Fatigue Analysis of Main Shaft with Split Collar - Utmattingsanalyser av deler til boring

Abstract

The task concerns a connection between a Top Drive shaft and a two part split collar resting on it. The main purpose of the connection is to transfer drilling-related loads from surrounding parts to the main shaft. Due to old design, heavy loads and a high number of cycles, the life of the connection is short where small cracks are initiated at the shaft shoulder grooves. The desire is to delay or in best case prevent crack initiation. The goal of the study is to find an optimal solution for the connection based on FE-testing and relevant theory. A more detailed geometry can help reduce the high stress concentrations and then increase the fatigue life. The FE-Analysis software Abaqus 6.13 is selected to evaluate the existing design and to carry out the proposed new design. The theory of the report consists of relevant information about the topic of fatigue and methods to assess the fatigue life of the connection. The stress-based and strain-based approaches to fatigue life assessment are thoroughly discussed before later used to evaluate the two designs. The FE-Analysis of the original design shows high stress concentrations at the upper and middle shoulder grooves where several locations are close to or exceeds the yield limit. The upper shoulder groove stands out with a large high-stress area. The analysis indicates possible improvements where increased groove radiuses and better load distribution between the three shoulders, may reduce the high stress values. Tests showed that plastic deformation is eliminated by increasing the groove radius above 8mm and that additional stress reduction is possible by increasing the contribution from the lower shoulder. The new design showed positive results where the critical stress values from the original design were strongly reduced. The worst location at the upper shoulder showed a von Mises stress reduction of 31%. The fatigue life of the two designs is evaluated according to the stress-based and strain-based approach. The original design suffers from too high stress values, making the strain-life curve better suited. The proposed new design is evaluated according to S-N curves due to elastic behavior, making the stress-based approach well suited. The result indicates large improvements to the fatigue life where the proposed new design shows improved life compared to the original design. The fatigue analysis suffers from lack of relevant information about the material and its fatigue properties. This makes it difficult to obtain an accurate estimate of the life. The mean stress, notches, fatigue parameters and reduction factors, are all important factors which alter the fatigue life. The result shows a large scatter between different assumptions where only additional information about the material and fatigue testing, may give a more accurate estimate of the life.