Conceptual Design Process Questions and Answers

Answer: C

Correct answer: Unsound; it commits most lifetime cost

Skipping the phase does not remove the decisions it makes; it only makes them implicitly, without exploring alternatives, while still committing the great majority of the lifetime cost. The schedule saved is a few per cent of the budget, and the exposure is the other sixty-five. A derivative programme can legitimately run a much shorter conceptual phase, since the configuration is inherited and only the changes need exploring - but that is compressing the phase, not skipping it.

Answer: D

Correct answer: The mission exceeds the technology assumed

Divergence means the growth factor exceeds unity: each added kilogramme demands more than a kilogramme in return, so the aircraft has no finite size at which the mission closes. Iterating further only produces larger numbers, and the result is not merely a heavy aeroplane but no aeroplane at all. The response is to change an input - relax range or payload, or assume better aerodynamics or fuel consumption and record the risk.

Answer: A

Correct answer: It is computationally cheaper than optimising disciplines separately

This is the false statement. Coupling the disciplines means the whole system must be evaluated at every point in the search, which is far more expensive than optimising one discipline at a time with the others frozen. The expense is accepted because the interactions dominate the answer, and it is mitigated with surrogate models and designed experiments rather than avoided.

Answer: B

Correct answer: Iterating the loop many more times

If the loop is diverging, further iterations produce larger numbers and nothing else; the failure is in the inputs, not in the arithmetic. The legitimate responses all change an input: reduce the payload, relax the range, or assume a more efficient technology and record the readiness risk that assumption creates. Each is a negotiation with the requirement rather than an attempt to compute a way out of it.

Answer: C

Correct answer: Meaningless; the classes differ entirely

A regression carries no information beyond the aircraft it was fitted to. A jet transport is a pressurised, swept-wing, high wing-loading machine with an empty weight fraction near 0.5; a sailplane is unpressurised, unpowered, of enormous aspect ratio and very low wing loading. Nothing in the fitted relationship applies, and the answer will be wrong by a large and unknown factor rather than conservatively so. Unconventional or out-of-class designs must be estimated from physics much earlier.

Answer: D

Correct answer: Replacing all physical testing

A digital mock-up finds clashes between systems, structure and equipment, rehearses maintenance access with a human model and a tool envelope, and checks that the assembly sequence is physically possible - all things that were once discovered on the first article at enormous cost. What it cannot do is replace testing: material behaviour, fatigue, fastener joints, and the many effects that emerge only in real hardware still have to be measured.

Answer: A

Correct answer: Fidelity rises as breadth falls

This is the fundamental shape of the process. Conceptual design examines many configurations with fast, approximate methods; preliminary design narrows to one and applies wind tunnel testing and structural analysis; detail design works on that single configuration in complete detail. Breadth and fidelity trade against each other because the effort per configuration rises as the number falls, and a programme that tried to do both at once would never finish the first sweep.

Answer: B

Correct answer: The value of recording readiness

Each of the three advances was presumably reasonable when assumed, and the difficulty is that nobody added them up. A design can accumulate optimistic assumptions silently and end up depending on all of them arriving together, which is a far larger risk than any one of them alone. Recording each assumption with its readiness level makes the stack visible while there is still time to fund maturation or plan a fallback.

Answer: C

Fitting a response surface to a modest number of high-fidelity runs allows the optimiser to search thousands of points at negligible cost. Its accuracy holds only within the region sampled, so the fit must be checked before its predictions are trusted.

Answer: D

Correct answer: Choose which combinations to evaluate

A full factorial sweep grows exponentially with the number of variables, so design of experiments selects a much smaller set of points that still resolves the main effects and the important interactions - Latin hypercubes, orthogonal arrays and similar schemes. Twenty-five well-chosen points can characterise a space that would need hundreds by brute force, and the results usually feed a surrogate model that stands in for the full analysis during the search.