What is Computer-Aided Engineering (CAE)?

Computer-aided engineering (CAE) is a subdiscipline of engineering that uses software tools to digitally simulate and optimize product designs. CAE is used across industries to aid in failure analysis, improve performance, decrease development costs, shorten design cycles, and give engineers insight into their product’s performance.

Initially developed in the aerospace industry, CAE tools have evolved to become an integral part of the engineering process, often driving product development decisions well before physical prototypes are available. The ability to virtually test real-world scenarios, answer critical design questions, and validate features and performance early in the design process makes an investment in CAE tools and skills one of the strongest returns on investment in engineering.

CAE tools build a mathematical representation of how a given set of geometry, materials, connections, and constraints behave under applied loads. The goal of any CAE simulation is to define known system quantities and calculate unknowns. It is used instead of physical testing when possible or to reduce iterations when real-world testing is required.

Common Terms Used in CAE

Before diving into the three-step process engineers use for CAE, let’s review some common terms.

• Load: An external force or energy change that acts on the digital model.
• Constraint: A characteristic value in the model, using a degree of freedom, that is fixed by the user.
• Boundary condition: Loads and constraints applied to the model.
• Connections: A definition of how physical values are transferred between each object in the model. These are usually loads or constraints, such as pressure or displacement.
• Material property: A definition of how a given material behaves when boundary conditions are applied to it. Properties like stiffness, conductivity, and density are material properties.
• Model: A mathematical representation of the known properties and how they respond to applied loads.
• Mesh and meshing: Many CAE tools partition the model domain into discrete geometric areas or volumes, called cells or elements. The mathematical definition of those elements and their connectivity is a mesh, and meshing is the process of discretizing.
• Solver: Refers to the algorithm used to solve for unknowns, given a set of knowns.
• Results: A numerical representation of the solved-for unknowns.

Regardless of where CAE is applied, practitioners typically follow three steps: pre-processing, solving, and post-processing. The complexity of each step varies with the physics captured by the simulation, required level of accuracy, complexity of the product, and complexity of the operating environment captured in the digital mock-up. In addition, the process is always driven by knowing what information the engineering team is looking to gather from the simulation.

Below are the standard three steps.

Pre-Processing: Describe the Problem

Pre-processing is the first and most important step in the CAE workflow. This is where engineers document known values, use CAE software to discretize the geometry, and capture all required data in a database. Most model building begins with taking geometry or the components in the system and discretizing, or meshing, the geometry. The user must then apply constraints that define specific physical values and define the loads acting on the geometry. They must also specify the properties of each material used, connections between components, and how boundary conditions change over time. The final task for engineers conducting a CAE simulation is to specify how they want the mathematical problem solved by specifying the inputs and variables the solver needs to do its job.

Pre-processing is important because the solver computes the problem's outputs, and if the inputs are incorrect, the outputs will not reflect the real-world situation. A classic garbage-in, garbage-out (GIGO) situation. It is also important to know that automation and tight connections to computer-aided design (CAD) geometry tools speed up and increase the accuracy of pre-processing.

Solving: Build a Mathematical Model and Solve It

The actual processing of the mathematical representation built by the software during pre-processing is called solving. First, the solver converts the mathematical definition into a set of equations, usually partial differential equations, with known and unknown values. Then, numerical methods are used to solve a large set of equations for unknown values. Although software is used to solve for the unknowns, some algorithms can consume large amounts of memory, disk space, and CPU cycles. Access to efficient solver methods and high-performance computing resources is important when solving many CAE models.

Post-Processing: Look at the Solution

The results from the solving step are stored as numbers in a database. To leverage those values, engineers need to use their CAE software to convert them into useful representations. Some of the more common result representations created during post-processing are:

• Graphs
• Geometry with result values represented as colors
• Results in fields as surfaces at a given value
• Motion of fluids as streamlines, field values as arrows, or light paths as rays
• Distorted geometry
• Tables
• Computer files with result values and the coordinate location of those values

Plots of geometry with values represented as colors are the most common type of output created in post-processing. In most cases, engineers use post-processing to provide information to verify and validate designs or aid in decision-making in the design or manufacturing process.

An important part of CAE is modifying the model and re-solving it to assess how those changes affect the results. This can be done by the engineer following a manual workflow or by automating loops, referred to as optimization, that parametrically vary the inputs using an algorithm that converges to the desired output values. Most modern CAE software includes scripting capabilities that give engineers the power to automate and control these iterations. Increasingly, this is done with a Python API, as with the PyAnsys pythonic access tool for Ansys software across the Ansys family of CAE tools.

Technically, CAE is any type of simulation where the use of computers plays a role in calculating the behavior of a product. Engineers can classify their simulations by the type of physics they are solving or by the type of solver they use.

CAE Types by Physics

• Structural and Mechanical:    
  • Structural analysis: Deflection and deformation of materials under static loading.
  • Dynamic analysis: Deflection of materials under loads that vary over time and the natural frequency response. Also includes acoustic simulation.
  • Explicit dynamics: Deflection and deformation over short durations of time.
  • Multibody dynamics (MBD): Displacement of bodies in a system relative to one another, modeling kinematics if momentum is not considered.
• Thermal and Fluids
  • Thermodynamics: Heat generation and energy transfer in a system.
  • Combustion and molecular dynamics: Heat generation from chemical reactions.
  • Heat transfer or thermal analysis: Heat movement through a system due to convection, conduction, and radiation.
  • Computational fluid dynamics (CFD): Fluid flow in a system, solving the Navier-Stokes equations.
• Electronics and Electromagnetics:
  • Circuit simulation: Current flow and voltage values in electrical systems.
  • Logical and RTL simulation: Digital signals in circuits.
  • Low-frequency electromagnetics: Electromagnetic fields where the wavelength is much larger than the object.
  • High-frequency electromagnetics: Electromagnetic fields where the wavelength is smaller than the object.
  • Signal integrity simulation: Signals interacting between different circuits.
• Other:
  • System simulation (1-D): A “block” representation of the system where each block represents a component in the system.
  • Process simulation: 1D simulation of any type of process, from chemical processing to the supply chain.
  • Technology CAD: Physical operation of transistors in an integrated circuit (IC).
  • Optical simulation: Light movement, including refraction, reflection, radiation, and absorption.
  • Multiphysics: Any two physics combined in one simulation. The most common are fluid-structure interaction (FSI) and electro-thermal.

CAE Types by Solver

Different physics types can be solved with multiple solver types. For example, heat transfer simulations can use finite element, finite difference, or finite volume solvers to compute heat flow.

Here are the most common types of solvers that engineers refer to when explaining the CAE approach they are using:

• Finite element analysis (FEA): Calculates in elements using a discretized representation of the geometry using linear algebra. FEA is the most common solver approach in CAE and is used primarily for structural, electromagnetic, and thermal analysis.
• Finite difference analysis (FDM): Calculates unknown values at discrete points in a model. Often used for system, circuit, and process modeling, it is used for simplified representations and can sometimes be solved in real time.
• Finite volume analysis (FVM): Calculates unknown values across small volumes, conserving energy or mass. This is the most common solver type of CFD.
• Boundary element analysis (BEM): Solves for fields using a discretized representation of the surface of the objects in the system.
• Discrete element method (DEM): Calculates the position and momentum of individual particles. Used to model particle behavior for bulk materials like rocks, grains, or powders.
• Smoothed particle hydrodynamics (SPH): Solves for unknowns in fluids using blobs instead of volumes. Effective for sloshing and waves.
• Ray tracing: Calculates the velocity, energy, reflection, refraction, and absorption of light and sound.

The effective use of CAE solutions is enhanced by leveraging computers to support other steps across the entire product life cycle, including the product development process, manufacturing, and maintenance. Not only does each area benefit in many of the same ways through a digital engineering approach, but data can more easily flow to and from CAE tools.

The most common forms of computer-aided domains are:

• Computer-aided design (CAD): The use of software for product development began with the shift from 2D physical drawings to CAD drawings and later to solid models. The use of CAD has been a de facto standard for decades. Almost all CAE tools support tight coupling with CAD tools to exchange geometry and dimension parameters.
• Electronic computer-aided design (ECAD): Similar to CAD in civil and mechanical engineering, ECAD is a set of computer software tools that aid product development, focusing on the configuration of electrical systems. ECAD is used for product definition, ranging from the transistors in an integrated circuit to the layout of a multilayer printed circuit board (PCB). The term electronic design automation (EDA) refers to tools that aid engineers in the design, simulation, verification, and manufacturing of PCBs and integrated circuits.
• Computer-aided manufacturing (CAM): CAD is also coupled to tools to aid in the planning and execution of manufacturing using CAM software. CAM tools can be used to do everything from programming computer numerical control (CNC) machining systems to robots and automated assembly lines.
• Computer-aided process planning (CAPP): Manufacturing and process engineers also use CAPP tools to plan and optimize processes, sometimes integrating CAE tools, for virtual prototyping of options.

The only constant in the CAE world has been consistent improvements in capability and speed, taking advantage of improvements in computer hardware, numerical methods, and user interface design. This trend continues with the same focus on improving accuracy, ease of use, and solver performance.

Here are some advances that should provide improvements in the coming years.

Generative Artificial Intelligence (AI)

CAE software has used AI, particularly machine learning (ML) and expert systems, for decades. Current research is around neural networks and incorporating large language models into the user experience and solvers. Users can already leverage this technology in the user interface with tools like Ansys Engineering CoPilot during pre- and post-processing across a wide range of Ansys tools and on the solver side with the Ansys SimAI AI platform for simulation, the Ansys GeomAI AI platform for geometry, and Ansys TwinAI AI-powered digital twin software.

Graphical Processing Units (GPUs)

CAE solvers benefit from the same massive parallelization of vector operations used in graphics processing and generative AI model training. Programmers can streamline linear-algebra algorithms, enabling larger models and faster solves.

Integrated Circuit Miniaturization and Improved Power Management

The same is true for improvements in CPUs, GPUs, and memory ICs. Smaller feature sizes enable more transistors and higher clock speeds, which CAE tools benefit from. It is worth noting that CAE tools are critical to the design of these improved chips.

The aerospace industry is where CAE tools first started adding value, but their use has grown across industries. Here are some industries where CAE workflows have become part of the design workflow:

• Aerospace and defense (A&D): CAE has been an integral part of aerospace and defense product development and life cycle management since the 1970s, and it remains a key component of workflows. Usage is driven by the industry’s high costs, low volume, difficult-to-test environments, and significant safety concerns.
• Automotive: Over time, the automotive industry has adopted CAE to the same extent as A&D. Use is driven by equally significant safety concerns, as well as by the increased complexity of automobiles, significant competitive forces, and cost pressures.
• The Build Environment: Engineers building large structures for civil engineering applications also benefit greatly from applying CAE throughout the process. CAE has always been a foundation in structural engineering, but usage has grown with system simulation of buildings, highways, sewage systems, and more. The industry was an early adopter of digital twins.
• Electronics: The modern integrated circuit could not be designed without the use of CAE tools to drive and validate designs. The cooling, packaging, and manufacture of complex electronics systems also rely heavily on engineers using CAE software for design and optimization.
• Power Generation: Another group of early adopters of CAE are the engineers who design, build, and maintain power generation equipment and plants. From steam turbines to modular nuclear reactors, almost every component and system is modeled and tested virtually.
• Mining, oil, and gas exploration and extraction: With high costs and high returns, engineers engaged in finding and extracting natural resources use simulation tools to find deposits and then design the machines and systems used to extract the resource.

The value of CAE software tools is exemplified by the overwhelming number of choices engineers have. In deciding which tools to use, teams should consider the following:

• Does it provide the information we need for the products we work on?
• Does it interact seamlessly and with bidirectional parametrics with my CAD?
• Is it easy to learn and use?
• Does it have useful online documentation and support, along with knowledgeable human support when needed?
• Are there a large number of users we can hire or contract?
• Are the solvers modern and efficient, and do they take advantage of the latest hardware technologies?
• Is the development team working on new and improved features all the time?
• Does it support multiphysics?
• Does it have robust post-processing that quickly transforms solution results into a usable form?
• Can our team easily customize the pre-processor or automate tasks?
• Are optimization and parametric studies built in from the ground up?

The Ansys, part of Synopsys, suite of tools answers all these questions affirmatively.

Taking a closer look at the flagship products helps show how far CAE tools have come and the value they deliver to engineering teams.

CAE in 3D Design

Ansys Discovery 3D product simulation software is an industry-leading CAE tool for design teams that works hand in hand with CAD. In a single intuitive interface, it offers geometry modeling and modification; structural, thermal, and fluid analyses; and optimization. It is also a notable example of how to incorporate GPUs into advanced solvers to deliver near-real-time results and, most recently, AI tools to guide non-CAE experts through the simulation workflow. Once engineers are done simulating in the 3D design space, they can transfer their simulation to detailed physics and multiphysics solves. For example, users of Discovery software can transition to flagship products, such as Ansys Mechanical structural FEA software, Ansys Fluent fluid simulation software, and Ansys HFSS high-frequency electromagnetic simulation software.

A CFD in Ansys Discovery 3D product simulation software that shows near real-time solving of a ventilation system

Structural, Dynamic, and Thermal

Engineers around the world count on Mechanical software as their go-to FEA powerhouse. Although primarily focused on simulating structural, vibration, and thermal situations, it also supports acoustics, voltage, fracture mechanics, and many other physics. Simulation can include non-linearities and time dependency, all in an open, scriptable platform with built-in parametrics and optimization.

Fluids Simulation

Fluent software is a prime representative of a comprehensive and robust CFD platform used to model fluid dynamics across industries. Because solving CFD problems tends to be mathematically demanding, it is also an example of leveraging high-performance computing (HPC) and GPUs to support larger, more accurate models.

High-Frequency Electromagnetics

The industry gold standard for high-frequency electromagnetic CAE is HFSS software. From PCBs to antennas in deep space, engineers have used this FEA-based tool to drive the design of electronics and communication products that have shaped our modern economy. It is a prime example of advanced capability, ease of use, and efficient execution.

Electronics Thermal Analysis

The final example to look at is what the industry calls a vertical application — a CAE tool focused on a specific use case. Ansys Icepak electronics cooling simulation software is a pre- and post-processing tool built on top of the Fluent solver, purpose-built for electronics cooling and PCB thermal simulation and analysis.

Related Resources