The Engineering Method

I watched a YouTube video that gave a great definition of the Engineering Method.

Solving problems using heuristics that cause the best change in a poorly understood situation using available resources.

Oddly, this was never explained to me during the course of obtaining a Bachelor of Science in Aerospace Engineering as I detailed in my post Science vs Engineering

The video explains how medieval European stonemasons sized a wall when designing those scenic, wide-open cathedrals. As an ex-stress analyst, I can certainly appreciate this task. After all, what is a launch vehicle but an axially-loaded, circular cylinder?

One excellent example of heuristics used by aerospace engineers to size the walls of load bearing cylinders is NASA SP-8007, Buckling of Thin-Walled Circular Cylinders.

Section 2.5.1 of SP-8007, “Traditional Empirical Design Approach” says:

The traditional approach for the preliminary design of a thin-walled buckling-resistant shell is to predict the buckling load of the shell using a classical linear eigenvalue analysis or approximate closed-form solution and then apply an empirical correlation factor, commonly known as a knockdown factor, to account for the difference between the predicted buckling load and the actual buckling load determined from tests.

I mainly worked on two projects during my years as an aerospace engineer, the Harpoon anti-ship missile, and the Commercial Titan launch vehicle.

We didn’t use a “knockdown factor”, but we did do other superstitious I mean empirical things.

On both of them, we used an “effective axial load”, which is a combination of the design axial load, plus the peak axial compression component caused by the design bending moment. At no time would a rocket or missile fuselage experience the effective axial load, except at one point on the fuselage’s circumference. In fact, 180° around the circumference, the bending moment would relieve the axial load.

Commercial Titan was the only one of the two vehicles that had substantial burst pressure loads, so when checking CT’s buckling strength, we ignored any helpful effects of burst pressure. Burst pressure tends to push out any tiny imperfections that lead to buckling, and it pre-loads the fuel and oxidizer tanks in tension. CT carried only about 50 psi burst pressure, but it was 10 feet in diameter, so the tension pre-load was around 500,000 pounds.

Naturally, we only ignored helpful tensile loading when calculating buckling strength, which is relevant only to compression loads. When looking at membrane stress in fuel and oxidizer tanks, you better believe we considered axial tension.

The general heuristic exhibited in these design practices is to ignore helpful effects, and combine detrimental effects.