Research On Fatigue Reliability Prediction Methods For Main Drive Bearing Of EPB

Shield machine main drive bearing is one of the core components of the cutter drive system. It has a direct impact on the efficiency, stability and fatigue life of the shield machine. Due to the irreversibility of the shield construction, the tunneling process does not exist a wide range of maintenance space and time, so we must ensure the reliability of the main bearings. Fatigue reliability design of the main bearing has been the most important issues of the design stage. Existing approaches for bearing fatigue reliability design entirely depending on the bearing life formula, can not consider the impact factors such as randomness of the load and the dispersion of cumulative damage. If a prediction method used in the design phase of the main bearing fatigue reliability, considering the structure and working conditions for the main bearing,can be proposed, the accuracy of the reliability design can be greatly improved. This paper makes a research on the main drive bearing of EPB, taking full account of the time-varying of strata and location of the tools, combined with cutterhead load calculation model under the compound strata, forming a boring load dynamic simulation method of cutterhead. Using finite element analysis and intelligent algorithms to transform the load spectrum into a stress spectrum. On this basis, analyzing the failure modes of components, then a reliability of fatigue life prediction method, considering the the system failure correlation, based on the theory of fracture mechanics and strength degradation theory was proposed.Firstly, strata median model of the whole construction section is established and the dynamic simulation of the cutter boring load is conducted. Statistical treatment of the load under different conditions is completed, then, the variety of conditions load were synthesized as the input of the main bearing system.Secondly, the bearing capacity characteristic under single load and coupling load is analyzed, and the impact of different loads on the main bearing contact stress based on the probability of the finite element design method is discussed. Taking into account the defects of bearing stress spectrum from the theoretical method and the finite element simulation, a BP neural network is established to achieve bearing stress spectrum prediction, making finite element analysis results as samples.Thirdly, based on the analysis of failure modes of each component, a stochastic process is introduced to established fatigue reliability model of components, using fracture mechanics and strength degradation. According to the mutual relations of the various parts of the main bearing, a fatigue reliability model is proposed with failure correlation of series system.Finally, this paper conducts an engineering example analysis of Shenyang Subway Line. The results verify that the fatigue reliability prediction method under random loads is with high accuracy. It can provide a design reference at the early design stage, and also can forecast and track the dynamic reliability of the main bearing system.