Engineering Trade Studies & Architecture Selection
A trade study organises engineering judgement — it does not eliminate it. This article covers the technical criteria that drive architecture selection, the danger of false objectivity in numerical weighting, sensitivity to weighting choices, and why the selection must follow programme priorities and requirements rather than a computed score.
What a Trade Study Is — and What It Is Not
A trade study is a structured comparison of candidate architectures or design concepts against a set of criteria that reflect the engineering requirements and programme priorities. Its purpose is to organise the engineering judgement of the team into a transparent, traceable and challengeable comparison, so that the selection of a preferred concept is based on stated criteria and stated reasoning rather than on unexamined preference. A trade study is not a formula that produces a correct answer. It does not replace engineering judgement; it structures it. The criteria, the weights and the scoring all reflect judgements made by the engineers performing the study. A trade study that is performed well makes those judgements visible, so that they can be challenged, tested for sensitivity and revised. A trade study that is performed badly hides the judgements behind numerical scores and presents the result as an objective ranking, which it is not. The distinction is critical: a good trade study is an aid to decision-making; a bad trade study is a substitute for thinking that produces a defensible-looking but potentially wrong answer.
A TRADE STUDY ORGANISES ENGINEERING JUDGEMENT — IT DOES NOT ELIMINATE IT. The criteria, the weights and the scores are all judgements. A well-executed trade study makes those judgements visible and challengeable. A poorly executed one hides them behind a numerical score and pretends the answer is objective.
The Criteria That Drive Structural Architecture Selection
The criteria used in a structural trade study depend on the programme, the customer requirements and the engineering discipline, but a representative set for a structural architecture selection includes mass, stiffness, performance, cost, risk, manufacturability, testability, maintainability and maturity. Each criterion measures a different attribute of the candidate concepts, and each interacts with the others. Mass and stiffness are often in tension — a stiffer structure typically requires more material and therefore more mass. Performance may be in tension with cost — a higher-performance concept may be more expensive. Risk may be in tension with maturity — a novel concept may offer performance advantages but carry higher technical risk because it is less mature. Manufacturability may be in tension with testability — a concept that is easy to manufacture may be harder to instrument or test. The criteria are not independent; they interact, and the interactions are themselves engineering judgements. The engineer performing the trade study must understand these interactions, not merely score each criterion in isolation.
| Criterion | What it measures | How it is assessed | When it dominates | How it interacts with other criteria | Risk of over-weighting |
|---|---|---|---|---|---|
| Mass | The structural mass of the candidate concept, typically a primary driver in aerospace and automotive structures | Estimated from preliminary sizing, material selection and configuration; may include mass growth allowance for immature concepts | In weight-critical applications — aerospace launch, unmanned aerial vehicles, racing — where every kilogram affects performance or range | Lower mass often means lower stiffness or lower margin; advanced materials reduce mass but may increase cost and risk | Over-weighting mass can drive selection toward flimsy, low-margin or high-risk concepts that fail on other criteria |
| Stiffness | The resistance to deformation under load — static stiffness, dynamic stiffness, natural frequencies | Estimated from preliminary structural models, beam or plate theory, or coarse finite element models of each concept | When deflection limits, frequency requirements or aeroelastic boundaries govern the design rather than strength | Higher stiffness typically requires more material (mass) or stiffer materials (cost); affects thermal stresses in constrained structures | Over-weighting stiffness can drive over-design and unnecessary mass when strength, not stiffness, is the governing requirement |
| Performance | The functional performance of the concept against the system requirement — load capacity, thermal performance, life, accuracy | Estimated against the requirement; may be a margin, a capability envelope or a functional metric specific to the application | When the requirement is performance-driven and the concept's performance margin is the primary discriminator | Higher performance often requires more mass, higher cost or advanced materials; performance margins reduce risk | Over-weighting performance can select a concept that exceeds the requirement at the expense of cost, mass or risk that is not justified |
| Cost | The recurring and non-recurring cost of the concept — materials, manufacture, assembly, tooling, qualification | Estimated from material costs, manufacturing process, tooling requirements, assembly complexity and qualification scope | In cost-constrained programmes where affordability is a primary driver; in production programmes with high unit count | Lower cost may mean higher mass (cheaper materials), lower performance, higher risk (less mature process) or lower maintainability | Over-weighting cost can drive selection toward concepts that are cheap to build but expensive to operate, difficult to verify or risky in service |
| Risk | The technical risk associated with the concept — novelty of the approach, immaturity of the technology, uncertainty in performance | Assessed qualitatively or semi-quantitatively; considers technology readiness, analysis confidence, test heritage and consequence of failure | In safety-critical or schedule-critical programmes where a technical failure or delay is unacceptable; for primary structure with high consequence of failure | Lower risk often means higher mass (conservative design), higher maturity (proven approach) and lower performance; risk reduction adds cost and time | Over-weighting risk can drive selection toward conservative, mature concepts that forgo performance or mass advantages that the programme actually needs |
| Manufacture | The manufacturability of the concept — ease of production, process availability, tooling complexity, tolerance capability | Assessed by manufacturing engineers; considers process maturity, supplier capability, tolerance achievement and inspection requirements | When production volume is high, when manufacturing capability is limited, or when tolerance control is critical to function | Easier manufacture may mean higher mass (simpler shapes), fewer joints (monolithic but harder to inspect) or limited material choice | Over-weighting manufacturability can select a concept that is easy to build but suboptimal in mass, stiffness or performance |
| Testability | The ease with which the concept can be verified by test — access for instrumentation, ability to apply representative loads, ability to detect failure | Assessed by test engineers; considers access, instrumentation locations, load introduction, failure mode visibility and test rig complexity | When test verification is mandatory and the test article must be representative; for safety-critical structure where test evidence is the primary verification | Better testability may require design features (bosses, flats) that add mass; some concepts are easier to test than to analyse and vice versa | Over-weighting testability can drive design changes that compromise the in-service structure to make the test easier |
| Maintainability | The ease with which the concept can be inspected, repaired and maintained in service — access, inspectability, repairability | Assessed by maintainability analysis; considers inspection access, damage detectability, repair procedures and tool access | In long-life structures where in-service inspection is required; for damage-tolerance designs where inspectability governs the inspection interval | Better maintainability may mean more access features (mass), more joints (complexity, potential leak paths) or redundant load paths | Over-weighting maintainability can add mass and complexity that the service requirement does not justify |
| Maturity | The technology readiness and heritage of the concept — how proven the approach is in similar applications | Assessed against technology readiness levels, service history, demonstrated performance in similar structures and availability of design data | In certification-driven programmes where demonstrated heritage is required; when the schedule does not allow technology development | Higher maturity means lower risk but may mean lower performance or higher mass; mature materials and processes have established allowables and design data | Over-weighting maturity can prevent the adoption of beneficial new technologies and lock in suboptimal legacy solutions |
Comparing Three Structural Concepts
To illustrate how a trade study works in practice, consider the selection of a structural architecture for a representative component — say, a wing carry-through structure for a small aircraft. Three candidate concepts might be considered: a conventional aluminium built-up assembly (riveted skins, spars and ribs), a carbon-fibre composite sandwich construction, and a hybrid aluminium-composite design with composite skins over aluminium substructure. Each concept has different attributes across the criteria. The purpose of the trade study is not to compute a score that declares a winner, but to make the attributes of each concept visible side by side so that the engineering team and the decision-makers can see what is being traded and why. The selection follows the programme priorities: if mass is the primary driver and the programme can accept the manufacturing investment and the technical risk of composite construction, the composite concept may be preferred. If cost and manufacturing maturity dominate and the mass penalty is acceptable, the aluminium built-up concept may be preferred. If a balance is needed, the hybrid may be preferred. The trade study does not make this decision; it provides the structured comparison on which the decision is based.
TRADE STUDY — THREE STRUCTURAL CONCEPTS COMPARED ON QUALITATIVE ATTRIBUTES Criterion │ Concept A: Al built-up │ Concept B: CFRP sandwich │ Concept C: Hybrid Al/CFRP ───────────────────────┼────────────────────────┼──────────────────────────┼────────────────────────── Mass │ Higher │ Lowest │ Intermediate Stiffness │ Moderate │ Highest │ High Manufacturing │ Well established │ Complex, high tooling│ Moderate complexity │ │ │ Testability │ Good (accessible) │ Moderate (embedded) │ Moderate Technical risk │ Low (proven) │ Higher (process) │ Moderate Maturity │ High (heritage) │ Moderate (newer) │ Moderate ───────────────────────┴────────────────────────┴──────────────────────────┴────────────────────────── No concept is automatically "best". Selection follows programme priorities and requirements: • Mass-critical, risk-tolerant, investment available → Concept B (composite) • Cost-critical, maturity-driven, mass acceptable → Concept A (aluminium) • Balanced mass/cost/risk → Concept C (hybrid) The trade study makes the trade-offs VISIBLE. The decision follows the programme priorities, not a computed score.
Weighting and the Illusion of Objectivity
Many trade studies assign numerical weights to the criteria and numerical scores to each concept, then compute a weighted total that ranks the candidates. This approach has a legitimate purpose: it forces the team to be explicit about which criteria matter most and it makes the comparison reproducible. But the approach also creates a serious risk: the numerical result can appear objective when it is not. The weights are not measured quantities — they are judgements about which criteria matter most to this programme. The scores are not measured quantities either — they are judgements about how well each concept performs against each criterion, often on an arbitrary scale. When these judgements are multiplied and added, the result carries the precision of arithmetic but the subjectivity of the inputs. A weighted score of 8.7 versus 8.3 looks like a clear ranking; it is actually the sum of several judgements, each uncertain, and the difference may be well within the uncertainty of the scoring. Treating such a result as an objective ranking hides the subjectivity in the weighting and scoring choices, and it can drive a selection that is not actually supported by the engineering reasoning.
ASSIGNING NUMERICAL WEIGHTS TO CRITERIA AND THEN TREATING THE RESULTING SCORE AS OBJECTIVE RANKING HIDES THE SUBJECTIVITY IN THE WEIGHTING CHOICE. The weights are engineering judgement, not measured quantities. The scores are engineering judgement, not measured quantities. The arithmetic is precise; the inputs are not. A weighted total is a summary of judgements, not an objective answer.
Sensitivity to Weighting
Because the weights are judgements, the ranking produced by a weighted trade study can change when the weights change. A concept that wins when mass is weighted highest may lose when cost is weighted highest. A concept that wins when risk is weighted low may lose when risk is weighted high. This sensitivity is not a defect of the trade study; it is a feature. It reveals that the selection depends on the programme priorities, which is exactly what it should reveal. The engineer performing the trade study should explicitly test the sensitivity of the ranking to the weights: if the ranking changes when the weights are varied over their plausible range, the selection is sensitive to the weighting, and the decision must be made with that sensitivity in mind. The most important outcome of a sensitivity analysis is not the identification of a winner but the identification of the conditions under which each concept is preferred. If concept A is preferred when mass dominates and concept B is preferred when cost dominates, the decision reduces to a question about programme priorities — which is a question that the engineering team can present to the decision-makers clearly and transparently, rather than hiding it inside a computed score.
- Test the sensitivity of the ranking to the weights — vary the weights over their plausible range and observe whether the ranking changes
- If the ranking is insensitive to the weights, the selection is robust — any reasonable weighting gives the same answer
- If the ranking is sensitive to the weights, the selection depends on programme priorities — make that dependence explicit, do not hide it
- Identify the conditions under which each concept is preferred — this is more useful to the decision-maker than a single ranking
- Present the sensitivity to the decision-makers, not just the baseline ranking — they need to know that the selection depends on priorities they set
When to Use a Trade Study
A trade study is most valuable when there are multiple viable candidate architectures and the selection depends on how the criteria are prioritised. It is less valuable — and potentially misleading — when there is only one viable concept, when the requirements are so constraining that only one architecture is feasible, or when one concept is so clearly superior on the dominant criterion that the other criteria do not matter. In the first case, a trade study is unnecessary; the requirements have already made the selection. In the second case, a trade study is unnecessary; the dominant criterion has made the selection. A trade study is also less valuable when the criteria and weights are imposed without engineering understanding — the study becomes a paperwork exercise that produces a number without engineering content. The value of a trade study comes from the engineering thinking that goes into the criteria, the weights, the scoring and the sensitivity analysis. Without that thinking, the study is a formality.
Trade Study Checklist
The following checklist supports the conduct of a trade study. It is not a mandated procedure; it is a set of engineering questions that, if addressed, produce a trade study that is transparent, challengeable and useful to the decision-maker. The specific format and documentation requirements depend on the organisation and the programme.
- Candidate concepts are genuinely distinct — If the candidates are minor variations of the same approach, the trade study adds little value
- Criteria reflect the actual programme requirements and priorities — Not generic criteria copied from a template — criteria specific to this decision
- Each criterion has a clear definition of what is being assessed — The team must agree on what "stiffness" or "risk" means in this context before scoring
- Weights are stated explicitly and justified — The weights are judgements — they must be visible, not hidden inside a formula
- Scores are based on engineering assessment, not guessed — Each score should be traceable to a preliminary sizing, a manufacturing assessment, a risk analysis or a maturity evaluation
- Sensitivity to weighting has been tested — If the ranking changes when the weights change, that sensitivity is reported to the decision-maker
- Interactions between criteria are understood — Mass and stiffness, cost and performance, risk and maturity — the interactions are engineering judgements, not independent numbers
- The selection is presented as following programme priorities, not as an objective ranking — The trade study informs the decision; the decision-maker sets the priorities
- The trade study is documented and traceable — A reviewer should be able to follow the criteria, weights, scores, sensitivity and reasoning to the conclusion
Key Takeaways
A trade study is a structured comparison of candidate concepts against criteria that reflect programme requirements and priorities. It organises engineering judgement; it does not eliminate it. The criteria — mass, stiffness, performance, cost, risk, manufacture, testability, maintainability, maturity — interact, and the interactions are engineering judgements. Numerical weighting is useful for making priorities explicit but dangerous when the resulting score is treated as an objective ranking. Sensitivity to weighting should be tested explicitly, and the conditions under which each concept is preferred should be identified. The selection follows programme priorities and requirements, not a computed score.
- A trade study organises engineering judgement — it does not eliminate it
- The criteria interact — mass, stiffness, cost, risk, maturity are not independent
- Numerical weights and scores are judgements, not measured quantities
- Treating a weighted score as objective ranking hides the subjectivity in the weighting
- Sensitivity to weighting should be tested — if the ranking changes, the selection depends on programme priorities
- Identify the conditions under which each concept is preferred, not just a baseline winner
- The selection follows programme priorities and requirements — the trade study informs the decision, it does not make it