Structural Idealisation Questions and Answers

Answer: C

A quarter model represents a quarter of the structure and must receive a quarter of the load. Applying the whole load overstates everything by about four times, and it is a mistake that reaction checking would catch immediately.

Answer: D

Correct answer: It preserves what matters, discards the rest

That is the whole definition. Retaining every detail would not be an idealisation at all, and would defeat the purpose. Nor is idealisation automatically conservative - discarding a load path can be unconservative if it was carrying load away from the region of interest. And far from removing the need to understand the load path, it requires that understanding, since deciding what to discard is deciding what carries what.

Answer: A

Correct answer: Refining the mesh always improves the answer, without limit

This is false, and its two exceptions are important. At a geometric singularity - a re-entrant corner, a point load - the exact solution is infinite, so refinement produces ever larger stresses rather than convergence. And refinement addresses only discretisation error; it does nothing about an incorrect idealisation, wrong boundary conditions or wrong loads. The other three statements are all true.

Answer: B

Correct answer: A solid mesh of every fastener

Preliminary sizing needs answers in hours, on a configuration that is still changing, so the model must be coarse enough to be built and rebuilt quickly. Booms and shear panels, their finite element equivalents of rods and shear panels, and a beam representation of the box are all appropriate. Meshing every fastener as a solid would take longer than the design phase and would answer a question nobody is yet asking.

Answer: C

Correct answer: Too strong; convergence is not validity

Convergence shows that the equations the model poses have been solved accurately, and nothing more. If the idealisation is wrong, the boundary conditions unrepresentative or the loads mistaken, refinement converges faithfully on the wrong answer. Verification - are the equations solved correctly? - and validation - are they the right equations? - are separate questions, and convergence addresses only the first.

Answer: D

Correct answer: They are more accurate than finite element models

Hand methods are not more accurate; a well-built finite element model will beat them on any geometry of consequence. They survive for the other three reasons: they make the load path explicit and so force the analyst to think about it, they give an independent order-of-magnitude check on the computer’s answer, and they are quick enough to keep pace with a design that is still moving.

Answer: A

The factor is two plus the stress ratio, so it falls from three to one across the practical range. Physically, a panel whose stress reverses along its length carries much of its load internally and contributes less to either boom.

Answer: B

The real part has a radius and the model does not, so the number is an artefact of the geometry idealisation rather than a prediction. Refining will only raise it further, since the underlying elasticity solution is unbounded there.

Answer: C

The panel gives up its shear stiffness on buckling but continues to carry load as a series of inclined straps. The uprights and flanges must then be sized for the inward pull those straps produce, which is a load case that would not exist otherwise.

Answer: D

The individual flanges buckle locally at a stress below the material's yield, and the section's capacity is governed by that rather than by column buckling or yield. Crippling stress is estimated from empirical methods based on flange width-to-thickness ratios.