Axial Stress
Stress is force spread over area, and it peaks where the area is smallest.
Before this Lesson
- Equilibrium of a Particle · Statics
- Trusses: Method of Joints · Statics
A structural steel rod half an inch thick, about as big around as your little finger, can hang a 5,000 lb pickup truck, and it doesn't start to give way until about 7,000 lb. A quarter-inch rod, half as thick, gives way at about 1,800 lb. Half the thickness, a quarter of the load. It's the same steel, so the total force can't be what decides this. What decides it is how many pounds each square inch has to carry.
Force shared out over area
Picture the rod as a bundle of thin strands, all pulling together, with the truck's weight shared out among them. Four times the cross-section means four times the strands, so each carries a quarter as much. A strand doesn't know how heavy the truck is. It only feels its share.
That share, force per unit of area, is stress. When the force runs straight through the section, perpendicular to it, it's normal stress. The Handbook prints it under Uniaxial Loading and Deformation:
The Handbook's area table in Statics gives a round rod , with the radius. Questions usually give the diameter, which makes it . The size gets squared, and that explains the two rods: 5,000 lb over the half-inch rod's 0.196 in² is 25,500 psi, and over the quarter-inch rod's 0.0491 in² it's 102,000 psi. Structural steel starts to yield at about 36,000 psi.
That even sharing holds away from where the load enters and away from holes. Right at a clamp or beside a hole, the stress bunches up.
A pulled member is in tension and a squeezed one is in compression, with the same formula. The Handbook's entry has no sign rule, so this topic uses the one from its Mohr's circle entry: tension is positive and compression is negative. Here you only find the stress in a compression member. A slender one can buckle first, and that waits for Euler Buckling.
The units come straight from what you divide:
| Force | Area | Stress |
|---|---|---|
| lb | in² | psi |
| kip | in² | ksi (1,000 psi) |
| N | mm² | MPa |
The last row saves work. A megapascal is a million N/m², and a square meter is a million mm², so 1 MPa is 1 N/mm². Between systems, the Handbook's conversion table gives 1 psi = 6,895 Pa, so 1 ksi is about 6.9 MPa.
Cut the member, and the force shows itself
The in that formula is the force inside the member, at the section you care about. You can't see it from outside, so cut.
Saw through the rod with the truck still hanging, and the lower piece falls. Before the cut it sat still, so the steel across the cut was pulling up exactly as hard as the truck pulls down. Treat that piece as its own body, with put back at the cut, and equilibrium gives . The upper piece gives the same , because whatever one piece does to the other, the other does back.
That's the method of sections:
- Cut through the member at the section you want.
- Draw either piece, with the unknown pulling on it: the arrow leaves the cut face and points away from the piece.
- Sum forces along the member. A positive is tension, and a negative one is compression.
The Handbook words this only for trusses, under Plane Truss: Method of Sections, but the move works on anything. It's the same cut you make in Trusses: Method of Sections and Internal Shear and Moment in Beams.
It also splits every question in this topic in two. The statics half finds the internal force at the section that's worst off. The materials half divides it by the right property of the section, here the area.
- From Equilibrium of a Particle: isolate, draw, sum to zero. A cable only pulls, and a strut pushes or pulls.
- From Trusses: Method of Joints: a member pinned at both ends with nothing in between is a two-force member. One that rises 3 over a run of 4 splits its force in the ratios and .
- From Beam Reactions: moments about a pin leave out the pin's own force.
In a bracket or on a wire, the statics half usually takes one equilibrium step. Isolating a joint is a cut too: it exposes the force in each member. On a rigid beam held by a hanger, the step is moments about the pin.
Each segment carries its own force
You'd expect a stepped bar to be worst off at its thinnest part. Try one: fixed to a wall, thick (2 in²) near the wall and thin (1 in²) at its free end. Pull the free end with 10 kip and the shoulder with 30 kip, both away from the wall.
Cut through the thin part and keep the piece away from the wall. Only the 10 kip acts on it, so the stress is 10 ksi. Cut through the thick part, and both loads act on the piece: 40 kip over 2 in², or 20 ksi. The thick part is working twice as hard.
So a member is only as strong as its worst combination of force and area, and you have to check each segment. A segment's force isn't the load beside it. It's whatever balances everything on the piece. The force changes only at a load, so one cut per segment does it, and keeping the piece that doesn't touch the wall means you never need the reaction. "Largest" means largest magnitude, unless a question says otherwise.
Pressure on one area, stress on another
In a hydraulic cylinder, oil pushes on the whole face of a piston, and a much thinner rod carries that push to the load. Pressure and stress are both force per area, with the same units, which is where the trouble starts. There are two areas here, and each number belongs to one of them.
The pressure acts on the piston's area and makes a force. The rod carries that force through its own area. That's the statics half and the materials half again:
The Handbook has no piston formula, only pressure as force per area in its Fluid Mechanics section, printed as a capital . Lowercase here keeps it apart from the internal force.
Put together, the rod's stress is the pressure times the ratio of the areas, . A 6 in. piston on a 1.5 in. rod makes the rod's stress 16 times the oil pressure.
A plate works the same way, over whatever area the pressure covers. A design stress is the highest stress the rod is allowed to reach, so set to it and solve for what's missing.
How close to failure?
A stress means little until you compare it with what the material can take. The yield strength, , is the stress at which a ductile metal starts to stretch for good. The ultimate strength, , is the highest stress it reaches before breaking. Strain, Hooke's Law, and Poisson's Ratio puts them on a curve. For now, they're table values.
| Material | Yield, | Ultimate, |
|---|---|---|
| Structural A36 steel | 36 | 58 |
| Stainless 304 steel | 30 | 75 |
| 6061-T6 aluminum | 37 | 42 |
| 2014-T6 aluminum | 60 | 68 |
| C86100 bronze | 50 | 95 |
Source: FE Reference Handbook, Table 2 - Average Mechanical Properties of Typical Engineering Materials (adapted)
Table 2 prints U.S. units only, so an SI question will state the strength or leave you to convert it: A36's 36 ksi is about 248 MPa. These metals have the same values in compression.
The factor of safety compares what a member can take with what it's taking. Since , you can say that in stresses or in loads:
The Handbook prints factors of safety for retaining walls, slopes, and shafts, but no general definition, so carry this one in. A question will say whether it means against yield or against ultimate, and the same member has a different factor for each.
For a largest load, run it backward. Strength over FS gives the allowable stress, the same kind of cap as a design stress, and that times the area gives the load:
A hole leaves less to carry the load
Drill a hole in a plate and pull. The force has to get past that spot through the material that's left, so cut there, straight across the hole at its widest. The whole section is the gross area, . What survives the cut is the net area, . The force is the same all along the plate, so the smaller area governs.
For a plate of width and thickness with a hole of diameter , the piece missing from the cut is a rectangle, not a circle, because the cut sees the hole edge-on:
A 6 in. by 1/2 in. plate with a 1 in. hole has and , so at an allowable stress of 20 ksi it can carry 50 kip, not 60.
A hollow tube loses material all along its length. The Handbook's area table prints , with outer and inner radii. With diameters, that's , and a wall of thickness makes .
The stress right at a hole's edge runs higher than this average, but a net-area question uses the average. And here you deduct the hole as dimensioned. The rule changes in Steel Tension Members, where the Handbook's net-area formula adds 1/8 in. to the bolt diameter.