Reduction of chatter using a probabilistic approach

Industrial robots are widely used in numerous industrial plants usually to perform repetitive, difficult or hazardous tasks such as polishing, grinding, milling, deburring, and welding. However, their relatively low stiffness prohibits the robots from machining of metallic materials accurately. Besides, the resulting vibration causes reduced tool life. One of the main problems with robot machining is chatter. Although many researchers have studied this problem for a while, they were not able to fully eliminate chatter due to the complexity of the problem which arises from many factors including cutting parameters, cutting tool and work-piece material. Several studies have been conducted to predict the occurrence of chatter. However, many of them lack the important element in predicting chatter: the element of uncertainty. It is important to consider uncertainty in predicting chatter since the parameters associated are inherently uncertain. This study implements probabilistic approach to predict chatter considering the uncertainty in machining parameters. The magnitude of cutting force is used to determine whether a chatter has occurred or not. The research question of this study is 'Can a probabilistic analysis be used to predict chatter occurrence?' To answer the research question, the following specific aims were developed: (1) develop a framework for predicting chatter occurrence, and (2) perform probabilistic analysis of chatter occurrence. The framework consists of a multibody modeling software (ADAMS), probabilistic analysis software (NESSUS) and MATLAB which is used as an interfacing platform. The probabilistic analysis implemented in the framework provides two important results. Firstly, it determines the probability of occurrence of chatter. Secondly, it provides sensitivity analysis that shows most important parameters which instigate chatter.