How to Turn Engineering Assumptions into Tests
Every assumption in a design is an implicit test waiting to be written. Making that explicit is what validation planning is.
An assumption is a statement about the world that the design depends on being true. The engineer who writes it does not know it is true. They believe it is probably true, or they have reasoned that it must be true if other things they know are true.
None of that is evidence. Evidence comes from testing.
The gap between assumption and evidence is the basic problem that validation planning is designed to close. The mechanism for closing it is straightforward: for each assumption, write the test that would confirm or refute it.
The structure of an assumption-to-test conversion
A useful assumption has three parts.
First, a claim: The bonded joint will maintain sufficient stiffness at the specified operating temperature.
Second, a condition: For this claim to hold, the adhesive modulus must remain above the threshold value across the full thermal range.
Third, a consequence of failure: If the condition does not hold, the load path changes and the secondary structure carries load it was not designed for.
Given those three parts, the test writes itself:
- Subject: the bonded joint, or a representative specimen
- Method: measure stiffness across the thermal range
- Accept/reject criterion: modulus above the threshold value at every temperature in the range
- Failure mode: if the test fails, the load path analysis must be rerun with the actual measured modulus
That is a validation test. It is traceable to the assumption. It has a pass criterion. It produces a definite answer.
Why assumptions stay implicit
The reason assumptions are rarely converted to tests is that writing the test makes the assumption visible. Once visible, it becomes a programme commitment: someone has to run the test, at some cost, by some date.
Implicit assumptions are free. They carry no schedule impact, produce no resource demand, and create no awkward conversation about whether the design can afford to be tested. The cost appears later, usually when the condition turns out to be false and the design work built on top of the assumption has to be re-examined.
A practical discipline
The most effective place to surface this is in design review. The question the review panel should ask, for each major design decision:
What would have to be true about the world for this decision to be correct?
Each answer to that question is an assumption. Each assumption that carries structural weight in the design — that affects load paths, interfaces, requirements compliance — should have a corresponding entry in the validation plan.
Not every assumption needs a physical test. Some can be closed by analysis against measured material data. Some can be closed by reference to prior programme evidence. The important thing is that each one has a plan for closure that is not simply “we will assume it is fine.”
The validation plan is the record of how you intend to convert assumptions into evidence. If an assumption does not appear in the validation plan, it is not being managed.