4.4Strength Requirements

NASA-STD-5020B, Requirements for Threaded Fastening Systems in Spaceflight Hardware

Last Updated 29 Oct 2026

Hover any section to copy its link or text, ask Copilot about it, or bookmark it.

4.4.1 Ultimate Design Loads

[TFSR 8] Threaded fastening system hardware shall withstand ultimate design loads (limit load times the ultimate factor of safety and fitting factor) without failure when subjected to:

a. The accompanying service environments (for example, temperature) and
b. A coefficient of friction between clamped parts equal to zero (applicable only to verification by analysis, not applicable to verification by test) (see TFSR 13).

Withstanding ultimate design loads can be demonstrated by test or by analysis where a non- negative margin of safety is calculated.

Use of friction to react applied shear forces between the clamped parts is not permissible for ultimate load assessment.

[Rationale: Withstanding the ultimate design loads validates the integrity of the structure against failure in the expected service environment. Under ultimate load conditions, it is assumed that static friction is overcome and shear forces are reacted through bearing contact.]

Failure, as used in TFSR 8, is the inability to meet any performance requirements applicable at ultimate load. Examples of failure include rupture of the joint and leaking of a joint that can credibly lead to a catastrophic hazard.

Assessment for ultimate design loads addresses potential rupture in all elements of the threaded fastening system, including the fastener, the internally threaded part such as a nut or an insert, and the clamped parts. Assessment of a procured item such as a nut or a threaded insert should be based on the strength specified for that item rather than on thread-stripping analysis. Such items can expand under load, reducing the thread engagement areas.

Each internally threaded part used for design should have either a specified and controlled allowable tensile load, or an allowable tensile load derived from dedicated testing. Many threaded inserts have two allowable tensile loads that should be considered: the minimum allowed tensile load capability of the insert internal threads, and the minimum allowed tensile load for pullout of the insert from the parent material. One or both may be provided in the insert specification (or the procurement specification). The lower value should be used for strength analysis.

When performing ultimate analysis, the analysis should account for load redistribution due to geometric design features, localized yielding, separation, or slipping. The loads acting on individual members of the fastened joint are dependent on features of the joint design (e.g., fastener to bearing surface clearance), materials (e.g., a bearing-critical versus shear-critical joint design), and friction (resistance to slipping is highest under a preloaded member and drops with distance from the member). Examples of load redistribution are shown in Appendix A.12.

An insert’s allowable pull-out load depends on the material in which the insert is installed (parent material). The most common failure mode associated with pull-out is shear of internal threads in the parent material. The specifications (or procurement specifications) for most threaded inserts define how the allowable pull-out load is calculated such as by multiplying a specified minimum shear engagement area by the allowable ultimate shear stress of the parent material. Such an allowable pull-out load applies when the insert is installed in a solid, homogenous material. For inserts installed in nonhomogeneous or nonmetallic materials or in sandwich panels, allowable pull-out loads should be derived from test. Nuts should be limited to the load rating of the nut.

The margin of safety indicates how much the applied load can increase before the criteria are no longer satisfied. Simplistic equations for calculating the margin of safety for ultimate under axial load are

MS_u = \frac{P_{tu-allow}}{FF \cdot FS_u \cdot P_{tL}} - 1 \qquad (Eq. 6)

when separation occurs before rupture, and

MS_{u} = \frac{P'_{tu}}{FF \cdot FS_{u} \cdot P_{tL}} - 1 \tag{Eq. 7}

when rupture occurs before separation, where P′tu is the applied tensile load that causes the fastener load to exceed the fastening system’s allowable ultimate tensile load if rupture occurs before separation, and Ptu-allow is the allowable ultimate load for the fastening system.

Linear theory, empirical, or semi-empirical methods can be used to assess ultimate tensile strength of the fastening system and if rupture occurs before separation. Appendix A.6 provides further detail on this approach. With linear theory, the tensile load in a preloaded bolt, Ptb ,

increases proportionally with the applied tensile load per Eq. 8 until either any part in the fastening system ruptures or separation occurs.

P_{tb} = P_{p} + n\phi P_{t} \qquad (Eq. 8)

where Pp is the preload, Pt is the applied tensile load, n is the load-introduction factor (see Appendix A.4) that depends on the joint design and accounts for where the load is applied to the joint, and is the stiffness factor. The stiffness factor is

\phi = \frac{k_{b}}{k_{b} + k_{c}} \tag{Eq. 9}

where kb is the stiffness of the bolt and kc is the stiffness of the clamped parts local to the fastener.

Based on Eq. 8 and the assumption of maximum preload, Pp-max , the applied tensile load that causes the bolt load to exceed the allowable ultimate tensile load for the fastening system, P′tu , is

P'_{tu} = \frac{1}{n\phi} (P_{tu-allow} - P_{p-max}) \tag{Eq. 10}

and the linearly projected load that causes separation when at maximum preload is

P'_{sep} = \frac{P_{p-max}}{1 - n\phi} \qquad (Eq. 11)

If P′sep is less than P′tu , linear theory predicts that separation would occur before rupture, and the ultimate margin of safety for tensile loading is calculated per Eq. 6. Conversely, Eq. 7 should be used when P′sep is greater than P′tu , as linear theory predicts rupture would occur before separation.

The allowable ultimate shear load for a fastener depends on whether or not the threads are in the shear plane. If threads are not in the shear plane, the allowable ultimate shear load per shear plane is commonly assumed to be half the value given in the fastener specification for double-shear joints, when applicable, or is calculated by

P_{su-allow} = \frac{\pi D^{2} F_{su}}{4} \qquad (Eq. 12)

where Fsu is the allowable ultimate shear strength for the fastener material. If threads are in the shear plane, the allowable ultimate shear load for a fastener is calculated by

P_{su-allow} = F_{su} A_{m} \quad (Eq. 13)

where Am is the minimum minor-diameter area for the fastener threads. The ultimate margin of safety for shear loading of a fastener is

MS_{u} = \frac{P_{su-allow}}{FF \cdot FS_{u} \cdot P_{sL}} - 1 \tag{Eq. 14}

where $P_{sL}$ is the limit shear load acting on the shear plane.

4.4.2 Yield Design Loads

[TFSR 9] Threaded fastening system hardware shall withstand yield design loads (limit load times the yield factor of safety and fitting factor) without detrimental yielding or detrimental deformation when subjected to the accompanying service environments (for example, temperature).

[Rationale: Withstanding yield design loads validates the integrity of the structure against yielding or deformations that could adversely affect the functionality of critical systems in the expected service environment.]

Withstanding yield design loads can be demonstrated by test or by analysis where a non-negative margin of safety is calculated.

Yield strength analysis is not required if there is no detrimental yielding or detrimental deformations at the yield design load. The assessment for yield design loads will address all elements of the threaded fastening system, including the fastener, the internally threaded part such as a nut or an insert, and the clamped parts.

Examples of detrimental deformations include

Fastener yielding causing the joint to separate under an applied tensile load that is less than the design separation load.
Fastener yielding causing the joint to suffer detrimental slip under an applied shear load that is less than the applicable design shear load.
• Deformations that degrade mission performance.
Some other design-specific reason exists for why fastener yielding is detrimental (e.g., any fastener yielding that adversely affects the form, fit, or function of the design).

In a typical joint with preload generated by torque control, fastener yielding under combination of preload and applied load is not detrimental for separation (see first bullet above), given the separation analysis criteria specified in section 4.4.3. Rationale and exceptions for the above statement, as well as a discussion of the effects of fastener yielding on joint slip (see second bullet above), are provided in Appendix A.9.

When performing yield analysis, the analysis should account for load redistribution due to geometric design features, localized yielding, separation, or slipping. The loads acting on individual members of the fastened joint are dependent on features of the joint design (e.g., fastener to bearing surface clearance), materials (e.g., a bearing-critical versus shear-critical joint design), and friction (resistance to slipping is highest under a preloaded member and drops with distance from the member). Examples of load redistribution are shown in Appendix A.12.

In a linear analysis, the margin of safety indicates how much the applied load can increase before the minimum design factor of safety criterion is no longer satisfied. In some instances, it may be difficult to demonstrate a positive margin of safety under yield design loads. Care must be taken when calculating the margin of safety to avoid being overly conservative in the analysis. For example, due to the geometry of fastener threads, even under nominal initial preload, a portion of material at thread roots could be above yield. It is not intended that highly localized peak stresses that may be predicted by computer-aided methods of analysis and may be blunted by confined yielding must demonstrate a positive margin of safety.

Simplistic equations for calculating the margin of safety for yield under axial load are

MS_{y} = \frac{P_{ty-allow}}{FF \cdot FS_{y} \cdot P_{tL}} - 1 \tag{Eq. 15}

when separation occurs before yield, and

MS_{y} = \frac{P'_{ty}}{FF \cdot FS_{y} \cdot P_{tL}} - 1 \tag{Eq. 16}

when yield occurs before separation, where

P'_{ty} = \frac{1}{n\phi} (P_{ty-allow} - P_{p-max}) \qquad (Eq. 17)

where P′ty is the applied tensile load that causes the fastener load to exceed the fastening system’s allowable yield tensile load if yielding occurs before separation, PtL is the limit tensile load, Pty-allow is the allowable tensile load of the material, n is the load-introduction factor that depends on the joint design and accounts for where the load is applied in the joint, and is the stiffness factor. See Appendix A.4 for an explanation of the load-introduction and stiffness factors. The quantity Pp-max is the maximum initial preload (see section 4.3.1). The allowable yield tensile load can be estimated using the equation below when a value is not explicitly defined in the corresponding fastener specification or relevant test data is unavailable, as

P_{ty-allow} = \left( \frac{F_{ty}}{F_{tu}} \right) P_{tu-allow} \tag{Eq. 18}

where Ptu-allow is the allowable ultimate load of the fastener in tension, and Fty and Ftu are the yield and ultimate strength of the fastener material, respectively. Ptu-allow will have considerations for the fastener geometric features and will typically be the capability of the threaded section of the fastener. Eq. 15 and Eq. 16 are not applicable to all load conditions and are provided as examples only.

Because shear yield strength is not a standard material property, when evaluating the margin of safety under yield design loads and performing combined loads analysis, the normal and shear components of stress should be transformed into principal stresses; and a failure theory (e.g., von Mises or Tresca) should be used that is compatible with the concept of tensile yield strength.

4.4.3 Separation Loads

[TFSR 10] Threaded fastening system hardware shall withstand design separation loads (limit load times the separation factor of safety and fitting factor) without loss of compression between the joint members or detrimental deformation due to separation when subjected to the accompanying service environments (for example, temperature).

[Rationale: Withstanding the separation design loads validates the integrity of the structure against the detrimental effects of separation in the expected service environment.]

Withstanding design separation loads can be demonstrated by test or by analysis where a non- negative margin of safety is calculated.

Separation is not strictly prohibited in all applications. Separation may be permissible when:

Separation does not cause detrimental deformations; and
Separation does not cause a critical or catastrophic hazard; and
Strength requirements in sections 4.4.1 and 4.4.2 are met in the separated condition; and
The life requirement in section 4.5 is met when accounting for the separated condition.

Examples of detrimental separation deformations include:

Reduction in compression of a seal or gasket that enables a liquid or gas to leak at an unacceptable rate; or
Separation that interrupts a critical electrical path; or
Separation that degrades the thermal contact conductance between mated parts; or
Separation that causes a change in stiffness or load path credibly leading to a load condition that causes detrimental yield or collapse of the structure; or
Separation that reduces the fatigue life of the fastener due to an increase in the amplitude of the cyclic load.

Separation analysis should account for load redistribution due to joint geometry, localized yielding, partial or full separation, or slipping. The loads acting on individual joint members are dependent on features of the joint design (e.g., fastener to bearing surface clearance), materials (e.g., a bearing-critical versus shear-critical joint design), and friction (resistance to slipping is highest under a preloaded member and drops with distance from the member). Examples of load redistribution are shown in Appendix A.12.

See Appendix A.11 for additional information on using minimum preload to calculate the margin of safety under separation loads. A simplistic equation to evaluate the margin of safety for separation, MSsep , is

MS_{sep} = \frac{P_{p-min}}{FF \cdot FS_{sep} \cdot P_{tL}} - 1 \qquad (Eq. 19)

where FSsep is the separation factor of safety per section 4.2.3. The equation is applicable to systems under axial loading only. Other equations or methods may be used to evaluate the margin of safety when combined loading is considered.

When a joint maintains a seal (e.g., to maintain pressure or contain a fluid), the analysis may be complex, particularly if the joint has a seal or gasket. In such cases, Eq. 19 does not accurately predict the margin of safety for separation; and the analysis approach depends on the design of the joint and the seal.

It is advisable to assume that minimum preload exists for all the fasteners in the joint to maximize robustness of the joint.

4.4.4 Combination of Loads

[TFSR 11] The limit, yield, ultimate, and separation loads shall account for interaction of the combined loading (simultaneously applied tensile, shear, and bending loads) and under all design environmental conditions.

[Rationale: Depending on the configuration of the fastened joint members and the applied mechanical loading, the members could experience a combination of forces and moments. Neglecting the combined effects of forces and moments that are physically reacted by joint members is non-conservative.]

Interaction equations for the ultimate condition are shown in this section. These interaction equations may be used, but other interaction equations based on test data or test-substantiated analysis methods may be substituted.

When assessing the strength of the fastener due to combined loading, if shear is not transferred across gaps or non-load carrying spacers, or if interference or close tolerance fits are used, then typically there is no need to account for bolt bending caused by the shear loading. However, if the shear is transferred across gaps or non-load carrying spacers, or if there are clearances between the bolt and joint, interaction of loads, including non-negligible bending, should be considered.

For fasteners under simultaneously applied tensile and shear loads, along with any applicable bending, analysis should account for interaction of the combined loading. Preload may be omitted in the interaction check at the ultimate condition.

For the case in which the full-diameter body is in the shear plane rather than the threads, satisfying either of the criteria given by Eq. 20 and Eq. 21 is acceptable for the ultimate condition. The criterion given by Eq. 20 should be satisfied when not accounting for plastic bending, or the criterion given by Eq. 21 should be satisfied when accounting for plastic bending.

\left(\frac{P_{su}}{P_{su-allow}}\right)^{2.5} + \left(\frac{P_{tu}}{P_{tu-allow}} + \frac{f_{bu}}{F_{tu}}\right)^{1.5} \leq 1 \qquad (Eq. 20)

\left(\frac{P_{su}}{P_{su-allow}}\right)^{2.5} + \left(\frac{P_{tu}}{P_{tu-allow}}\right)^{1.5} + \left(\frac{f_{bu}}{F_{bu}}\right) \leq 1 \tag{Eq. 21}

where Psu-allow is calculated using Eq. 12, Psu is the ultimate design shear load per shear plane, fbu is the design ultimate bending stress based on linear-elastic theory, Ftu is the allowable ultimate tensile stress, and Fbu is the allowable ultimate flexural stress (bending, that applies for ductile materials as a result of plastic bending. Use of Fbu , which depends on the material and the cross section, accounts for the additional moment-carrying capacity of a section in bending over and above the capacity calculated based on linear-elastic theory.

Inclusion of the bending stress term in the criteria given in Eqs. 20 and 21 is considered to be conservative. These criteria, with omission of the bending stress term, are based on tests of A- 286 3/8-24 (NAS1956C14) fasteners performed at NASA MSFC in 2010, as documented in NASA/TM-2012-217454, Aerospace Threaded Fastener Strength in Combined Shear and Tension Loading.

For the case in which the threads are in the shear plane, as is typically the case for a bolt going into a tapped hole or a threaded insert, satisfying either of the criteria given by Eqs. 22 and 23 is acceptable. The criterion given by Eq. 22 should be satisfied when not accounting for plastic bending, or the criterion given by Eq. 23 should be satisfied when accounting for plastic bending.

\left(\frac{P_{su}}{P_{su-allow}}\right)^{1.2} + \left(\frac{P_{tu}}{P_{tu-allow}} + \frac{f_{bu}}{F_{tu}}\right)^{2} \leq 1 \tag{Eq. 22}

\left(\frac{P_{su}}{P_{su-allow}}\right)^{1.2} + \left(\frac{P_{tu}}{P_{tu-allow}}\right)^{2} + \left(\frac{f_{bu}}{F_{bu}}\right) \leq 1 \qquad (Eq.\ 23)

where Psu-allow is calculated using Eq. 13. The shear capability of a fastener is greatest when the full diameter body is in the shear plane and is reduced when the threads are in the shear plane due to cross-sectional area reduction.

Note the difference in exponents used in Eqs. 22 and 23 (threads in shear plane) as compared with those used in Eqs. 20 and 21 (full diameter body in shear plane). Tensile and shear stresses peak at the same cross section when the threads are in the shear plane; when the full diameter body is in the shear plane, the tensile and shear stresses do not peak at the same cross section. Inclusion of the bending stress term in the criteria given in Eqs. 22 and 23 is considered to be conservative.

4.4.5 Inclusion of Preload in Yield and Ultimate Load Strength

[TFSR 12] If rupture occurs before separation, preload shall be included in the determination of the total tensile load in the preloaded bolt.

[Rationale: When joint members separate prior to rupture, the tensile load in the bolt is equal to the applied mechanical load; whereas if rupture occurs prior to separation, there is a contribution to failure from preload.]

When a fastened joint is completely separated prior to rupture, the total axial component of the load acting on the bolt is equal to the axial component of the applied load only. If rupture occurs before separation, preload also acts on the bolt and should be included when assessing the performance of the fastened joint. In this case, a factor of safety is not applied to preload. Whether separation would occur before rupture can be determined based on test or analysis. Refer to Appendix A.5 and A.6 for guidance.

4.4.6 Use of Friction at Limit and Yield Load

a. [TFSR 13] Use of friction to react shear loads shall only be permissible for analysis at limit load (alignment, fatigue, and fracture) or yield load.

Beneficial effects of friction are not permissible in analysis for ultimate load conditions.

[Rationale: When a bolt begins to yield under the combination of preload and applied load, preload relaxes and the shear force reacted by friction at the faying surfaces decreases. It is not conservative to assume that friction carries any of the shear load when assessing ultimate strength.]

b. [TFSR 14] Unless otherwise substantiated by test, the coefficient of friction for joint- slip analysis shall be no greater than:
(1) 0.20 for uncoated, non-lubricated metal surfaces that are cleaned by a qualified process and visibly clean at and after assembly.
(2) 0.10 for all other surfaces, including nonmetallic (coated or uncoated) surfaces and metallic surfaces that are coated with any substance, including lubricant, paint, and conversion coating.

[Rationale: The coefficient of friction is highly dependent on the materials in contact, environmental conditions, contact force distribution, and presence of coatings, lubricants, or contamination. Test data indicates that dry and wet conditions across a variety of designs can have a minimum value as reported above.]

When friction is used as a load path for shear loading, the analysis should show no slippage at the applicable design load (see Appendix A.10 for guidance on joint-slip analysis). If the margin of safety for joint slip is negative, then friction cannot be used in analysis as a load path for shear loading.

Note that positive margins of safety are required for ultimate design loads without reliance on friction to ensure high structural reliability regarding catastrophic failure.

When selecting the value of the coefficient of friction, values lower than those presented above should be used if the design warrants it. The requirement sets the maximum value allowed without test data. For example, in some designs such as with Teflon™-coated surfaces, the coefficient of friction may be less than 0.10. A value consistent with such a design should be used instead of the value of 0.10.

Higher values may be used with program- or project-approved testing in a relevant environment. When values are substantiated by test, the tests should mimic the environmental conditions expected in the service environment and using materials, coatings, and cleanliness consistent with the hardware design.

Friction may be included as a shear-load path for the following types of analysis for joints using preloaded fasteners:

Yield-strength analysis of the joint members (e.g., bearing).
Analysis for ensuring a structural assembly maintains required alignment of mechanical interfaces.
Fatigue or fracture mechanics analysis.

Relying on friction to carry shear across a joint is a departure from past NASA policy. This departure is intentional for the purpose of encouraging good design practice. In structural

assemblies subject to cyclic loads or vibration such as aerospace structures, joints using threaded fasteners in oversized holes are most dependable if shear is transferred by friction, by close-fit or interference-fit shear pins (dowel pins), or by other dedicated shear-transfer features rather than by the fasteners themselves. This is especially true for fasteners going into tapped holes or threaded inserts, with threads in the shear plane. When shear pins (dowel pins) are not practical, the fasteners should be highly preloaded; and care should be taken to clean the faying surfaces and prevent contamination from oils or other lubricants to maximize the coefficient of friction.

If the joint members slip within the limits of the fasteners’ clearance holes:

A nonlinearity is introduced within the assembly that can compromise the accuracy of dynamic loads analysis and predicted load distribution.
A damaging shock can be introduced when the fasteners contact the sides of the clearance holes.
The first fasteners within a pattern to contact the sides of the clearance holes are potentially overloaded.
The structural assembly may suffer detrimental misalignment.
If the slip is cyclic, the fasteners may lose preload despite the presence of locking features.
If the slip is cyclic and the fastener threads are in the shear plane, as is typically the case for fasteners threading into tapped holes or threaded inserts, the fastener may fail in fatigue.

If friction is inadequate for preventing slip at limit load and no shear pin (dowel pin) is present, the severity of the above effects should be reduced by ensuring minimal clearance fits, achieved by specifying precision fasteners having a closely controlled or “close tolerance” full diameter body in conjunction with precision holes per NAS 618. Fasteners installed into tapped holes or threaded inserts typically require relatively large-clearance fits between fasteners and their holes to account for location tolerances. As such, shear pins (dowel pins) should be used unless friction is adequate for preventing slip at limit load.