Mechanical Engineering Personal Statementfor Oxford & Cambridge

News coverage of the Titan submersible implosion in 2023 made mechanical engineering feel less like a tidy set of formulas and more like an argument about what can be trusted. I had thought about strength in simple terms: either a part survives a load or it does not. Reading more about pressure vessels, fatigue and safety factors made that view feel naive. Failure is rarely about one number being too small; it comes from the interaction of material behaviour, geometry, manufacturing and judgement. I became interested in how engineers decide what counts as safe when a structure may seem reliable before it is not. That question kept pulling me toward mechanical engineering because it sits at the point where mathematical models have to face real materials, real manufacturing limits and real consequences.

A-level Physics gave that interest a clearer shape, especially through stress, strain and Young's modulus. I was less interested in memorising the graph than in what it leaves out: two materials can tolerate the same load and still be very different engineering choices once stiffness, fracture behaviour and repeated loading matter. J. E. Gordon's Structures: Or Why Things Don't Fall Down helped me think about bending, buckling and fatigue, while sections of Engineering Materials 1 by David R. H. Jones and Michael F. Ashby showed me that material selection is a compromise between competing requirements rather than a search for the strongest material. My EPQ then let me test those ideas. I investigated how far low-cost prototyping can be trusted when evaluating stiffness and failure in student designs, using aluminium strips as a control against 3D-printed PLA samples of the same geometry. The aluminium data followed theory more closely, while the PLA results shifted with print orientation and varied more between nominally identical samples. The hardest part was deciding what counted as an explanation. At first I treated disagreement with theory as measurement error. Repeating the tests forced me to see that the discrepancy was the result. That helped me understand why prototypes are not just cheaper versions of finished parts; the manufacturing process can be a main reason they behave differently.

Outside formal education, I wanted to test these ideas directly, so I built a project for my school engineering club and entered it into the Big Bang UK Young Scientists and Engineers Competition. I investigated how closely simple beam theory predicts the behaviour of 3D-printed cantilever beams. Using Fusion 360, I designed identical PLA specimens and printed them with different layer orientations. I then clamped each beam, added masses at a fixed distance, measured deflection with digital calipers, and compared the results with values I calculated in Python using the Euler-Bernoulli beam equation. The model gave me a useful baseline because it predicts how deflection varies with load and flexural rigidity, but it also assumes a much simpler material than a printed polymer actually is. Beams loaded across the layer lines failed earlier and showed more scatter than the calculation suggested. That made me focus less on whether the theory was 'right' and more on where its assumptions stopped matching the material. Working part-time in a cycle repair shop has reinforced the same lesson. Components are not judged only by the maximum load they can survive once; wear, fatigue and ease of replacement affect whether a design is durable in use. That has made me want to study mechanics, materials and structures in greater depth, particularly how engineers make sound decisions about failure, reliability and design trade-offs when real use does not behave as neatly as a model would like.