Engineering & Technology: Assumption surfacing

What you'll be able to do: Find the assumptions you made without noticing, which is where your marks go and where real structures fail.

Why this is the first method for engineering

Read any inquiry report into a serious engineering failure. The proximate cause is usually mundane. The interesting part is always an assumption that nobody wrote down: that the load would be static, that the material would be at room temperature, that the operator would follow the procedure, that the component would be replaced on schedule, that water would drain.

Unstated assumptions are the failure mode of the discipline. And the same thing is true in miniature on every exam paper: marks are allocated to stating and justifying assumptions, and students treat that section as throat-clearing before the calculation.

The calculation is the easy part. Anyone can execute it. The engineering is deciding what to assume, and knowing that you decided.

The assumptions engineers make without noticing

These are the ones that recur, across disciplines:

Loading. Static when it's dynamic. Uniform when it's point. A single load case when there are several. Nobody standing on it.

Material. Linear elastic. Isotropic. At room temperature. Not corroded. Not fatigued. The properties on the datasheet rather than the ones in the batch.

Boundary conditions. Perfectly rigid support. Perfectly pinned joint. No friction. Infinitely stiff foundation. These are the ones that bite hardest because the idealised versions don't exist.

Steady state. That transients don't matter. That startup and shutdown are the same as running.

Use. That it'll be used as intended. That it'll be maintained. That someone will notice the warning.

Environment. Room temperature, dry, clean, no vibration, no UV, no salt.

Scale. That behaviour at test scale predicts behaviour at full scale.

How to run it

1. Present the reasoning, not the answer. The setup is where the assumptions hide.
2. Ask for what you're relying on but haven't said.
3. Sort them: which you'd defend, which you'd drop if challenged, which you didn't realise you'd decided. The third category is the valuable one.
4. For each, ask what changes if it's false. Does the answer change by 5% or by a factor of ten? Does it fail safely or catastrophically?
5. State the important ones in the work. A stated, justified assumption is strength; an unstated one is a hole.

Here's my analysis: [paste your setup and working]. Don't tell me if the answer is right. List the assumptions I'm relying on but haven't stated. For each: what happens to my result if it's false, does it change the magnitude, or does it change whether the thing works at all?

That second question is the engineering one. Some false assumptions move a number; others mean the whole approach was wrong.

Across the disciplines

Structural. Support conditions. A connection modelled as pinned that's actually semi-rigid changes the moment distribution entirely.

Mechanical. Friction, tolerance stack-up, and thermal expansion. The assumption that parts are nominal size is the one that produces assemblies that don't assemble.

Electrical. Ideal sources, zero wire resistance, and that components are at their nominal value. Tolerance analysis is the discipline of not assuming this.

Chemical. Steady state, perfect mixing, isothermal operation, and that the reaction goes as written. Runaway reactions are usually an unstated assumption about heat removal.

Civil. Ground conditions, drainage, and that the loading in the code covers what will actually happen. Ground is the classic, it's assumed uniform and never is.

Software. Input is well-formed, the network is reliable, the clock is monotonic, the disk has space, nothing runs concurrently. The fallacies of distributed computing are a list of unstated assumptions.

Industrial. That demand is stationary, that setup times are fixed, and that people will follow the process.

The exam version

Here's an exam question and my answer. Which assumption marks would I have earned, and which available ones did I not? What should I have stated that I treated as obvious?

Students lose marks here in volume, and it's the cheapest category to recover, you knew the assumption, you just didn't think it was worth writing.

Pitfalls

1. Confusing assumptions with errors. Most are legitimate choices. The goal is knowing you made them.
2. Listing trivia. "Gravity is 9.81" is not the interesting kind. Ask for the load-bearing ones.
3. Stopping at the list. The what-if-it's-false test is the method.
4. Only doing it when stuck. Run it on analyses you're confident about, that's where unstated assumptions sit most comfortably.
5. The tell: every assumption surfaced is one you'd already have named. Give it your actual working rather than a tidied version.

Try this today

Take the last piece of analysis you did. Paste the setup and ask only: what am I assuming here that I haven't stated?

Then for the most important one: does my answer change by a few percent if it's false, or does the whole approach stop being valid?