Abstract
Studies have shown that the connection details used for cable anchorage blocks on cable-stayed bridges have the potential for fatigue damage due to fluctuating stresses generated by the cyclic traffic loads passing over the bridge. To investigate the fatigue damage and determine the remaining fatigue life of a cable anchorage block used on a cable-stayed bridge, finite element (FE) analyses were undertaken by using the Fatigue Load Model 4 (FLM 4) proposed by the Eurocodes to identify the most fatigue-critical locations within the details. One of the main objectives of this research was to identify the critical area prone to fatigue in the anchorage block due to the response in traffic loads. Therefore, two types of numerical models of a typical cable anchorage block were analysed as a three dimensional sub-model which was driven by global cable forces obtained from the global analysis of a three-span cable-stayed bridge. These models are of the cable anchorage block without the longitudinal girder modelled and the cable anchorage block with the longitudinal girder modelled. The cable anchorage blocks without the longitudinal girder model were classified into three categories of model types namely; model types 0, A and B. Similarly, the cable anchorage blocks with the longitudinal girder model were classified as model types A-G, B-G and C-G. These model classifications are based on several boundary conditions simulated in the analyses. In addition to this, the fatigue behaviour of the cable anchorage block was analysed by using three different approaches namely; by using the nodal stresses at the location of the stress concentration (node stress concentration), by using a stress averaged over an area in the vicinity of the stress concentration (average elements) and by using the hot-spot method, in order to identify the stress ranges that adversely affect the remaining fatigue life of cable anchorages. Each approach was analysed with three different mesh sizes; 5mm by 5mm, 10mm by 10mm and 20mm by 20mm in order to carry out a mesh sensitivity analysis of the resulting stresses and associated stress ranges. The 10mm by 10mm mesh size was found to be most appropriate for this fatigue appraisal. This finding is supported because the 10mm by 10mm mesh size is specified in several code of practices such as the International Institute of Welding (IIW) and BS 7608 as guidance for use when determining hot-spot stress when using the hot-spot method for the fatigue analyses of a welded detail. The critical stresses from model type C-G were used in the fatigue appraisal as the behaviour of this model represented more accurately the actual cable anchorage block on the cable-stayed bridge compared to the other types of models used. Model type C-G were selected for further fatigue appraisal as this model include the correct boundary conditions and applied load that represented the actual condition of the anchorage behaviour on the cable-stayed bridge. This included the movement of the top anchorage block due to the displacement of the cable and in addition the deck movement. Also, non-uniform pressure was applied on the bearing plate which was included to model possible construction tolerances which was one of the important properties of the model type C-G. In evaluating the possible fatigue damage in the cable anchorage block, the cumulative model for fatigue failure expressed in terms of Miner’s rule was used. In addition to this, the condition of the structural detail due to fatigue with increasing traffic loading was determined by projecting traffic volume increases of up to 20%. Based on the results calculated, if the long distance traffic characteristic was used in fatigue appraisal, the cable anchorage block was justified to be not ‘safe’ as the damage accumulation for fatigue, Dd at the top gusset was recorded as 1.270, which exceeded the limiting value of 1.0 corresponding to a 120 year design life. However, if medium distance traffic characteristic was used in the fatigue appraisal, the cable anchorage block will remain ‘safe’ except when a 20% increase in traffic volume was included in the analysis, which resulted in Dd value of 1.016. Also, if a more conservative value of Dd = 0.5 as suggested by IIW (2008) was used, the cable anchorage block appraised by using both the long distance and medium distance traffics was found not safe from fatigue damage and would not survive its design working life without structural repair. For future fatigue appraisals of anchorage blocks (and other important structural details), it is strongly recommended that the numerical model of anchorage block is analysed together with the longitudinal girder using the hot-spot method. A 10mm by 10mm finite element mesh size is suggested and it is also necessary to specify the displacement at the top of the anchorage block to simulate the cable movement together with the girder movement both of which are obtained from the global analysis of the whole bridge structure.