From Textbook Formulas to Design Decisions
Mechanics of materials is a subject that engineering students study carefully and working designers apply imperfectly. The gap between the textbook version — controlled boundary conditions, idealized materials, static loads — and the real version — uncertain loads, material variability, stress concentrations, dynamic effects — is where design failures live.
This article does not repeat the textbook derivations. It focuses on the concepts that practicing designers must understand to use material strength data correctly: the distinction between yield and ultimate strength, how safety factors are selected and applied, how allowable stress is calculated, and where the standard approaches break down.
Yield Strength vs. Ultimate Tensile Strength
The stress-strain curve of a ductile metal shows two critical points that appear in material data sheets:
- Yield strength (Sy or Rp0.2): the stress at which permanent plastic deformation begins. The 0.2% offset convention is used for materials without a sharp yield point — it is the stress at which permanent strain equals 0.2%. Below this value, the material behaves elastically and returns to its original shape when unloaded. Above it, permanent deformation occurs.
- Ultimate tensile strength (Su or Rm): the maximum engineering stress the material sustains before fracture. Actual fracture occurs slightly below this value due to necking, but for design purposes, Su is the failure-by-fracture reference.
For a static load on a ductile part, yielding is typically the governing failure mode — a part that yields plastically has failed in the sense that it no longer meets dimensional requirements, even if it has not fractured. For a brittle material (cast iron, ceramic, some hardened steels), there is no significant plastic region; fracture occurs at the ultimate strength with little warning.
Safety Factors: What They Account For
A safety factor (also called a design factor or factor of safety, abbreviated SF or N) is a multiplier applied to the failure stress to obtain the allowable stress. The allowable stress is the maximum stress permitted in the design:
Allowable stress = Failure stress / Safety factor
The safety factor must be large enough to account for all uncertainties in the analysis, including:
- Uncertainty in load estimation (applied loads are often approximated)
- Material property variability (published values are typically minimum or mean; actual parts vary)
- Simplified stress analysis (real stress distributions differ from beam-theory assumptions)
- Stress concentrations not captured in the analysis (holes, fillets, surface scratches)
- Residual stresses from manufacturing (welding, machining, heat treatment)
- Environmental degradation (corrosion, temperature, fatigue)
A safety factor of 1 means the part is expected to be at the edge of failure under the design load — no margin for any of the uncertainties above. This is never acceptable.
Selecting Appropriate Safety Factors
| Application Condition | Typical Safety Factor (on Sy) | Notes |
|---|---|---|
| Well-known static load, ductile material, accurate analysis | 1.5–2.0 | Minimum for non-critical applications |
| Uncertain load, ductile material, simplified analysis | 2.0–3.0 | Most industrial machinery |
| Impact or dynamic load, ductile material | 3.0–4.0 | Shock load multiplier already in load estimate |
| Brittle material, static load | 3.0–4.0 (on Su) | No plastic redistribution; sudden fracture |
| Safety-critical application (pressure vessel, lifting gear) | 4.0 and above or per code | Governed by regulation (ASME, EN 13445, etc.) |
| Fatigue loading (cyclic stress) | 1.5–2.5 (on Se, endurance limit) | Separate fatigue analysis required; SF on static limit is insufficient |
Stress Concentrations: The Most Common Source of Premature Failure
The theoretical stress concentration factor (Kt) represents the ratio of the actual peak stress at a geometric discontinuity to the nominal stress calculated by simple formulas. A shaft with a shoulder fillet might have Kt = 2.5, meaning the actual stress at the fillet root is 2.5 times the nominal bending stress. A safety factor of 2 on nominal stress provides no margin if Kt = 2.5.
For static loads on ductile materials, stress concentrations are partially relieved by local yielding — the material yields at the peak stress point, redistributes load, and the part may survive if overall equilibrium is maintained. This is why theoretical stress concentration factors are sometimes not applied in full for ductile static loading.
For fatigue loading, stress concentrations are critical and must always be accounted for. The fatigue stress concentration factor (Kf) is related to Kt but modified by the material’s notch sensitivity. High-strength steels are more notch-sensitive than mild steels — a stress concentration that a mild steel shaft survives indefinitely may cause a high-strength steel shaft to fail in fatigue.
Combined Loading: When Bending and Torsion Act Together
Real machine shafts typically carry both bending moments and torsional loads simultaneously. The principal stresses from combined loading must be combined using a failure criterion — von Mises (distortion energy) is the standard choice for ductile metals:
von Mises stress = sqrt(σ² + 3τ²)
where σ is the normal (bending) stress and τ is the shear (torsional) stress. The von Mises stress is compared to the yield strength (with the appropriate safety factor). Using the maximum normal stress or maximum shear stress alone under combined loading gives unconservative or overly conservative results depending on the load combination.
FAQ
Q: Is it conservative to use ultimate tensile strength as the basis for safety factor calculation instead of yield strength?
For ductile materials under static loading, using ultimate strength as the failure criterion is unconservative — the part will have permanently deformed long before it fractures. Always use yield strength as the failure criterion for ductile materials under static loads. Ultimate strength is the appropriate basis for brittle materials, for fracture mechanics calculations, and for certain code-based calculations where regulations specify it.
Q: How do I handle a design where the load is partially unknown?
Estimate the maximum possible load using the worst credible scenario, not the average or expected load. Then apply a safety factor appropriate for uncertain loading (2.5–3.5 for most mechanical applications). Document your load assumptions clearly. If operating conditions change later and loads increase beyond the assumed maximum, the documented assumption alerts the responsible engineer to reassess the design.
Q: My FEA result shows stress peaks well above yield in a small region near a hole. Should I be concerned?
Possibly, but not necessarily. In a ductile material under static load, localized stress above yield in a small region is often acceptable — the material yields, the stress redistributes, and the part reaches equilibrium. The concern arises if: the yielding region is large enough to change overall geometry, the load is cyclic (fatigue), the material is brittle, or the part operates at elevated temperature where creep is a factor. For static ductile applications, check the volume of material above yield and confirm that the overall load path remains intact.
コメント