Fatigue Life Analysis of Crane K-Type Welded Joints Based on Non-Linear Cumulative Damage Theory

摘要

There are large numbers of welds on cranes' lattice booms. Due to notably complicated force conditions, the fatigue life analysis of the lattice boom is difficult when designing the limit life. The fatigue cycle of the utilization level of a lattice boom crane is in the overlap zone of low-cycle fatigue and high-cycle fatigue. The poor accuracy of measurement of the fatigue life of lattice booms indicates that the calculation dispersion of the linear Palmgren-Miner (PM) rule was high. To obtain real crane stress spectra inexpensively and conveniently, a new stress spectra acquisition method based on the 'measured & simulated & compared & statistics' integrated strategy of crane K-type welded joints is proposed. The errors of the maximum stress amplitudes between the measured stress spectra and the simulated stress spectra were less than 10%. The fatigue test results also indicated that errors of the test fatigue life were less than 10% under both the simulated and the measured stress spectra. A new, simplified Huffman non-linear cumulative damage theory is proposed to calculate the fatigue life of crane K-type welded joints based on the notch stress and strain approaches. The calculation results indicated that the accuracy of the non-linear damage accumulation was higher than that of the PM rule, although the calculation result based on the non-linear damage accumulation method was slightly un-conservative when the initial damage was not considered in the calculation. By setting different initial damage conditions, various results were analysed, which revealed that the calculation errors of fatigue life based on the non-linear theory were less than 10% when the initial damage levels were set from 0.02 to 0.04. Such results are appropriate for engineering applications. When the fatigue life calculation needs to be more conservative, the initial damage levels may be set from 0.04 to 0.07; the resulting calculation errors could be less than 25%. As the Huffman non-linear cumulative damage theory requires data from only a few material properties, such as the cyclic stress-strain curve and the constant amplitude strain-life curve, it could therefore be more suitable for engineering applications with higher calculation accuracy and fewer costs.