Iterative design is the practice of repeating the design-build-test cycle multiple times, with each iteration producing a better version based on data from the previous round. Rather than trying to get everything right on the first attempt, engineers deliberately plan for multiple rounds of prototyping and testing. Each iteration narrows the gap between the current design and the requirements. The key principle is that testing reveals problems that analysis alone cannot predict, and each round of testing provides data that drives targeted improvements.
Assign a multi-round design challenge where students build, test against quantitative criteria, record results, identify the weakest performance area, redesign only that area, and test again. Track performance across iterations on a graph to visualize improvement. Discuss diminishing returns -- early iterations produce big gains, later ones produce smaller refinements. Compare to how video games release patches or how cars improve across model years.
The Wright brothers did not build one airplane and fly it perfectly. They built gliders, tested them, measured lift and drag, redesigned the wings, tested again, built a wind tunnel to get better data, redesigned again, and only after years of iterating did they achieve powered flight. This is iterative design -- the practice of deliberately cycling through design, build, test, and improve multiple times to converge on a solution that works.
The power of iteration comes from a fundamental truth: you cannot predict everything from analysis alone. No matter how carefully you calculate stress in a beam or airflow over a wing, real materials behave in ways that surprise you. Joints loosen. Heat warps components. Users hold the product differently than you expected. Testing reveals these surprises, and each round of iteration addresses them.
A disciplined iterative process has a clear structure. After each test, engineers ask three questions: What worked? (keep it), What failed? (change it), and What do we still not know? (test for it next time). This prevents random tinkering. You do not change everything at once -- you change the weakest link, retest, and see if performance improves. Changing one variable at a time also makes it clear what caused any improvement.
One important concept is diminishing returns. Your first prototype of a solar water heater might achieve 30% efficiency. After one round of testing and redesign, you might jump to 55%. Another round gets you to 65%. Then 70%. Then 72%. Each iteration costs time and money, but the improvements shrink. At some point, the cost of further iteration exceeds the value of the improvement. Engineers use their requirements to define "good enough" -- when the design meets all requirements, iteration stops, even if further improvement is theoretically possible.
Professional engineering uses different fidelity levels across iterations. Early prototypes might be rough cardboard models that test only the basic concept. Middle iterations use more realistic materials and test specific subsystems. Late iterations are near-final versions tested under realistic conditions. This progression from low-fidelity to high-fidelity prototyping saves enormous resources -- it is far cheaper to discover a fundamental flaw using a cardboard model than a precision-machined prototype.