Skip to content
Queensferry

Module 8 · Lesson 8.1

Trusses and frames

What the pin-jointed idealisation omits, and where a connection stops being pinned.

Why this matters

Two idealisations carry most of structural steelwork: that a truss is pin-jointed, and that a frame connection is either rigid or pinned. Both are extremely useful and neither is true.

The interesting question is not whether they are true but how far wrong they are, and in which direction — which is answerable by building the model both ways.

By the end of this lesson you should be able to

  • Quantify what the pin-jointed idealisation omits
  • Explain why a heavier chord attracts more secondary bending
  • Sweep a connection stiffness and find where the moment actually changes
  • Use the free bending moment as a check on any portal result

What you should already know

  • Module 7's releases — a truss is a frame with every member released
  • Module 5's Warren truss, which is the same model used again here

The pin-jointed truss

Every textbook truss has frictionless pins at every joint. Every real truss has chords that run continuously through the panel points, welded or bolted, and are bent by the joint rotations.

Model the same 24 m Warren truss both ways — pin-jointed bars, then continuous chords with pin-ended web members — and the comparison is reassuring in one respect and instructive in another:

Pin-jointedContinuous chords
Largest bottom chord force340.0 kN339.1 kN
Bending in that chord01.41 kN·m
Midspan deflection20.30 mm20.25 mm

The axial force is within 0.3 % and the deflection within 0.3 %. The idealisation is excellent for the things it was meant for.

What it omits is the bending, and 1.41 kN·m in a member sized for 340 kN axial is not nothing: through the section it accounts for about 7.7 % of the peak stress. A member checked at 95 % utilisation on axial force alone is over.

The counter-intuitive part

Make the chord stiffer and the secondary bending goes up, not down.

Bending follows stiffness. A continuous member picks up moment in proportion to how strongly it resists the rotation being imposed on it, so a heavy chord attracts more of it than a light one. A truss cannot be made safe against secondary bending by making the chords bigger; it can only be made safe by detailing the joints so the rotation is not imposed, or by checking the combined stress.

Where a connection stops being pinned

Module 7 established that a release is a spring of zero. Sweeping that spring on a 12 m portal with fixed bases at 15 kN/m:

Connection stiffnessEaves momentMidspan momentShare of rigid
Pinned (0)0270.00 %
1 000 kN·m/rad38.2231.826 %
3 00075.8194.251 %
10 000115.5154.578 %
30 000135.8134.291 %
Rigid (∞)148.9121.1100 %

Two things are worth taking from that table.

The eaves and midspan moments always add to 270 kN·m, which is wL²/8 — the free bending moment of the span. That is not a property of this frame; it is statics, and it holds for every row. It is therefore the check to run on any portal result before reading anything else.

Half the rigid moment arrives at about 1.65 EI/L. The beam's EI/L here is 1 750 kN·m/rad, and the 50 % point is at 2 896. Ninety per cent needs 14.9 EI/L — nine times as much stiffness for the last 40 %.

So the curve is steep exactly where real connections are, and flat well above them. A connection an engineer would describe as 'quite stiff' is often at 50–70 % of the rigid moment, and calling it rigid over-predicts the eaves moment and under-predicts the midspan by a similar margin — in opposite directions, on the same member.

The linkage

One combination is not a frame at all. Pinned bases and pinned connections give four pins in a four-bar arrangement, which is a mechanism: nothing resists sway. A solver should refuse it, and it is worth knowing that some will instead return a very large number without comment, because the mechanism's pivot survives a naive positive-definiteness test.

The free bending moment identity

What it calculates: The check that any portal result must satisfy, whatever is assumed about the connections

w
uniformly distributed load on the beam (kN/m)
L
span (m)
M
moment (kN·m)

This assumes

  • A uniformly loaded beam between two supports, however those supports are restrained

In plain terms: This is statics rather than a property of any particular frame, so it holds at every point on the connection sweep. If a model's two moments do not add to wL²/8, the problem is the model rather than the connection assumption.

Try it

Truss modelling workshop

A 24 m Warren truss as pin-jointed bars, and again as a frame with continuous chords.

Bending follows stiffness — try increasing it.

Chord force, pin-jointed
340.0 kN
Chord force, continuous chords
339.1 kN
Secondary bending in that chord
1.41 kN·m

not present in the pin-jointed model at all

Bending as a share of peak stress
7.7 %
Midspan deflection, pin-jointed
20.30 mm
Midspan deflection, continuous
20.25 mm

And the beam analogy, for the same truss

Midspan moment, actual point loads
1080 kN·m
Midspan moment, loads smeared to a UDL
900 kN·m
Error from smearing
16.7 %
Chord force ≈ M/d
360 kN

What this shows: The idealisation gets the axial force right and omits the bending entirely — and a stiffer chord attracts more of it, not less.

Try it

Frame connection explorer

Sweep the beam-to-column connection from a pin to full rigidity. Their sum is the free bending moment at every point.

The beam's own EI/L is 1750 kN·m/rad — this is 1.71 × that.

eaves (solid)midspan (dashed)stiffness, log scale →
Eaves moment
75.78 kN·m
Midspan moment
194.22 kN·m
Sum
270.00 kN·m

wL²/8 = 270.00 — equal at every stiffness

Share of the rigid eaves moment
50.9 %
Sway at eaves
0.023 mm
Rigid connection would give
148.93 kN·m
Pinned connection would give
270.00 kN·m at midspan

The two moments always sum to the free bending moment. That identity is the fastest check available on any portal result — if a model’s moments do not add to wL²/8, the problem is the model rather than the connection assumption.

What this shows: Rigid and pinned are the ends of one parameter, and the curve is steep exactly where real connections are.

Worked example

A connection assumed rigid that is half-stiff

Given

  • A 12 m portal, fixed bases, 15 kN/m on the beam
  • The design assumes rigid connections
  • The connections as detailed have a rotational stiffness of about 3 000 kN·m/rad

Find

What the design used and what the frame will do

    Practice

    A portal's eaves moment is 75.8 kN·m and the free bending moment wL²/8 is 270 kN·m. What is the midspan sagging moment?

    Practice

    A truss chord carries 340 kN axial. The pin-jointed model reports 340 kN and the continuous model 339.1 kN. What is the percentage difference in axial force?

    Check yourself

    A truss chord is upgraded to a heavier section. What happens to the secondary bending in it?

    Worked example

    Checking a portal result in ninety seconds

    Given

    • A model reports an eaves moment of 96 kN·m and a midspan sagging moment of 210 kN·m
    • The beam spans 14 m under 12 kN/m

    Find

    Whether the result can be right

      Check yourself

      Why does the eaves-plus-midspan identity hold at every connection stiffness?

      Check yourself

      The pin-jointed truss idealisation gets the axial forces within a fraction of a per cent. What does that justify?

      Check yourself

      What makes a portal with pinned bases and pinned connections a mechanism?

      Summary

      • The pin-jointed truss gets the chord force within 0.3 % and the deflection within 0.3 %
      • It omits bending entirely — here 7.7 % of the chord's peak stress
      • A stiffer chord attracts more secondary bending, not less
      • Eaves plus midspan moment equals wL²/8 at every connection stiffness
      • Half the rigid moment arrives at about 1.65 EI/L; ninety per cent needs 14.9
      • Pinned bases with pinned connections is a linkage, not a frame
      Progress is kept in this browser only.

      This is educational material. It uses simplified examples to teach principles, and must not be relied on for real design or safety-critical decisions. Module overview and checkpoint