Biaxial and Unsymmetrical Flexure
Key Takeaways
- Resolve moments about the same centroidal axes used to calculate section properties; an x offset creates bending about y and a y offset creates bending about x
- For principal axes, biaxial elastic stress is the algebraic sum of both bending terms, so each corner needs its signed x and y coordinates
- The governing tensile and compressive corners can be missed when moment magnitudes are added without tracking signs
- For an unsymmetrical section with nonzero product of inertia, transform properties and moments to principal axes or use the full coupled formula consistently
- The neutral-axis direction generally does not match the resultant-moment direction when the two principal inertias differ
Biaxial and Unsymmetrical Flexure
Draw the axes: A biaxial stress calculation is a signed stress plane, not
M_x/S_x + M_y/S_ywith every term assumed positive. Label the centroid, positive axes, positive moments, and every candidate point before substituting numbers.
Biaxial Bending About Principal Axes
For centroidal principal axes x and y, the product of inertia is zero and the two bending contributions can be superimposed directly. Adopt positive x to the right, positive y upward, tensile normal stress positive, and a moment convention giving
σ(x,y) = -M_x y/I_x + M_y x/I_y.
Under this stated convention, positive M_x compresses the +y side and positive M_y tensions the +x side. Another handbook convention can be used, but moments, coordinates, and signs must all change consistently. Section-modulus shortcuts work at selected extreme fibers for one-axis bending; for simultaneous biaxial bending, evaluate both terms at the same physical point.
| Item | Axis discipline |
|---|---|
I_x | Measures y²; pairs with M_x and coordinate y |
I_y | Measures x²; pairs with M_y and coordinate x |
Eccentricity e_y | Creates a moment about the x axis |
Eccentricity e_x | Creates a moment about the y axis |
Signs of eccentric-load moments come from the cross product and the chosen axial-load direction. Do not attach e_x to M_x merely because the subscripts match. A transverse load that misses the shear center can also produce torsion; keep that action separate from the axial-eccentricity mapping above.
Worked Four-Corner Stress Check
A solid rectangle is 8 in wide along x and 12 in deep along y. It carries M_x = +80 kip-ft and M_y = +30 kip-ft under the convention above. Its centroidal properties are
I_x = bh³/12 = 8(12³)/12 = 1,152 in⁴
and
I_y = hb³/12 = 12(8³)/12 = 512 in⁴.
Convert moments: M_x = 960 kip-in and M_y = 360 kip-in. At a corner, x = ±4 in and y = ±6 in, so
σ = -(960/1,152)y + (360/512)x
σ = -0.8333y + 0.7031x in ksi.
Evaluate signs instead of adding magnitudes blindly:
Corner (x,y), in | -M_xy/I_x, ksi | +M_yx/I_y, ksi | Total σ, ksi |
|---|---|---|---|
(+4,+6) | -5.000 | +2.813 | -2.188 |
(-4,+6) | -5.000 | -2.813 | -7.813 |
(+4,-6) | +5.000 | +2.813 | +7.813 |
(-4,-6) | +5.000 | -2.813 | +2.188 |
The governing compression is -7.81 ksi at (-4,+6); governing tension is +7.81 ksi at (+4,-6). At the other two corners the contributions partly cancel. Adding both extreme magnitudes everywhere would falsely report 7.81 ksi at all corners.
The neutral axis satisfies σ = 0:
y = (M_y I_x)/(M_x I_y) x = 0.844x.
Its slope is not M_y/M_x = 0.375 because bending stress also depends on unequal inertias. The neutral axis still passes through the centroid for pure elastic bending without axial force.
Eccentric Axial Loading
An axial force P applied away from the centroid produces axial stress plus bending. An offset in y produces moment magnitude P e_y about x; an offset in x produces P e_x about y, with signs determined from the force vector. Taking P algebraically positive in tension, the elastic stress becomes
σ = P/A - M_x y/I_x + M_y x/I_y
under the same sign convention. This superposition is a service-level elastic stress calculation, not a material-code strength interaction equation. Keep that distinction when later checking AISC 15th design resistance.
Unsymmetrical Sections
An angle, channel about non-symmetry axes, or arbitrary built-up section may have product of inertia I_xy ≠ 0. Then bending about the chosen geometric x axis can create stress variation involving both coordinates; the independent principal-axis equation cannot be used on those axes unchanged.
A robust exam method is:
- Find the centroid of the actual section.
- Calculate
I_x,I_y, and signedI_xyusing one coordinate convention. - Use the current PE Civil handbook's transformation relationship to find principal axes
x_p,y_pand propertiesI_xp,I_yp, whereI_xpyp = 0. - Resolve the applied moment vector into signed
M_xp,M_ypcomponents. - Transform each candidate point into
(x_p,y_p)coordinates consistently. - Evaluate
σ = -M_xp y_p/I_xp + M_yp x_p/I_ypat all extreme vertices and compare tension and compression.
The coordinate labels in step 5 must match the adopted principal-axis equation; sketching them prevents a silent swap. Alternatively, use the handbook's full coupled unsymmetrical-bending formula with the same I_xy sign convention. Do not rotate properties but leave moments or point coordinates unrotated.
Axis and Corner Audit
For a polygonal elastic section, stress is a linear plane, so an extreme value occurs at a boundary extreme point; check all relevant vertices rather than only “top” and “bottom.” Mark each term + or − before calculating its magnitude. Verify that stress changes linearly, the zero-stress line is plausible, and opposite corners of a centrally symmetric section have opposite stresses under pure bending.
For a 2026 exam, use the PE Civil handbook active for the test date and AISC Steel Construction Manual, 15th edition, where steel provisions are required. The April 2027 edition set is future-only.
In the worked 8-in by 12-in rectangle, which pair correctly identifies the governing pure-bending corner stresses?
What is the soundest procedure when an unsymmetrical section has nonzero product of inertia about the initially chosen centroidal axes?
A compressive axial load is applied with offsets in both centroidal directions. Which mapping of offset to bending-axis magnitude is correct?