Static Failure Theories and Safe Design

Introduction

Machine components rarely experience one simple load. Shafts transmit torque while bending, pressure vessels carry membrane stress and local loads, and bolted joints may experience tension, shear, and bearing stresses simultaneously. Safe design therefore requires a rational method for converting a multiaxial stress state into a failure prediction.

A static failure theory provides that method. It compares the actual stress state with a material-dependent failure condition, usually expressed through a factor of safety. For ductile materials, yielding is normally the governing concern; for brittle materials, fracture by tensile or compressive stress is often more appropriate. The maximum-normal-stress, maximum-shear-stress, and distortion-energy theories are central tools for these decisions.

This chapter addresses components subjected to static or slowly varying loads. Fatigue, creep, buckling, impact, and fracture mechanics require additional methods and are not replaced by static theories.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Describe the difference between yielding, fracture, and serviceability failure.
2. Determine principal stresses and maximum shear stress for common plane-stress states.
3. Apply the maximum normal stress, maximum shear stress, and distortion-energy theories.
4. Select a suitable theory for ductile and brittle materials.
5. Calculate dimensions or factors of safety under combined loading.
6. Recognize assumptions, limitations, and professional responsibilities in safe design.

1. Failure, Strength, and Factor of Safety

Failure does not always mean complete separation. A component may be considered failed when it yields excessively, loses dimensional accuracy, leaks, jams, becomes unstable, or cannot perform its intended function. A design criterion must therefore be linked to a defined limit state.

For a component with a predicted equivalent stress σ_eq and allowable strength S_allow, the factor of safety is commonly written as

where n is dimensionless. In yield-based design, S_allow is often the yield strength S_y; in fracture-based design, it may be the relevant ultimate tensile or compressive strength. The formula assumes that the chosen equivalent stress and material strength are compatible and measured under comparable conditions.

Alternatively, if a design factor n_d is specified, the requirement is

for a ductile material governed by yielding. Strength values must use consistent units, such as MPa with N and mm, or Pa with N and m.

Important design distinctions are:

• Yielding: permanent plastic deformation begins.
• Fracture: a crack propagates or the part separates.
• Serviceability failure: the part remains intact but no longer performs acceptably.
• Ultimate failure: the load-carrying function is lost.
• Factor of safety: the ratio of a selected limiting strength to the predicted demand.

The factor of safety is not a substitute for engineering judgment. Uncertainty in loading, material properties, manufacturing, stress concentration, corrosion, temperature, and consequences of failure must all be considered.

2. Stress Description for Combined Loading

At a point in a loaded body, normal stresses act perpendicular to a plane and shear stresses act within the plane. For a plane-stress condition, the nonzero components are σ_x, σ_y, and τ_xy; the out-of-plane stresses σ_z, τ_xz, and τ_yz are assumed negligible.

The principal stresses are the normal stresses on planes where shear stress is zero. For plane stress,

The third principal stress is σ_3 = 0 when plane stress is an adequate idealization. The ordering convention is σ_1 ≥ σ_2 ≥ σ_3.

The maximum in-plane shear stress is the radius of the Mohr-circle representation:

For a complete three-dimensional state, the absolute maximum shear stress is

where σ_max and σ_min are the largest and smallest of the three principal stresses. This distinction matters: in plane stress, σ_3 = 0 must be included when identifying the absolute maximum shear stress.

The mean or hydrostatic stress is

Hydrostatic stress changes volume but, in the idealized yielding behavior of many ductile metals, contributes little to yielding. Distortional stress changes shape and is central to the von Mises theory.

Common stress components in machine elements

For a circular shaft of diameter d subjected to torque T, the torsional shear stress at radius r is

where J = πd⁴ ÷ 32 is the polar second moment of area for a solid circular section. At the outer surface r = d ÷ 2,

For a solid circular shaft subjected to bending moment M, the surface bending stress is

These equations assume elastic behavior, a circular prismatic section, and loads applied in a manner consistent with elementary beam and torsion theory. Stress concentrations, keyways, shoulders, and nonuniform geometry require local stress modification.

3. Maximum Normal-Stress Theory

The maximum-normal-stress theory, often called the Rankine theory, predicts failure when the largest principal stress reaches the relevant uniaxial strength. For a tensile or compressive brittle material, the design condition may be expressed as

and

where S_ut is ultimate tensile strength, S_uc is ultimate compressive strength, and n_d is the design factor. The appropriate tensile or compressive limit is selected according to the sign of the principal stress.

A more general statement is that failure is predicted when any principal stress reaches its corresponding limiting strength. For materials with approximately equal tensile and compressive strengths, a single magnitude limit may be used cautiously.

Applicability

The maximum-normal-stress theory is historically associated with brittle materials because brittle fracture is strongly related to normal stress and because brittle materials often have different tensile and compressive strengths. It is generally conservative or inaccurate for ductile metals under multiaxial loading because it ignores the beneficial or harmful interaction of the other principal stresses in a way that does not match observed yielding.

For a uniaxial tensile test, the theory correctly reduces to σ_1 = S_ut at failure. For pure shear, principal stresses are +τ and −τ, so the predicted tensile limit is reached at τ = S_ut. This prediction is often too strong for ductile materials and illustrates why theory selection matters.

4. Maximum-Shear-Stress Theory

The maximum-shear-stress theory, or Tresca theory, predicts yielding when the maximum absolute shear stress equals the shear stress at yield in a uniaxial tensile test. In a uniaxial tensile test at yield, the principal stresses are S_y, 0, and 0. Therefore,

The Tresca yielding condition is consequently

or, with a design factor,

Using ordered principal stresses, this becomes

for the safe condition. The intermediate principal stress does not appear explicitly in this expression, although it influences the ordering.

Interpretation

Tresca is physically connected to slip along planes of maximum shear stress. It is relatively simple and generally conservative for ductile metals. In pure shear, σ_1 = τ, σ_2 = 0, and σ_3 = −τ, so

and yielding is predicted at τ = S_y ÷ 2. Tresca is particularly useful for shafts, keys, pins, and other ductile components in which torsion is significant.

Its yield surface in principal-stress space has a hexagonal shape, whereas the von Mises surface is circular in the deviatoric plane. Thus, Tresca and von Mises predictions are identical in uniaxial tension but differ under many combined stress states.

5. Distortion-Energy Theory

The distortion-energy theory, commonly called the von Mises theory, states that yielding begins when the distortion energy per unit volume reaches the value at yield in a uniaxial tensile test. In terms of principal stresses, the equivalent stress is

The safe design condition is

For plane stress with σ_3 = 0,

This expression is dimensionally consistent because every term has units of stress squared.

Derivation from stress decomposition

The normal stress tensor can be separated conceptually into hydrostatic and deviatoric parts. The hydrostatic part is σ_m times the identity tensor and produces volumetric change. The deviatoric part produces shape change. The distortion-energy density is proportional to the second invariant of the deviatoric stress tensor.

For principal stresses, the distortion-energy density is

where u_d is distortion energy per unit volume and G is the shear modulus. Equating this value to the corresponding uniaxial-yield value produces the von Mises criterion. This derivation assumes small elastic strains, isotropic material behavior, and a yield response that is reasonably represented by distortional energy.

For pure shear, σ_1 = τ, σ_2 = 0, and σ_3 = −τ. Substitution gives

Thus, von Mises predicts yielding at

which is about 15.5% higher than the Tresca prediction S_y ÷ 2. Both are accepted engineering approximations; Tresca is more conservative.

6. Comparing the Principal Static Theories

No theory is universally correct. Material test data, loading history, geometry, temperature, residual stress, and failure consequences determine the appropriate method. For ductile steels and aluminum alloys under static loading, von Mises is widely used in modern design codes and finite-element post-processing. Tresca remains valuable as a conservative check. For brittle cast iron, ceramics, and many hardened materials, normal-stress-based criteria are often more suitable, although brittle failure can also depend on flaws and fracture toughness.

7. Worked Example: Shaft Under Combined Bending and Torsion

Problem

A solid steel shaft has diameter d = 30 mm. At a critical surface location it carries a bending moment M = 1.20 kN·m and torque T = 0.80 kN·m. The yield strength is S_y = 360 MPa. Determine the factor of safety using Tresca and von Mises theories. Neglect stress concentrations and assume static loading.

Solution

Convert moments:

M = 1.20 × 10⁶ N·mm, T = 0.80 × 10⁶ N·mm.

The surface bending stress is

Since d³ = 30³ = 27,000 mm³,

σ_b = (32 × 1.20 × 10⁶) ÷ (π × 27,000) = 452.7 MPa.

The torsional shear stress is

τ = (16 × 0.80 × 10⁶) ÷ (π × 27,000) = 150.9 MPa.

At the surface, take σ_x = 452.7 MPa, σ_y = 0, and τ_xy = 150.9 MPa. The principal stresses are

Thus,

σ_1 = 497.6 MPa, σ_2 = −44.9 MPa, σ_3 = 0 MPa.

The ordered values are σ_1 = 497.6 MPa, σ_2 = 0 MPa, and σ_3 = −44.9 MPa.

For Tresca,

Therefore,

The factor of safety is below one; the shaft is predicted to yield.

For von Mises,

because σ_y = 0. Hence,

σ_vm = √(452.7² + 3(150.9)²) = 522.6 MPa,

and

Both criteria reject the design. A redesign could increase the diameter, reduce the loads, change the material, or introduce a more efficient geometry. Increasing diameter is particularly effective because bending and torsional surface stresses for a solid circular shaft vary with 1 ÷ d³, subject to the validity of the elastic model and the absence of new stress concentrations.

8. Design for a Required Diameter

For a solid shaft under surface bending stress σ_b and torsional stress τ, the von Mises condition is

Using the circular-shaft relations gives

Therefore, the minimum theoretical diameter is

Here M and T must be expressed in consistent force-length units, S_y in force per area, and d in the corresponding length unit. This is an ideal elastic sizing equation. A practical design must also check deflection, angle of twist, fatigue, keyways, shoulders, bearing loads, critical speed, manufacturability, and applicable standards.

The corresponding Tresca sizing equation can be obtained from the principal stresses or, for the combined surface state, from the condition 2τ_max ≤ S_y ÷ n_d. It may produce a slightly larger diameter than von Mises because it is generally more conservative.

9. Brittle and Ductile Design Selection

Ductile components

Ductile metals usually exhibit substantial plastic deformation before rupture. For static loading:

• Use von Mises when a realistic yielding estimate is desired.
• Use Tresca when a conservative yield check is preferred.
• Include stress concentrations through local elastic stresses or validated concentration factors.
• Do not assume that ductility eliminates fracture; flaws, low temperatures, high strain rates, and welding can reduce ductile behavior.

Brittle components

Brittle materials may fracture with little visible plastic deformation. Their tensile and compressive strengths can differ substantially. A normal-stress approach should distinguish the signs of the principal stresses. For cast iron, ceramics, and brittle polymers, additional considerations include:

• tensile stress concentrations and surface defects;
• notch sensitivity;
• fracture toughness and crack size;
• residual stresses;
• moisture, temperature, and environmental degradation;
• statistical scatter in strength.

A component that passes a nominal principal-stress check may still be unsafe if a crack-like defect is present. In such cases, linear elastic fracture mechanics or a material-specific brittle-failure criterion is required.

10. Stress Concentration and Local Failure

The nominal formulas for beams and shafts describe idealized stress distributions. A shoulder fillet, keyway, hole, thread, weld toe, or abrupt section change produces a local concentration. The elastic concentration factor is defined by

for normal stress, with an analogous definition for shear. The maximum local elastic stress is therefore K_tσ_nominal. The factor depends on geometry and loading mode and must be obtained from reliable charts, standards, analytical solutions, or validated numerical analysis.

For ductile components, local yielding may redistribute stress, so the elastic K_t is not automatically the final plastic response. For fatigue, a fatigue concentration factor may differ from K_t because notch sensitivity and cyclic behavior are involved. Static and fatigue concentration factors must not be interchanged without justification.

11. Case Application: Preliminary Design of a Drive Shaft

The following is an illustrative engineering case, not a certified design.

Background

A gearbox output shaft transmits torque to a chain sprocket. The sprocket produces a transverse force, causing bending. The shaft material is normalized alloy steel with a specified yield strength of 420 MPa. The design factor for static yielding is 2.0.

Problem

At the critical section, the estimated bending moment is 0.95 kN·m and torque is 0.60 kN·m. A 25-mm trial diameter is proposed. A keyway is located near the section, and a conservative elastic concentration factor of 1.6 is estimated for the combined local stress during preliminary sizing.

Analysis

For d = 25 mm,

σ_b = 32(0.95 × 10⁶) ÷ [π(25³)] = 619.0 MPa,

τ = 16(0.60 × 10⁶) ÷ [π(25³)] = 195.6 MPa.

Without concentration, the von Mises stress is

σ_vm = √(619.0² + 3(195.6)²) = 706.8 MPa.

Applying the preliminary local factor to the stress components gives an estimated local equivalent stress of

σ_vm,local ≈ 1.6(706.8) = 1131 MPa.

The allowable yield stress is S_y ÷ n_d = 420 ÷ 2 = 210 MPa. The trial diameter is plainly unacceptable.

For preliminary redesign, neglecting the concentration temporarily, the diameter required by von Mises is

Substitution gives d ≥ 35.8 mm for the ideal smooth shaft. Because the keyway remains, selecting a standard diameter somewhat larger than this value and reducing the concentration through a generous fillet or alternative torque-transfer method would be appropriate. The final design must recalculate the actual geometry rather than simply multiplying a smooth-shaft result by a guessed factor.

Outcome and learning points

The preliminary study rejects the 25-mm shaft, identifies the keyway as a major risk, and shows that a smooth-shaft calculation alone can be misleading. The final design process should include a detailed stress analysis, shaft deflection and twist limits, fatigue from rotating bending, key and hub checks, bearing reactions, assembly requirements, and inspection provisions.

12. Limitations and Professional Practice

Static failure theories are models, not universal laws. Their limitations include:

1. Material idealization: isotropy, homogeneity, and a single yield strength may not represent forged, cast, welded, or additively manufactured material.
2. Load uncertainty: actual torque, shock, misalignment, and overload may exceed nominal values.
3. Residual stress: welding, heat treatment, machining, and assembly can alter the local stress state.
4. Temperature and environment: strength and ductility can change with temperature, corrosion, hydrogen, or chemicals.
5. Multiaxial fracture: brittle failure may depend on crack orientation and fracture toughness rather than only nominal stress.
6. Elastic analysis near yielding: once substantial plasticity occurs, linear elastic stress fields are no longer sufficient.
7. Numerical modeling: finite-element results are sensitive to mesh quality, boundary conditions, contact assumptions, and singularities.

Engineering ethics requires conservative communication of uncertainty, traceable calculations, independent review for safety-critical parts, and compliance with applicable standards. A designer must never report a factor of safety without stating the failure criterion, material property, load case, geometry assumptions, and exclusions from the analysis.

13. Contemporary Tools and Best Practices

Modern finite-element software reports principal stresses, maximum shear stress, and von Mises stress. These outputs are useful only when the model is physically credible. Best practice is to:

• verify reactions and equilibrium;
• compare simple regions with hand calculations;
• refine the mesh near stress gradients while recognizing singularities;
• distinguish averaged, nodal, and integration-point stresses;
• examine multiple load combinations;
• use material data appropriate to temperature and manufacturing condition;
• validate important predictions with testing or established analytical solutions.

AI-assisted design and optimization can search geometry and material combinations rapidly, but generated recommendations require engineering verification. AI does not determine whether a load case is complete, whether a material certificate is valid, or whether a failure consequence is acceptable. Human accountability remains essential.

Chapter Summary

Static failure theories convert combined stress states into usable design criteria. Principal stresses describe the normal stresses on planes of zero shear, while the maximum absolute shear stress depends on the largest and smallest principal stresses, including the zero third principal stress in plane-stress analysis.

The maximum-normal-stress theory compares principal stresses with tensile and compressive strengths and is most appropriate as a basic model for brittle materials. Tresca theory predicts ductile yielding when the maximum shear stress reaches S_y ÷ 2 and is generally conservative. Von Mises theory compares distortion energy with the uniaxial yield condition and is widely used for ductile metals. Neither theory replaces attention to stress concentrations, defects, fatigue, fracture toughness, serviceability, or standards.

Safe design requires a stated failure mode, reliable material data, complete loading information, consistent units, an appropriate factor of safety, and verification of geometry-specific effects. A sound preliminary calculation is the beginning of design, not its conclusion.

Key Terms

Principal stress: Normal stress on a plane where shear stress is zero.

Plane stress: Stress state in which out-of-plane stresses are negligible.

Hydrostatic stress: Equal normal stress in all three principal directions.

Deviatoric stress: Stress component associated primarily with shape change.

Yielding: Onset of permanent plastic deformation.

Brittle fracture: Fracture occurring with little macroscopic plastic deformation.

Ductile material: Material capable of substantial plastic deformation before fracture.

Factor of safety: Ratio of selected limiting strength to predicted demand.

Maximum normal stress theory: Criterion based on the largest tensile or compressive principal stress.

Rankine theory: Another name for the maximum-normal-stress theory.

Maximum shear stress theory: Criterion that predicts yielding from the largest principal-stress difference.

Tresca theory: Another name for the maximum-shear-stress theory.

Distortion energy: Elastic energy associated with shape change.

Von Mises stress: Equivalent stress derived from distortional energy.

Stress concentration: Local amplification of stress caused by geometric discontinuity.

Nominal stress: Stress calculated using an idealized gross geometry.

Principal-stress ordering: Arrangement σ_1 ≥ σ_2 ≥ σ_3.

Serviceability failure: Loss of required function without necessarily causing fracture.

Fracture toughness: Resistance to crack extension under specified conditions.

Load case: Defined combination of forces, moments, pressures, and constraints used in design.

Review Questions

Short Answer Questions

1. What is the difference between yielding and fracture?
2. Why must the zero third principal stress be considered in a plane-stress problem when finding absolute maximum shear stress?
3. State the Tresca yielding condition in terms of ordered principal stresses.
4. What physical component of stress is emphasized by the von Mises theory?
5. Why are stress concentrations important in static design?

Long Answer Questions

1. Explain the assumptions and typical applications of the maximum-normal-stress, Tresca, and von Mises theories.
2. Derive the plane-stress expression for von Mises equivalent stress from the principal-stress form.
3. Compare the predictions of Tresca and von Mises theories for pure shear.
4. Discuss why brittle components require attention to tensile and compressive strengths separately.
5. Describe a professional workflow for validating a finite-element static failure analysis.

Analytical / Application Questions

1. A solid shaft is subjected to bending and torsion. Develop a sizing equation using the von Mises criterion and explain every assumption.
2. For principal stresses σ_1 = 180 MPa, σ_2 = 40 MPa, and σ_3 = −60 MPa, calculate the Tresca stress measure and von Mises stress, then explain which criterion is more conservative relative to a specified yield strength.
3. A cast component passes a von Mises check but contains a sharp surface defect. Explain why the design may still be unsafe and identify a more appropriate analysis approach.

Multiple Choice Questions with Answers

Question 1

For a plane-stress state, which value must be included when determining the absolute maximum shear stress?

A. The temperature

B. The zero third principal stress

C. The elastic modulus only

D. The nominal diameter only

Correct Answer: B. The zero third principal stress

Question 2

The Tresca theory predicts yielding when which condition is reached?

A. σ_1 + σ_2 + σ_3 = 0

B. σ_vm = 0

C. σ_1 − σ_3 = S_y

D. σ_1 − σ_3 = S_y ÷ n_d for a design check

Correct Answer: D. σ_1 − σ_3 = S_y ÷ n_d for a design check

Question 3

Which theory is most directly associated with distortion energy?

A. Von Mises theory

B. Maximum normal stress theory

C. Euler buckling theory

D. Coulomb friction theory

Correct Answer: A. Von Mises theory

Question 4

For pure shear stress τ, the von Mises equivalent stress is:

A. τ ÷ √3

B. τ

C. √3τ

D. 2τ

Correct Answer: C. √3τ

Question 5

For a solid circular shaft, the maximum torsional shear stress is:

A. 8T ÷ (πd²)

B. 16T ÷ (πd³)

C. 32T ÷ (πd³)

D. T ÷ (πd⁴)

Correct Answer: B. 16T ÷ (πd³)

Question 6

Which theory is generally most suitable as a basic static model for brittle materials?

A. Maximum normal stress theory

B. Von Mises theory only

C. Tresca theory only

D. S-N fatigue theory

Correct Answer: A. Maximum normal stress theory

Question 7

A ductile steel has S_y = 300 MPa and a calculated von Mises stress of 150 MPa. With no additional design factor specified, the yield-based factor of safety is:

A. 0.5

B. 1.0

C. 2.0

D. 450

Correct Answer: C. 2.0

Question 8

The stress concentration factor K_t is defined as:

A. Nominal stress divided by maximum stress

B. Maximum elastic local stress divided by nominal stress

C. Yield strength divided by ultimate strength

D. Torque divided by diameter

Correct Answer: B. Maximum elastic local stress divided by nominal stress

Question 9

In a uniaxial tensile test at yield, the Tresca shear stress at yield is:

A. S_y

B. S_y ÷ 2

C. S_y ÷ √3

D. 2S_y

Correct Answer: B. S_y ÷ 2

Question 10

For a smooth solid shaft under bending and torsion, increasing diameter is highly effective because the nominal surface stresses vary approximately with:

A. d

B. 1 ÷ d

C. 1 ÷ d²

D. 1 ÷ d³

Correct Answer: D. 1 ÷ d³