Bolt+Fatigue

Bolt Connection Fatigue //Thaison Nguyen, BAE, Penn State 2013 //
 * Multiple Structures - 2012 **

Overview
toc Bolt connections are commonly designed and sized according to material strength, block shear, bearing/tearout, and prying. The design method, mentioned above, is insufficient for connections exposed to cyclical loading. Cyclical loading has a tendency to cause crack initiation and growth in the structural material. This phenomenon is commonly known as material fatigue. Material properties, fabrication, member geometry, and cyclical loading frequency all affect fatigue induced damage.

Though cyclical loading is lower in buildings, when compared aeronautical structures, buildings and other related infrastructure do fail from fatigue. A few cases of fatigue include: the Kemper Memorial Arena in Kansas City, U.S.A.; William H. Putnam Memorial Bridge in Connecticut, U.S.A.; and Sayano-Shushenskaya Dam in Sayanogorsk, Russia.

Topics discussed in this study includes: fatigue mechanics, impact of fabrication and bolt connection geometry in fatigue resistance. In addition, strategies to reduce fatigue induced failures will be discussed.

Fatigue Mechanics
Endurance of the structural member is what engineers need to know when designing for fatigue. To better understand endurance, the material properties and cyclical loading patterns must be understood first.

Pedalarm_Bruch.jpg ||   ||
 * || [[image:Micro-View of Material Surface.png width="295" height="156"]] ||  ||   || [[image:Fatigue-Crack Prop.PNG width="307" height="147"]] ||   ||   || [[image:failures/Fatigue Failure.PNG width="297" height="139" align="center"]] ||   ||
 * || Figure 2.1, Microscopic View of Material Surface  Source: http://en.wikipedia.org/wiki/File:CrystalGrain.jpg  ||   ||   ||  Figure 2.2, Crack Growth  Source: PD-OLD-70, [] 9/ 9d/Ewing_and_Humfrey_fatigue_cracks.JPG  ||   ||   ||  Figure 2.3, Metal Fatigue Failure  Source: http://upload.wikimedia.org/wikipedia/commons/9/96/

The prelude to fatigue failure is micro-sized discontinuities and deformations, as shown on Figure 2.1. Cross-section changes occur at the discontinuities and deformations, resulting in local stress increases. Once the maximum stress is exceeded small cracks form and grow. Figure 2.2 demonstrates fatigue induced crack growth. After numerous loading cycles the effective structural section is reduced and the member fails, as exemplified in Figure 2.3.

Surface discontinuities and deformations arise from multiple reasons, the two primary are: atomic structure and loading pattern.

On the atomic level multiple groups of atoms are orientated in directions. Groups of atoms orientated in the same direction and bonded to each other are defined as crystals or grains (Allen and Thomas 1999, pp. 316). At the grain boundaries there is no uniformity, atoms are orientated to connect with adjacent grains (Allen and Thomas 1999, pp. 327-329). The resulting chaos manifests itself as boundary discontinuity and susceptibility to pitting corrosion. To achieve high fatigue endurance, it is desirable to reduce sudden angle and twist between the grains (Wang and Zhang 2000, pp. 289). One way to achieve low angle and twist between grains is the increase the grain boundary width and reducing the grain size (Hong et al. 1998, pp. 8). At high temperatures, fatigue endurance can be improved by aligning all the grains parallel to the tension direction. This creates one large grain, thus eliminating the weakness stemming from the boundary bridging the angles and twists between adjacent grains.

As mention earlier, load patterns also influences a member’s fatigue endurance. Members exposed to loading cycles with high tension-to-compression ratios, have low fatigue endurance (Schijve 2009, pp. 168). Unlike compression loads which closes cracks, tension loads does the opposite quickly opening cracks. The cyclical loading’s stress amplitude determines whether a structural member will fail in a plastic manner or not. Only low cycle loading causes structural members to fail in plastic behavior (Schijve 2009, pp. 161). By definition low cycle loading failure occurs when the stress amplitude surpasses the yield and experiences 10^5 cycles or less (Munse and Yao 1961, pp. 1). The number of load oscillations before low cycle loading failure is predicted in the //Coffin-Manson// //relation// (Munse and Yao 1961, pp. 15-16).

On the other hand, high cycle fatigue the maximum stress experienced doesn’t leave the elastic range. As less strength degradation occurs at over each cycle, when compared to low cycle loading, it takes occur over a million cycles to reach failure. High cycle fatigue failure can be predicted through using the Paris Crack Propagation Law (Anderson et al. 1961, pp. 9-14). Though fatigue evaluation and design standards aren’t discussed is this study, listed below are some of the standards pertaining to fatigue:



Fabrication
Characteristics of each member in a bolted connection are influenced by the fabrication process. Fatigue endurance is no exception. Repeated exposure to cyclical loading will exacerbate fabrication induced damages. Damage stemming from the various methods in fabricating bolt holes, bolts, as well as thermal effects will be discussed in greater detail below.

Bolt holes can be fabricated through punching, flame cut, or drilling. Each hole fabricating method imparts a certain level of damage and varies in economy/speed. Most holes in members are punched, the popularity of punched holes is due to the process’ speed. Punched holes have a slightly conical section (AISC 2002) and experience some damage at the hole’s edge. Both of these characteristics increases the local stresses at the hole. Flame cut holes is the other option, but suffer from heat effects and results in weakened material surrounding the hole. AISC recommends that flame cut holes should be grounded smooth. This process removes the heat damage and stress concentrations (AISC 2002). Once the flame cut bolt holes have been grounded smooth, its strength is considered to be equivalent to punched holes (Iwankiw and Schlafly 1982). Drilled holes typically experience less fabrication damage. The uniform bolt diameter reduces the possibility of local stress concentrations. Experiments done at the Technical University of Hamburg, in 2004, demonstrated that drilled holes experience greater fatigue endurance than punched holes (Huhn and Valtinat 2004, pp 304). A problem of drilling holes is the slow production speed, when compared to punched holes.

The method to fabricate bolts influences the location where fatigue cracks be concentrated and fatigue endurance in low cycle loading (Bakhmet’ev et al. 1982). Machined bolts whose threads are cut using a thread rolling machine typically experience failure at the threads. Discontinuous grain parallel, to the bolt shank, caused the threads on the machined bolts shear off (Colton 2009). Another bolt fabrication method is hot upsetting the bolt’s head and rolling the threads. The process of rolling applies pressure to deform the previously smooth surface into threads. Unlike threads in machined bolts, rolled threads in hot upsetting bolts has an angled grain, as seen in Figure 3.1. Hot upsetting bolts with rolled threads primarily failed due to a combination of thread failure and head rupture (Bakhmet’ev et al. 1982). Though hot upsetting bolts has more than one failure plane, these bolts experience greater loading cycles than machined bolts. According to AISC it is not recommended to begin accelerated cooling until the structural member’s temperature drops below 600°F (AISC 2002). The primary reason is that accelerated cooling introduces significant residual stresses within the member. Increases in residual stresses reduce the amount of additional stress which can be applied, before failure. Residual stresses arise due to differential cooling of the material, as in Figure 3.2. A member’s exterior will cool at a faster rate, due in part to the surface area exposed to the temperate atmosphere. This causes the exterior surfaces to contract and experience tension stresses.
 * || [[image:failures/Bolt Thread Grain.PNG width="640" height="135"]] ||  ||   ||   ||   ||   || [[image:Steel Differential Cooling.PNG width="176" height="150" align="center"]] ||   ||
 * || Figure 3.1, Grain of Machined Threads (Top) and Rolled Threads (Bottom) Based on grain structure described in Bolt Manufacture: Process Selection by Colton  ||   ||   ||   ||   ||   || Figure 3.2, Differential Cooling  Source: http://upload.wikimedia.org/wikipedia/commons/archive/    1/13/20120710194746%21Fotothek_df_n-08_0000774.jpg  ||   ||

Bolt Connection Geometry
Geometry of the bolt connection components is a contributor to fatigue endurance. An engineer must be conscience of how the member, being designed, is loaded. The loading direction will affect cyclical loading patterns and member stress concentrations. In addition, bolt connections exposed to multiple environments are likely to be affected by fatigue. Discussed below are typical bolt connection geometries that have low fatigue endurance.

Bolted angle are commonly used to replace welds experiencing brittle fatigue cracks (Cousins and Stallings 1998). Though bolted angle connections experience longer fatigue endurance, most fail before their predicted lifetime. The cause of this is bolt prying and excessive angle flexure; as in T-hangers, seated connections, and façade clip angles. Angles typically fail around the angle fillets (Cousins and Stalling 1998). Fatigue endurance can be increased by increasing the stiffness of the angle, either through thickening the angle or reducing the reduced the gage. At first glance, staggering bolts could increase fatigue endurance by reducing the number of bolt holes per row. Experimentally staggered holes experience greater stress concentrations but have negligible reduction in fatigue endurance, when compared to non-staggered holes (Grondin et al. 2004).

Another common bolted connection with low fatigue endurance are bolted cover plates. Fatigue cracks initiate at the first row of bolts and end bolts, caused by the short end distance (Albrecht 1985). Two ways to increase the fatigue endurance include: increase spacing between end bolt and edge and thicken the cover plate. Fraying the interface between the cover plate and member is another solution. Roughening the interface works by transferring some of the loads through friction (Garlock 2003). This reduces the demand on the end bolt and hole. Similar to bolted angle connections, pre-loading the bolt reduces the number of loading cycles. Experiments done in 1989 demonstrated that the amount of pre-loading is inversely proportional to the fatigue endurance (Chao and Sih 1989).

Prevention
Managing bolt connection fatigue is best achieved by proper design in conjunction with regular maintenance and inspection, as well as understanding the cyclical loading conditions present at the building site. This section will explore non-destructive methods in assessing fatigue. Also two relevant failures will be studied where either lack of maintenance or understanding of cyclical loading on building site.

Structural members can be non-destructively analyzed for possible fatigue damage though Liquid Penetrants Test (PT), Ultrasonic Testing (UT), and Radiographic Testing (RT). PT utilizes liquids penetrate and identify cracks extending to the surface. The highly visibility dye points out cracks so that observer can record and inspect the previously unseen cracks. Any debris or contaminants present in the cracks will likely prevent detection (ASCE and SEI 2004, pp. 25). Only non-porous materials, like steel, are compatible with PT (ASCE and SEI 2004, pp. 25). Ease of use in the field and low cost are the reasons for the wide use of PT.

Cracks that have not propagated to the surface can be detected either by UT or RT. UT uses non-audible sound waves detect cracks. As sound waves progresses through the structural member changes in density are detected, due to the change in wave speed. Attenuating material will skew UT results (ASCE and SEI 2004, pp. 26). Though UT is compact and readily adapted in the field (ASCE and SEI 2004, pp. 26), the need for a skilled operator to calibrate the detector is a disadvantage. The third most common non-destructive method is RT. High energy radiation, like x-rays, is shot through the structural member. Any cracks encountered are developed in the film, on the opposing side. Larger equipment is required to implement RT, this confines the non-destructive method to the fabrication shop (ASCE and SEI 2004, pp. 25). Best detection results are achieved when the radiation wave is parallel to the crack (ASCE and SEI 2004, pp. 26).

__Case Study I: Sayano-Shushenkaya Hydroelectric Power Station__

A case fatigue is the 2009 Sayano-Shushenskaya Hydroelectric Power Station Accident. Constructed in the early 1970s, the dam is designed to withstand earthquakes with a magnitude up to 8, on the richter scale (Rushydro 2009). Approximately 70% of the electricity generated is supplied to the four aluminum smelters belonging to RUSAL, the world’s largest aluminum company. During the night of August 16, 2009 through the morning of August 17, 2009 turbine number 2 experienced increasing levels of vibrations (Kutina 2009). Turbine number 2, since it’s installation in 1979, had experienced vibrations, due to unbalanced turbine wheel. At 8:13 AM on August 17, 2009; the turbine number 2’s cover and respective 920 ton rotor dislodged (Kutina 2009). Both the water pressure and the detached turbine number 2 took out a portion of the turbine hall. Below are two video clips of the accident: the video on the left show the events unfolding, while the video on the right is a simulation of the failure. It was later determined that most of the bolts securing turbine number 2 experienced fatigue damage, with eight of the bolts having over 90% cross-section lose (Kutina 2009).

media type="youtube" key="egeABBr5hyA" height="315" width="420"media type="youtube" key="Xr5dNjcub9g" height="315" width="420" __Case Study II: Putnam Memorial Bridge__

Another case of fatigue is the 1993 inspection of the William H. Putnam Memorial Bridge. Built in 1958 the Putnam Memorial Bridge crosses the Connecticut River. In terms of structure the bridge consists of 14 simply supported spans (Bernard and DeWolf 1994). Beams span between the bridge girders and are connected with bolted angle connections (Bernard and DeWolf 1994). Fatigue cracking in the angles and failure of some A325 7/8” bolts were noticed in the 1993 bridge inspection (Bernard and DeWolf 1994). Additional fatigue cracks occurred at the girder’s flange and beam’s web, at the bolted angle connections. Bolt prying and angle flexure was determined to be the cause of the fatigue cracks. An investigation concluded that the traverse beams shouldn’t be connected to the bridge girder’s top flange. The rational behind removing the restraint is to allow the beam to deform freely and prevent fatigue crack formation in either the beam’s web or girder flange (Bernard and DeWolf 1994). Another solution, though not favored by the investigation, is to use a stiffer angle to replace the failing angles. The second solution is predicted to have a life time of approximately 15 years (Bernard and DeWolf 1994).

Conclusion
Bolt connection fatigue is a usually not considered for buildings, because most buildings experience small cyclical loading. Most fatigue problems are caught when the structural member(s) has failed, causing millions in damage and disrupting operational continuity. What should be taken away from the two fatigue cases is that proactive maintenance and fatigue inspections are essential to catching fatigue damage early. Maintenance and fatigue inspection should not be passed off, where the cost of replacing damaged components is nearly negligible when compared to replacing a section of the building.