Other engineering cheat sheet

Define the system

Choose a system boundary, inputs, outputs and assumptions before selecting equations.

A component model and a whole-system model answer different questions.

Dimensional consistency

Quantities added or equated must have compatible dimensions. Dimensionless groups compare relevant physical effects.

Energy cannot be added directly to power without a time factor.

Trade-offs

Engineering choices balance competing requirements. An optimum depends on the criteria and constraints.

Reducing mass may affect stiffness, cost or manufacturing options.

Verification and validation

Verification checks whether a solution follows its specification or model. Validation asks whether the model addresses the intended real-world purpose.

A correct calculation can still use an unsuitable model.

Requirements and constraints

A requirement describes what success must achieve; a constraint limits available choices. Make both measurable before comparing proposed designs.

“Carry 5 kg” is a performance requirement; “mass below 1 kg” is a design constraint.

Model fidelity

A model should include the effects important to the question. A more detailed model is not automatically more useful if its parameters are unknown or its assumptions remain unchecked.

Start with a simple estimate, identify its largest uncertainty, then refine that part.

Sensitivity analysis

Sensitivity examines how a result changes as an input changes. It helps identify which measurements or assumptions deserve the most attention.

For A = πr², a small relative change in radius produces approximately twice that relative change in area.

Uncertainty and margins

Measurements, materials and operating conditions vary. A robust comparison considers uncertainty and the consequences of failure, not only an ideal nominal calculation.

List the assumed input range before treating one numerical result as definitive.

Worked example: efficiency

A teaching model receives 500 J and delivers 350 J of useful output. Efficiency is 350/500 = 0.70, or 70%; the remaining 150 J must be accounted for elsewhere.

Useful output depends on the defined purpose and system boundary.

Common mistake: precision equals accuracy

More decimal places do not prove that a model is accurate. Report digits consistent with the evidence and check whether the assumptions dominate the result.

An output of 12.0000 does not remove an input uncertainty of several units.

Check your understanding

How do verification and validation differ?

Verification checks whether a model or product meets its specified implementation or requirements; validation checks whether it is suitable for the intended use.

Curriculum connections

Sources