Tools › Structures › Joints › Preloaded Joint (NASA-STD-5020) › Method
Preloaded Joint (NASA-STD-5020): Method
Open the Preloaded Joint (NASA-STD-5020) tool
Preloaded Joint Analysis by NASA-STD-5020
The standard and its revision
The page follows NASA-STD-5020B, Requirements for Threaded Fastening Systems in Spaceflight Hardware, approved 2021-08-06 and revalidated without change on 2026-01-05 (it superseded NASA-STD-5020A with Change 1). Section and equation numbers below are those of revision B. The equations of Section 4 are printed in the standard as guidance ("simplistic equations") that show one way to meet each requirement; the requirements themselves are the numbered TFSR statements. The standard contains no complete worked example of the margins: its printed numbers are the torque-tension example of Appendix A.2 and the separation loads of Appendix A.12.3.
Loads and factors (Section 4.2)
Every check is written on the limit load, the maximum expected applied load. Applied loads never include preload or its change with temperature, and the factors of safety never multiply preload. The ultimate design load is limit × FF × FSu, the yield design load is limit × FF × FSy, and the design separation load is limit × FFsep × FSsep. The standard does not set FSu and FSy: the program levies them (it names NASA-STD-5001 as an example). The page opens with this site's usual 1.5 and 1.15; enter the program values. The fitting factor FF is at least 1.15 unless Section 4.2.2 allows 1.0. The separation factor of safety follows Figure 1: the program's ultimate factor if separation could credibly lead to a catastrophic hazard, the larger of 1.2 and the yield factor for a critical hazard, otherwise the larger of 1.0 and the program's test factor. The Load Level radio says whether the entered loads are limit or ultimate; ultimate loads are divided by FSu to recover the limit load.
Maximum and minimum preload (Section 4.3.1, Appendix A.2)
Pp-max = Ppi-max + PΔt-max (Eq. 1) and Pp-min = Ppi-min − Ppr − Ppc − PΔt-min (Eq. 2), with the initial preloads
Ppi-max = cmax (1 + Γ) Ppi-nom (Eq. 3), Ppi-min = cmin (1 − Γ) Ppi-nom (Eq. 4), Ppi-min = cmin (1 − Γ / √nf) Ppi-nom (Eq. 5).
Eq. 4 serves separation of a separation-critical joint; Eq. 5, with the number of fasteners nf, serves joint slip and separation of a joint that is not separation-critical. With torque control Ppi-nom = T / (Knom D) (Eq. 24), and cmax, cmin are the ratios of the maximum and minimum effective torque to the nominal one. The effective torques are the specified maximum less the minimum breakaway torque and the specified minimum less the maximum locking torque (Eqs. 29, 30), or the specified values themselves when the torque is called out above running torque (Eqs. 27, 28: enter zero locking torques). The standard does not define the nominal effective torque when a running torque is subtracted; the page takes the mean of the two effective torques, which leaves Eqs. 25 and 26 unchanged. Ppr is the short-term relaxation, 5 percent of the minimum initial preload for all-metallic clamped parts (Table 1), and Ppc a creep loss.
The preload variation Γ comes from Table 3. For a joint that is not separation-critical it may be 0.25 (lubricated) or 0.35 (as received) with torque control, or the variation seen in tests if that is larger: Γa-max = Pmax / mean − 1 for the maximum preload and Γa-min = 1 − Pmin / mean for the minimum (Eqs. 33, 34). A separation-critical joint needs the statistical value Γ90/95 = s σ / mean (Eq. 36), where σ is the sample standard deviation and s the two-sided 90/95 tolerance factor of Table 5. The page computes all three from pasted test preloads. For a sample size between the rows of Table 5 it uses the factor of the next smaller tabulated size, which is never too small.
Stiffness factor, load-introduction factor and thermal change
Before separation the bolt load is Ptb = Pp + n φ Pt (Eq. 8) with φ = kb / (kb + kc) (Eq. 9). Both factors can be entered. Otherwise the clamped parts are modeled as two hollow cones that open from the bearing faces at a half-apex angle (30° by default; Appendix A.12.3 uses 25°) and meet where their diameters are equal; each piece between layer faces is a frustum of stiffness π E dh tan θ / ln[(2 t tan θ + Ds − dh)(Ds + dh) / ((2 t tan θ + Ds + dh)(Ds − dh))], and the pieces act in series (NASA/TM-20250005284, eqs. F.27 and D.6). The bolt is a rod over the grip. The stiffness-based load-introduction factor is the share of the compression-zone flexibility that lies between the two loading planes (Eq. 56 of the standard); the geometric factor Llp / L (Eq. 37) is printed beside it. The thermal preload change is Pth = [kb kc / (kb + kc)] L ΔT (αc − αb) (the same report, eq. E.6), evaluated at the hot and the cold temperature change; the largest increase enters Eq. 1 and the largest decrease Eq. 2.
Tension: does the joint separate before the fastener ruptures?
Section 4.4.5 (TFSR 12) requires preload in the bolt load only when rupture would occur before separation. The page uses the linear test of Section 4.4.1 at maximum preload: the joint separates at P′sep = Pp-max / (1 − n φ) (Eq. 11), and the bolt load would reach the allowable at P′tu = (Ptu-allow − Pp-max) / (n φ) (Eq. 10). If P′sep is smaller, the bolt carries only the applied load at rupture and MSu = Ptu-allow / (FF FSu PtL) − 1 (Eq. 6); otherwise MSu = P′tu / (FF FSu PtL) − 1 (Eq. 7). Yield is treated the same way with P′ty (Eqs. 15 to 17) and Pty-allow = (Fty / Ftu) Ptu-allow (Eq. 18); the standard states "when separation occurs before yield" without a test, and the page compares P′sep with P′ty by analogy. Ptu-allow is the allowable of the fastening system: the lower of the fastener and the nut or insert when the latter is entered.
Separation, shear, combined loading and slip
Separation: MSsep = Pp-min / (FFsep FSsep PtL) − 1 (Eq. 19). The separation load is deliberately set equal to the minimum preload, not to the larger linear value (Appendix A.11).
Shear: Psu-allow = π D² Fsu / 4 with the full diameter body in the shear plane (Eq. 12) or Fsu Am with the threads in it (Eq. 13), and MSu = Psu-allow / (FF FSu PsL) − 1 (Eq. 14). Friction carries no shear at ultimate load.
Combined loading at ultimate, preload omitted: with the body in the shear plane Rs2.5 + (Rt + fbu / Ftu)1.5 ≤ 1 (Eq. 20) or, with plastic bending, Rs2.5 + Rt1.5 + fbu / Fbu ≤ 1 (Eq. 21); with the threads in the shear plane the exponents are 1.2 and 2 (Eqs. 22, 23). The exponents come from tests of A-286 3/8-24 bolts (NASA/TM-2012-217454). The standard gives the criterion, not a margin; the page reports λ − 1, where λ is the factor by which shear, tension and bending can grow together until the criterion is met. The bending stress is 32 M / (π d³) on the body diameter, or on the minor diameter when the threads are in the shear plane (the page's choice of section).
Joint slip (Appendix A.10): for a concentric joint MSslip = μ nf Pp-min / [FS (PsL-joint + μ PtL-joint)] − 1 (Eq. 84), and per fastener μ Pp-min / [FF FS (PsL + μ PtL)] − 1 (Eq. 86). The coefficient of friction may not exceed 0.20 for clean, uncoated, non-lubricated metal and 0.10 for other surfaces unless shown by test (TFSR 14).
Earlier criteria and thread areas
NSTS 08307 Revision A (1998), the Space Shuttle criteria for preloaded bolts that NASA-STD-5020 names as the historical practice, has the same structure and differs in detail: preload variation 0.35 unlubricated, 0.30 cadmium-plated and 0.25 lubricated; a preload loss of 5 percent of the maximum preload; the bolt load PLDmax + n φ (SF × P) checked together with PAt / (SF × P) in every case, without a separation-before-rupture test; the separation margin PLDmin / [(1 − n φ) Psep] − 1; and the interaction (Ra + Rb / K)² + Rs³ ≤ 1. The page computes only the NASA-STD-5020B forms. Neither NSTS 08307A nor NASA RP-1228 gives a formula for the thermal preload change or a worked margin example.
The Unified tensile stress area 0.7854 (D − 0.9743 p)² is the FED-STD-H28 formula printed in NSTS 08307A Appendix A, which allows Ftu At as the tensile allowable when the fastener specification gives no load. The Unified minor-diameter area (π/4)(D − 1.299038 p)² reproduces the root areas of NASA RP-1228 Table V. MIL-HDBK-5E Table 8.1.5(b), reprinted in RP-1228 Appendix B, bases the tensile strength of steel fasteners on the minor area instead: 5,340 lb for a 1/4-28 fastener of 160 ksi with the MIL-S-7742 thread and 5,480 lb with the MIL-S-8879 thread, against 5,820 lb from the stress area. Enter the specified strength of the fastener where there is one. RP-1228 gives T = K F d with K = 0.2 as the commonly assumed nut factor and 0.15 as more typical of steel on steel.
Margin boxes
Strength boxes read ULT (limit × FF × FSu) and the yield box LIM (limit × FF × FSy), as everywhere on this site. Separation and slip are not strength checks; their boxes read SEP (design separation load against minimum preload) and SLIP (limit load times the entered slip factor against friction at minimum preload), the levels at which the standard defines them.
What it assumes
One fastener location with known loads; linear springs for the bolt and the clamped parts; an applied tension concentric with the bolt (no prying, no bending of the flanges); the same preload statistics for every fastener in the joint; shear carried by the fastener at ultimate and by friction only in the slip check; elastic moduli that do not change with temperature; thread areas of the basic profile unless entered.
When it misleads
When the load path bends the flange or pries on the bolt, the bolt load is larger than n φ Pt and the applied load must first be corrected. When rupture precedes separation, the combined-loading exponents were not measured for that case. A computed φ from frusta depends on the assumed cone angle and bearing diameters; the standard recommends test or finite element analysis where it matters. The default yield strength, moduli and layer expansion coefficients are illustrative values typed from memory, and the other defaults are typical values, not design allowables: use the fastener specification and test data. The frustum model of the page puts both bearing faces at the ends of the grip, so a bolt into a tapped hole or insert needs an engineering choice of effective grip. Bearing and shear-out of the clamped parts, thread stripping, fatigue, the Figure 8 ductility screen and locking features are not checked. Figure 8 of revision B labels its outcome boxes "Perform analysis per Eq. 10" and "per Eq. 11"; the text of Section 4.4.1 and Appendix A.6 assigns Eq. 6 to separation before rupture and Eq. 7 to rupture before separation, which is what the page does.