English is currently available.
Civil Engineering DISCUSSION
I am checking a 20 mm diameter solid steel rod, 2 m long and pinned at both ends, used as a compression strut. Based on a yield strength of 250 MPa and an area of 314 mm², it should carry about 78 kN. The Euler formula gives a buckling load of only about 3.9 kN.
If the load is applied exactly along the axis, where does the sideways bending come from? And what can I change to raise the buckling load without simply using much more steel?
Buckling is a loss of stability, not a material failure. Picture the strut with a tiny sideways bow of size δ. The axial load P now has a lever arm and produces a bending moment P × δ that tries to increase the bow, while the bending stiffness of the rod tries to straighten it. Below the critical load the stiffness wins and the rod springs back. Above it, the moment grows faster than the resistance, and any disturbance grows.
A perfectly straight rod with a perfectly central load is an idealisation. Real members have a slight initial bow and some load eccentricity, so the sideways deflection starts from the first newton and grows rapidly as the load approaches the critical value.
Your numbers are right. For a solid round bar I = π × d⁴ / 64 = π × 20⁴ / 64 = 7854 mm⁴, and P_cr = π² × E × I / L² = 9.87 × 200,000 × 7854 / 2000² = 3880 N.
The slenderness ratio tells you why. The radius of gyration of a solid round is d / 4 = 5 mm, so L / r = 400, which is extremely slender. The average stress at buckling is π² × E / (L / r)² = 12 MPa, about 5 percent of yield. Notice that yield strength does not appear in the formula at all: a higher grade of steel would buckle at exactly the same load, because E is the same.
The end conditions enter through the effective length K × L, which replaces L in the formula. Pinned at both ends, K = 1. Fixed at both ends, K = 0.5, which quadruples the load. Fixed at one end and pinned at the other, K is about 0.7, which doubles it. Fixed at one end and free at the other, a flagpole, K = 2 and the load falls to a quarter.
Real connections are never perfectly fixed, so design guidance uses more cautious values than the theoretical ones. A more dependable gain is a brace at mid-height that prevents sideways movement: it halves the buckling length and gives four times the load, provided the brace acts in every direction the strut could bow.
To raise the load without more steel, move the material away from the centre. A tube with the same cross-sectional area as your rod, 40 mm outside diameter with a wall of about 2.7 mm, has I = π × (40⁴ - 34.6⁴) / 64 = about 55,000 mm⁴, seven times that of the solid rod. The buckling load rises to about 27 kN for the same weight per metre.
That is why compression members are tubes, angles and box sections and tension members can be rods and cables. The limit to the approach is local buckling: a wall that is too thin relative to the diameter crumples locally before the member buckles as a whole.
One caution on the formula: Euler is valid only for slender members, where the buckling stress is well below yield. Setting π² × E / (L / r)² equal to 250 MPa gives L / r = π × sqrt(200,000 / 250) = 89. Around and below that slenderness, the member fails by a combination of yielding and buckling, and residual stresses and initial crookedness pull the real strength below both the Euler load and the squash load.
Design codes handle this with column curves that give a reduced design stress for each slenderness, and they add safety factors on top. Use the Euler load to understand the behaviour and the code curves to size anything that carries real load.
