Skip to main content

Numerical solution techniques, advantages and suitability

Short answer

FEM divides the whole domain into elements and suits complex geometry and materials but struggles with large deformation. FDM uses regular grids, is simple, but poor for irregular boundaries. BEM discretises only the boundary, good for infinite domains, weak for nonlinearity. DEM models particles or blocks for discontinuities but is computationally costly.

Describe the different numerical solution techniques used in solving civil engineering problems (with the help of mechanics), giving their advantages, disadvantages and suitability.

Civil engineering problems (stress analysis, seepage, flow in channels, soil-structure interaction) are governed by differential equations that rarely have closed-form solutions for real geometry and loading. Numerical techniques replace the continuous problem by a set of algebraic equations. The main techniques are FEM, FDM, BEM, DEM and mesh-free methods (SPH).

1. Finite Element Method (FEM)

The body is divided into elements joined at nodes; displacement inside an element is interpolated by shape functions and the equations [K]{U}={F}[K]\{U\}=\{F\} are formed from energy or virtual work (e.g. a beam with [k][k] from Hermite functions).

  • Advantages: handles irregular geometry, mixed materials, non-linearity, any boundary condition; systematic and general; large software base.
  • Disadvantages: whole domain must be meshed; large equations; poor for infinite domains, cracks and large deformation (mesh distortion).
  • Suitability: structures (frames, trusses, plates, shells), foundations, dams, soil stress analysis.

2. Finite Difference Method (FDM)

Derivatives are replaced by differences between grid-point values, e.g. ∂u∂x≈ui+1−ui−12Δx\dfrac{\partial u}{\partial x}\approx\dfrac{u_{i+1}-u_{i-1}}{2\Delta x}.

  • Advantages: simple concept and programming; efficient on regular grids; explicit and implicit time stepping.
  • Disadvantages: difficult for curved or irregular boundaries; grid refinement is global; stability limits on Δt\Delta t.
  • Suitability: groundwater flow, seepage, open-channel and pipe flow, heat flow, flow nets.

3. Boundary Element Method (BEM)

The governing PDE is converted into an integral equation on the boundary only; only the boundary is discretised.

  • Advantages: reduces dimension by one; small data; good for infinite or semi-infinite domains and stress concentration.
  • Disadvantages: full, unsymmetric matrices; needs a fundamental solution; poor for non-linear and heterogeneous media.
  • Suitability: soil-structure interaction, tunnels, fracture, potential flow.

4. Discrete Element Method (DEM)

Material is modelled as an assembly of separate blocks or particles; contact forces are found from force-displacement laws and Newton's second law is integrated in time.

  • Advantages: models large displacement, sliding and separation naturally.
  • Disadvantages: costly; contact parameters hard to calibrate.
  • Suitability: granular soils, jointed rock, masonry, rockfall.

Comparison

MethodDiscretisesStrengthWeakness
FEMWhole domain (elements)Complex geometry, materialsLarge deformation
FDMWhole domain (grid)SimplicityIrregular boundary
BEMBoundary onlyInfinite domainsNon-linear, non-homogeneous
DEMParticles/blocksDiscontinuaComputational cost

FEM is the most widely used because it suits both solids and fluids and complex geometry.

✓ Answer checked by Claude Opus 5.5 (high effort)

All Computational Techniques in Civil Engineering chapter 1 questions →