digilib@itb.ac.id +62 812 2508 8800

ABSTRAK Adi Ekaputra
PUBLIC Open In Flip Book Alice Diniarti

Remaining life assessment aimed at predicting the remaining life of a mechanical component after a certain time of operation is developed in order to determine whether the component may be operated for a certain time of operation in the future. A particular problem that exists in the planetary gear in dump truck final drive after a certain time of operation is surface pitting, which is a formation and propagation of small holes in the contact surface of gear teeth that can cause catastrophic failure. The goal of this research is to develop and apply a methodology to predict the remaining life of the gear tooth by calculating the pits’ life cycle to failure based on multi-pit model and variable engine torsional load. The pitting growth is assumed as propagating crack using fracture mechanics principles. The particular gear tooth with working contact load is modeled, analyzed using fracture mechanics in order to find the stress intensity factor, K and also linear static finite element method in order to find the nominal stress value, S. The values of K and S are then used to calculate the geometry factor ?, which enables the pit remaining life be calculated with Paris Law of crack propagation. The process is repeated for each pit crack geometry, ultimately creating a geometry factor database of a number of crack geometry variation. This geometry factor database is then used to calculate the remaining life, that is defined as the life until the given pits develop to the specified maximum allowable pit size. With an image processing software, pit geometry is obtained numerically and processed using geometry factor database, resulting in a each pit’s remaining life with varied gear loads. This set of each pit’s life values is then processed through statistical means to find the probabilistic remaining life with a certain confidence level value. With the aforementioned methodology, sample case study pit data is analyzed which results in 122 hours remaining life with 90% confidence level. This methodology presents a practical approach for remaining life estimation, and can be adopted to cases of other different mechanical components.