External Tension Loads on Bolted Joints

American Fastener Journal, Nov/Dec 2006 by Barrett, Richard T

Q Can I apply an external tension load on a bolt that exceeds the installation tension load?

A It is not a good idea, as bad things such as joint separation and/or bolt failure can occur. I can give you some guidelines to help you avoid such problems.

Background:

Although some manufacturers live dangerously and load bolts above yield, it is not a good thing to do. In the first place, there are too many variables in an ordinary (non-instrumented) bolt installation to allow the yield point to be accurately determined. Thread friction, nut/head friction, joint relaxation, shear loads, and thread locking mechanisms all have to be considered in net bolt tension load determination. Joint and bolt stiffnesses must also be factored in, as well as differential expansion/contraction of bolts and joints.

Bolt yield and/or joint separation:

If a bolt is subjected to an externally applied tension load, the preload must exceed the applied load to prevent joint separation. You do not want joint separation under any circumstances. Another rule is to keep the maximum bolt load below yield. Otherwise, you will yield and lose some of the clamping load. For fatigue loading, it is desirable to have the bolt load near yield to limit stress cycling, so the ideal fatigue installation is to be near, but still below yield.

Joint relaxation and embedment:

In many cases, the joint material surface is softer than the bolt head or the nut. Consequently, the head or nut will bite into the joint surface as it rotates. This is why I recommend a hardened washer under both the head and the nut. In addition, the joint surface may have high spots on it. These may yield both during and after tightening to allow a decrease in bolt load. The bolt can also wind up in torsion (angle = TL/JG) during installation, such that it will unload a small amount after it is released. Unless the bolt is re-torqued several hours later, the final load will be less than that indicated by the initial torque reading. If the joint and bolt materials are different, differential thermal expansion can add or subtract from the bolt load.

Joint friction coefficients and running torque:

If we go back to the definition of torque, it is simply a force times a distance (e.g., in. lbs.). The bolt axial load is the load left over after all of the friction loads have been overcome. The major torque losses are from thread friction and head/nut friction. The lubrication used in these areas determines the friction coefficients. The friction coefficients (times normal forces) determine the "K" nut factor in the formula, T = KDF, where T = torque, D = bolt diameter, and F = bolt axial load. Although "K" values can be calculated, they are more easily determined with instrumented equipment.

Running torque (a.k.a. prevailing torque) must also be added to the torque value calculated from the above formula. Running torque is the torque needed to seat a fastener with some type of locking device, such as deformed thread, nylon plug, or collar, etc.

Spring constants of bolt and joint:

In structural analysis, we usually look at the stiffness of each load path for its load carrying capability. For a bar, the elongation/compression is PL/AE, where P is the load, L is the length, A is the cross-sectional area, and E is the modulus of elasticity. This is also true for the joint, although "A" usually has to be approximated. The relative stiffnesses of the joint and fasteners need to be determined to get the relative distribution of an externally applied load.

Using John Bickford's (Introduction to the Design and Behavior of Bolted loints by Marcel Dekker) analogy of representing both the joint and bolt as a strong and weak spring assembly (Figure 1), and assuming that no joint separation can occur, we see that the bolt deflection and joint deflection are equal as an external load is applied. The joint is unloaded as the fastener is loaded further. However, the amount of load that each "spring" carries is per the ratio of their stiffnesses. A good value for this ratio is about 5 to 1, so the bolt takes about 16.7% of the added load. The rest of the applied load goes to unloading the joint. This means that the initial bolt preload should be far enough below yield that the added load still won't bring the total load above yield. (A study of the joint loading diagrams in the two books referenced here will make this concept easier to understand.)

Addition off shear loads:

If a bolt is loaded in both shear and tension, the stresses must be combined and compared to the overall strength of the bolt material. You can't simultaneously use up both the shear strength and tensile strength of the bolt. You can combine the loads and use stress/load ratios as a substitute for a Mohr's circle. Then you can calculate a Margin of Safety: M.S. = (I/R^sub S^ R^sub T^) - 1, where R^sub S^ = actual shear/allowable shear and R^sub T^ = actual tension/allowable tension. Note that R^sub S^ R^sub T^ should be less than 1.0 for a positive margin. (See NASA RP-1228 Fastener Design Manual by R.T. Barrett.)