top of page
Writer's pictureBhargava Krishna Marripati

What is Finite Element Analysis (FEA)? Understanding the Simulation Technique for Predicting Mechanical Behavior under Various Conditions



What is Finite Element Analysis (FEA)? Understanding the Simulation Technique for Predicting Mechanical Behavior under Various Conditions

Introduction to Finite Element Analysis


Finite Element Analysis (FEA) is a computer simulation technique used to predict how physical objects will behave under different loading conditions. FEA works by breaking down complex problems into small, simplified pieces called "elements". Mathematical equations help predict the behavior of each element, which are then combined to model the entire structure.


FEA emerged in the 1960s within the aerospace and nuclear industries. Engineers developed the method to analyze complex structures like airplane wings and pressure vessels which couldn't be calculated by hand. Advancements in computing brought FEA into the mainstream by the 1980s. Today it is used across every engineering field to optimize designs.


The basic premise of FEA is that a body or structure can be divided into smaller elements of finite size and simple geometry. The software then applies material properties and boundary conditions to construct simultaneous equations that model the behavior of the entire structure. FEA uses matrix algebra and numerical methods to solve these algebraic equations and approximate a solution.


Engineers use FEA to model mechanical responses like stress concentrations, temperature distribution, pressure loading, natural frequencies, fatigue life, heat transfer, fluid flows, and many other behaviors. FEA provides insights that can't be obtained through physical testing alone. It reduces costs by minimizing the need for physical prototypes and testing. FEA has become an indispensable tool for design optimization across automotive, aerospace, biomedical, civil, electrical, and mechanical engineering industries.


Approximating Solutions with Finite Element Analysis


Finite Element Analysis utilizes the divide and conquer strategy to break down complex problems into small and simple elements that can be solved efficiently. The geometry of the entire structure or domain is discretized into a finite number of elements, connected at points called nodes.


The field variables like displacement, temperature etc are approximated over these elements using shape functions. Simple functions like polynomials are used to represent the field variables over each finite element. These elemental approximations are then assembled to represent the behavior of the entire structure.


The complex partial differential equations representing the physics are also discretized and converted into simple algebraic equations for each element. These equations can then be efficiently solved using matrix algebra.


So FEA divides a complex problem into small simple problems, approximates solutions for each, and combines them to obtain the overall solution for the entire structure. This efficient numerical technique allows engineers to simulate reality and study complex phenomena.


Governing Equations: The Mathematical Foundation of FEA


At the core of finite element analysis are the governing equations that represent the physics of the problem being analyzed. FEA aims to find approximate numerical solutions to these equations over complex domains like the geometry of a 3D model.


The governing equations are generally partial differential equations (PDEs) that describe phenomena like structural mechanics, fluid flow, and heat transfer. For example, the Navier-Stokes equations govern fluid flow, Fourier's Law governs heat transfer, and equations like Hooke's Law govern structural mechanics.


These PDEs relate properties like velocity, pressure, temperature, and displacement to their rates of change over time and space. FEA software solves these PDEs by breaking the model geometry into many small elements and solving simplified versions of the equations for each element.


There are two main approaches to deriving the finite element equations from the governing PDEs:


  • Weak Formulation: The PDEs are converted to an integral form and solved via calculus of variations. This provides a more flexible way to handle complex geometries.

  • Strong Formulation: The PDEs are discretized directly in their differential form. This is less flexible but computationally more efficient.


Most commercial FEA solvers use the weak formulation as it offers more modeling capabilities for real-world engineering problems. The weak formulation also leads to a system of equations that connect the elements and can be efficiently solved using matrix methods.


Meshing


Meshing is one of the most critical steps in finite element analysis. It involves dividing the geometry into small elements to create a finite element mesh.


Creating the Mesh



The geometry is discretized into a number of nodes which are connected to form a mesh of elements. The mesh can contain different element shapes like tetrahedral, hexahedral, shell, or beam elements based on the geometry and analysis type.



Meshing software like ANSYS Meshing can generate meshes automatically for complex geometries. Mesh generation requires balancing the element size and number to achieve good accuracy without excessive computational cost.


Element Types


Different element types are used based on the physics of the problem. Tetrahedral elements can mesh complex shapes easily but may require more elements. Hexahedral elements provide better accuracy but can be difficult to generate for complex parts.


Shell and beam elements are commonly used to take advantage of their 2D or 1D formulations to reduce the mesh size. Choosing the right element type and meshing technique is important for an accurate and efficient FEA model.


Mesh Quality


Mesh quality metrics like aspect ratio, skewness, and element angles determine how well-shaped the elements are. A good quality mesh improves accuracy and simulation stability. Checking mesh metrics and resolving any issues like highly distorted elements is important.


Mesh Convergence


Mesh convergence refers to iteratively refining the mesh to determine the point at which the solution stops changing significantly. A converged mesh ensures the simulation has an accurate solution. Checking deformation, stress, or other results for convergence is necessary.


Loads and Boundary Conditions


In finite element analysis, loads and boundary conditions need to be applied to the model to simulate how it will react in the real world.


Loads represent forces, pressure, or other effects that will cause stress, deformation, temperature change, fluid flow or other behavior in the structure. Some examples of loads include:


  • Point forces

  • Pressure loads

  • Thermal loads from temperatures or heat fluxes

  • Fluid flow loads

  • Gravity or body force loads

  • Rotational speeds

  • Dynamic or time-varying loads


Constraints and boundary conditions restrict motion or degrees of freedom in the model. Common examples include:


  • Fixed constraints that restrict all translation and rotation at a point

  • Sliding constraints that allow translation only along one axis

  • Symmetry conditions that reduce the model size

  • Contact constraints between components

  • Boundary conditions like fixed temperatures


Properly applying loads and boundary conditions is crucial for obtaining accurate results from FEA. The analysis will only be as good as the inputs used. It's important to accurately determine the types, locations, directions, and magnitudes of loads the product will experience in reality.


Boundary conditions must also match the physical constraints in the real model. Assuming incorrect or incomplete boundary conditions is a common source of errors in FEA.


Material Models


The accuracy of any FEA simulation depends heavily on using the proper material models. Material models mathematically describe how a material will behave under various loading conditions. The most basic material model is linear elasticity, which assumes the material deforms proportionally to the applied load. While simple, linear elasticity only applies to a limited range of problems. More complex material behaviors require advanced nonlinear material models.


Linear Elastic Materials


The most basic material model used in FEA is linear elasticity. It assumes the material obeys Hooke's law, which states stress is proportional to strain. While all real materials show some nonlinear behavior, linear elasticity works well for small strains and simple loading. It requires only two material properties - Young's modulus and Poisson's ratio. Linear elasticity provides a good approximation for materials like aluminum, steel, and concrete under moderate loads.


Nonlinear Material Models


Real-world materials demonstrate complex nonlinear behaviors. As stress increases, materials can yield, harden, soften, and fracture. Advanced nonlinear material models are needed to capture these effects. Common nonlinear models include plasticity, hyperelasticity, viscoelasticity, and creep. These models require experimentally determined stress-strain curves and other advanced properties. Proper nonlinear models are crucial for accurate simulation of forming, impact, and high-temperature processes.


Anisotropic Materials


Many materials demonstrate directionally dependent properties, known as anisotropy. Composites like fiberglass and layered materials like wood require anisotropic models. These define unique material properties in different directions aligned with the material orientation. Accounting for anisotropy is vital for precise simulation of composite structures and directionally loaded members.


By using the appropriate linear or nonlinear, isotropic or anisotropic models, FEA can accurately predict how nearly any real material will perform. However, reliable models require comprehensive material testing and characterization. The precision of any FEA solution is ultimately limited by the fidelity of the selected material models.


Solvers


Once the finite element model is set up with a mesh, loads, boundary conditions, and material properties, it must be processed by a solver to obtain results. There are two main types of solvers used in FEA:


Direct Solvers


Direct solvers obtain the solution to the finite element equations directly in a single step. They factorize the stiffness matrix and solve the system of equations by substitution. Direct solvers provide accurate solutions and work well for simpler linear problems. However, they require extensive computer memory and processing power. The computational expense scales poorly as the problem size increases.


Iterative Solvers


Iterative solvers start with an initial approximation and improve the solution iteratively until convergence criteria are met. They solve the equations through a succession of approximate solutions, with less memory requirements than direct solvers. However, iterative techniques can suffer from slow convergence or fail to converge for complex non-linear problems.


Examples of iterative methods include the conjugate gradient method for linear systems and the Newton-Raphson method for non-linear systems. The computational expense of iterative solvers scales better for large complex models. However, care must be taken to ensure proper convergence, through metrics like the residual error.


Overall, direct solvers are preferred for smaller linear problems where accuracy is critical. Iterative solvers enable solving larger non-linear problems but require monitoring convergence to avoid inaccurate results. The choice depends on the particular FEA application, model size, and available computational resources.


Post-Processing Results


Once the finite element analysis is complete, engineers can visualize the results in various graphical formats to gain insights. Common post-processing outputs include:


Stress Visualization


  • Von Mises stress plots to see where yielding may occur

  • Principal stress plots to see maximum and minimum stresses

  • Safety factor plots to see the margin against failure


Deformation

  • Displacement plots to see how much deformation occurs

  • Contour plots to see overall deformation patterns

  • Deformed shape plots to visualize distortion


Animation


  • Animations of stress propagation

  • Animations of deformation over time

  • Mode shape animations for vibration analysis


Optimization


  • Topology optimization to minimize mass

  • Shape optimization to reduce stress

  • Parameter optimization to find optimal design variables


Post-processing allows engineers to thoroughly analyze the FEA results and gain actionable insights. The graphical outputs help assess design feasibility, troubleshoot issues, and guide design changes. Animations are particularly useful for understanding the physics. Optimization leverages FEA to automatically improve the design.


Applications of Finite Element Analysis


Finite Element Analysis has a diverse range of applications across various engineering disciplines. Here are some of the key areas where FEA is commonly used:


Structural Analysis


FEA allows engineers to simulate and predict the deformation, stresses, and strains on a structure under loading conditions. This application is most common in mechanical, civil and aerospace engineering. Some examples include:


  • Stress analysis on mechanical parts like gears, shafts, and pressure vessels

  • Vibration analysis to determine natural frequencies and mode shapes

  • Structural integrity analysis of buildings, bridges, aircraft, and spacecraft

  • Crashworthiness analysis for vehicles


Thermal Analysis


FEA enables the modeling of heat transfer systems to calculate parameters like temperature distribution and heat fluxes. This can be used to design and optimize heat sinks, microelectronics, and more. Examples:


  • Modeling heat dissipation in electronic devices like CPUs and PCBs

  • Analyzing thermal stresses in turbine blades, ovens, engines

  • Simulating injection mold cooling systems


Fluid Flow Analysis


Computational Fluid Dynamics (CFD) uses FEA principles to analyze fluid flow in applications like aerodynamics, weather prediction, and combustion modeling. Examples:


  • Modeling airflow over aircraft wings and car bodies

  • Analyzing turbomachinery like turbines, compressors, pumps

  • Simulating mixing and chemical reactions in engines and combustion chambers


Design Optimization


FEA enables rapid virtual testing of design changes to optimize parameters like shape, weight, strength, and cost. This application sees use across industries. Examples:


  • Automotive - optimizing car body stiffness and crashworthiness

  • Aerospace - reducing aircraft wing weight while preserving strength

  • Biomedical - customizing implants and prosthetics for patients

  • Manufacturing - minimizing tool wear and improving quality


As shown above, FEA is a versatile analysis method used extensively in mechanical, aerospace, civil, biomedical, and many other engineering domains. Continued advances in computing power and FEA software will enable more complex multiphysics simulations in the future.


Future of FEA


The field of finite element analysis continues to evolve with new methods and technologies. Some emerging trends in FEA include:


Isogeometric Analysis


The isogeometric analysis aims to integrate CAD and FEA models more tightly. Traditional FEA uses simple elements like triangles and rectangles to approximate CAD geometries. The isogeometric analysis uses the same mathematical basis (NURBS) for both CAD and FEA models, enabling more accurate simulations directly from CAD.


GPU Computing


Graphics processing units (GPUs) are being leveraged for parallel processing in FEA. Complex simulations can be accelerated by distributing computations across thousands of GPU cores. This allows larger, higher fidelity models to be solved interactively. GPU computing may enable real-time FEA in the future.


Model Order Reduction


For complex simulations like transient dynamics or multiscale models, model order reduction techniques are being used to reduce computational cost. These techniques build simplified meta-models that capture the dominant physics with far fewer degrees of freedom. The reduced models can be solved very efficiently.


Cloud Computing


FEA software and computing power are moving to the cloud. Cloud-based FEA enables on-demand access to advanced simulation capacities without expensive hardware investments. It allows engineers to easily collaborate on large simulation projects as well.


As computing power grows exponentially, FEA will push the boundaries in terms of problem size, geometry details, and multi-physics simulations. The technique will continue enabling engineers to accurately predict real-world performance and drive innovation.

4 views0 comments

Comments


Blogs last image.png
bottom of page