T. Jaglinski Materials Science Program, University of Wisconsin-Madison, 147 Engineering Research Building, 1500 Engineering Drive, Madison, WI 53706-1687 e-mail: tmjaglinski@engr.wisc.edu A. Nimityongskul Engineering Mechanics Program, University of Wisconsin-Madison, 147 Engineering Research Building, 1500 Engineering Drive, Madison, WI 53706-1687 e-mail: apnimityongs@engr.wisc.edu R. Schmitz Engineering Mechanics Program, University of Wisconsin-Madison, 147 Engineering Research Building, 1500 Engineering Drive, Madison, WI 53706-1687 e-mail: robert.schmitz@ata-e.com R. S. Lakes Department of Engineering Physics, Biomedical Engineering Department, Rheology Research Center, University of Wisconsin-Madison, 147 Engineering Research Building, 1500 Engineering Drive, Madison, WI 53706-1687 e-mail: lakes@engr.wisc.edu Study of Bolt Load Loss in Bolted Aluminum Joints Bolted joints are used widely in mechanical design and represent a weak link in a system where loss of joint clamping force can lead to degraded product performance or human injury. To meet current market demands, designers require reliable material data and analysis tools for their industry specific materials. The viscoelastic response of bolted aluminum joints used in the small die-cast engine industry at elevated temperatures was studied. Bolt load-loss tests were performed using strain gages in situ. It was found that after a week at temperature, most bolts lost 100% of their initial prestress. Nonlinear constitutive equations utilizing parameters obtained from uniaxial creep and relaxation tests were used in a simple one-dimensional model to predict the bolt load loss. The model cannot predict the detailed response and overpredicts retained bolt stress for bolt holes that are not preconditioned. For preconditioned holes, the behavior is intermediate between creep and relaxation. DOI: 10.1115/1.2400262 1 Introduction Many technological applications rely on bolted connections to connect components composed of dissimilar materials or compli- cated geometries as well as providing a means of access and maintenance. Although necessary, these connections provide a weak link where loss of joint clamp force can lead to degraded performance or serious injury in extreme temperature applica- tions. Bolt load loss arises from the time- and temperature- dependent nature of all materials, and in the case of bolted joints, time dependency is manifested by creep or stress relaxation of the structural components. Die-cast aluminum engine blocks are not immune to time- dependent problems. Small die-cast engines that rely on air cool- ing can experience transient operating temperatures in excess of half the homologous melting temperature T m , where time- dependent behavior is greatly accelerated. High-temperature areas in an engine block include exhaust ports, combustion chambers, and head joints. Loose joints in any of these areas can lead to inefficient operation, escalated emission of pollutants or prema- ture failure. Today’s domestic markets, coupled with increasing global product competition, demand low-cost products with long life cycles, where little or no consumer maintenance is required to preserve performance over the product’s lifetime. To meet these design goals, detailed knowledge of industry specific material and structural responses is required. Viscoelastic material behavior, in which we include all time- dependent behavior i.e., anelasticity or any other form of time- dependent inelastic behavior can be obtained from uniaxial creep or stress relaxation tests. It would be advantageous to utilize the acquired material parameters from these tests to predict the re- sponse of geometrically complex structures subjected to pro- longed load and varying time-temperature histories. However, proper use of time-dependent constitutive equations entails an un- derstanding of the detailed stress or strain history to which the structural element is subjected, which can be extremely complex, and the associated boundary conditions, which are heterogeneous in real structures. The actual relaxation behavior of a bolted joint differs from the observed response from a strictly defined creep or relaxation situation. In linear viscoelasticity, creep and relaxation are defined as a material’s time-dependent response due to a step input stress or strain, respectively. Furthermore, it is assumed that the boundary conditions remain the same throughout the time pe- riod of the step input condition. For a real bolted joint, neither the applied stresses nor strains are step functions due to fluctuating time-temperature histories as well as a partitioning of load be- tween the bolt and flange. For small engines the period of opera- tion is generally short and dominated by a pronounced thermal transient. Such a situation presents challenges for analysis since metals at elevated temperatures are nonlinearly viscoelastic, meaning the time-dependent behavior depends on the applied stress or strain. The majority of literature concerning experimental time- dependent bolted joint behavior is concerned with steam applica- tions 1 and bridges 2–4. Most of these studies concern steels of varying compositions 5–8 and some refractory metals 9,10. Besides aluminum alloys 11, the automotive industry has studied Contributed by the Materials Division of ASME for publication in the JOURNAL OF ENGINEERING MATERIALS AND TECHNOLOGY. Manuscript received April 28, 2005; final manuscript received May 2, 2006. Review conducted by Matthew P. Millar. 48 / Vol. 129, JANUARY 2007 Copyright © 2007 by ASME Transactions of the ASME