Beams with One Redundant
Remove the extra support, then push the beam back to where that support held it.
Before this Lesson
A steel beam cantilevers 20 ft from a wall under 2 kips per foot, and its tip sags 1.2 in. Jack the tip back to level and slide a post under it. How hard does the post push? With four reactions and three equations, statics can't say. But you know how far the post holds the tip up, 1.2 in., and a push at the bare tip lifts it 0.08 in. per kip. So the post pushes kips.
Take the support away, then push back
From Determinacy and Stability, a beam fixed at one end and propped at the other, a propped cantilever, is indeterminate to degree 1, so one reaction is a redundant: the beam would stand without it. The force method takes the redundant away, leaving a released structure (or primary structure) that statics can handle. It finds how far that beam moves where the support was, then makes the redundant push it back, just as the bar between two walls in Thermal Strain and Indeterminate Bars grew freely and was pushed back until it fit. The Handbook's Elementary Statically Indeterminate Structures by Force Method of Analysis says it in one line: the deflection from removing the support is equal and opposite to the deflection the redundant causes. Its sketch is the hook's beam.
Adapted from the FE Reference Handbook, Elementary Statically Indeterminate Structures by Force Method of Analysis
From Beam Deflection by Superposition: the Handbook's beam tables measure upward, so a sag prints negative. Four rows do most of the work:
- Cantilever, uniform load: at the tip.
- Cantilever, a distance from the wall: at the tip, or with the load at the tip.
- Simple beam at midspan: under a uniform load, or under a midspan load.
With the post's force positive upward, as in the Handbook, the load's deflection at on the released beam is , and the redundant's is :
That's 15 kips for the hook's beam. And cancelled: a stiffer beam sags less, but it's harder to push back up by the same factor. With supports that stay put, the reactions depend only on the loads and the lengths. Once is known, statics gives the rest, including the moment at the wall.
Two spans: the released beam is twice as long
A beam continuous over two equal spans , on a pin and two rollers, has one redundant too. Remove the middle support, , and the released structure is one simple beam long. That's the length that goes into the table:
Adapted from the FE Reference Handbook, Elementary Statically Indeterminate Structures by Force Method of Analysis
The Handbook's force-method page sketches this beam but calls the whole length , so solving its equations gives . Either way, the middle support takes 5/8 of the total load. Each end gets the rest, , which is the propped cantilever's prop reaction. That's no coincidence. By symmetry, the beam is level over , as if clamped there, so each span is a propped cantilever with its wall at .
So the moment over is the propped cantilever's wall moment: , with tension on top.
Choosing what to release
Any one restraint can be the redundant, as long as the released structure is stable and determinate. A propped cantilever can keep its roller and release the wall's grip on rotation instead, leaving a simple beam. The redundant is then the wall's moment , and compatibility is a slope: the load tilts the end by , and the table's end-couple row says tilts it back by , so . On a continuous beam, the same move puts a hinge over an interior support.
Watch three things.
- Stability. Release a propped cantilever's horizontal restraint at the wall, and nothing resists a sideways push. As in Determinacy and Stability, the released structure must pass the count and the look.
- Direction. The redundant pushes the beam back, so and have opposite signs.
- Settlement. If the support sinks, the beam has to meet it where it now sits, not at zero.
With a settlement , no longer cancels:
The settlement takes load off the support, and more of it the stiffer the beam.
Fixed at both ends
Clamp both ends, and the count says three redundants, but under vertical loads the horizontal reactions are zero, which leaves two: the wall moments. That's past the one-redundant method, but the Handbook's Fixed End Moments table has done the work. A fixed-end moment, FEM, is the twist a wall needs to hold its end level while the load tries to tilt it. Under gravity loads, both are negative, with tension on top.
| Load | Fixed–fixed, | Fixed–fixed, | Fixed–roller, |
|---|---|---|---|
| Uniform | |||
| at midspan | |||
| at |
Adapted from the FE Reference Handbook, Fixed End Moments
The table gives sizes, and its last column is the propped cantilever from earlier. For the moment anywhere else, superpose once more. A fixed-fixed beam is a simple beam under the load plus a moment at each end, and end moments make a straight-line diagram. So hang the simple beam's diagram from a line joining the end moments. At midspan:
For a uniform load, that's . The walls lower the parabola's baseline by two-thirds of its height.
Restraint pulls moment toward the supports
In the Handbook's sign convention, a positive moment bends a beam concave upward. So a moment diagram is a map of curvature: where is positive, the beam smiles, and where it's negative, it frowns. Where crosses zero, the curvature flips at an inflection point. The supports supply the rest of the shape:
- Every support holds the beam at zero deflection.
- A fixed end also holds it level.
- The beam passes over an interior support without a kink. Only a hinge makes one.
- On a line of symmetry, the slope is zero.
Now restrain one uniformly loaded span three ways.
Restraint pulls moment toward the supports. The simple beam carries at midspan. A wall at one end takes itself, leaving in the span, and walls at both ends take each, leaving . The sag shrinks with the midspan moment. Each sign change in is one inflection point: two in a span that's negative at both ends and positive between, one in a span that runs from negative to positive and ends at a simple support, and none where keeps its sign.
Load one span, and the next one lifts
Load just one span, and its neighbor moves the other way. On two spans with a load in only, sags and tips the beam over like a seesaw. The beam can't kink at , so the unloaded span bows up, enough that the roller at has to hold it down. Its moment runs straight from the negative back to zero at , a frown the whole way with no inflection point. Influence Lines uses this pattern to choose which spans to load.
Portal frames pushed sideways
A portal frame's corners are rigid, and a rigid joint keeps its angle: if a column top turns, the girder end turns with it. The Handbook says nothing about the shapes of frames, so this is reasoning you bring.
Push the top sideways with a force . Both column tops sway together, and the girder resists turning, so they turn only a little. A fixed base holds its column vertical, so the column starts and ends nearly vertical, but offset. That takes an S, with an inflection point near mid-height: exactly there if the girder were rigid, and a bit higher for a real one. Both girder ends turn the same way, so the girder is an S too, with its inflection point at midspan.
A frame has no top or bottom, so these diagrams plot each member's moment on its compression side, the concave side. For a beam, that puts positive moment above the axis, as usual. The moment changes sides at every inflection point.