both engineers and laypersons. As the model consists of nodes (representing physical components and/or phenomena) and connections (representing the flow of energy, mass or information), complex systems can be represented in a comprehensible graphical way and allow for easy monitoring, troubleshooting and tweaking.
Why you should employ (1D) simulation from the very start of your development
While 3D and 4D FEAs & CFD certainly yield lots of information and have their place in simulation-driven design, it is oftentimes not economical to employ these methods during very early development: the amount of work required for setting up and running such simulations is high while its usefulness is debatable as the number of unknowns is at its peak.
This is why we turn to 1D simulations: they are the ultimate tool for system design and we’ll show how they accelerate developments right from the first conceptual phase below:
- Think it through: while building the model, you need to think through each node or building block – its function and relation with the other building blocks. If the model is missing crucial components, interactions or boundary conditions, this will be noticeable in the simulated behavior of the system. Hence 1D simulation provides an excellent sanity check on the architecture you are envisioning and helps uncover unknowns very early in the design process.
Download the perspective to continue reading on how simulation can accelerate your development from the very start.