Study of the Usage Life for a Robotic Arm Based on Reducer Diagnosis

A robotic arm is an important equipment in an automated production line. The reducer in the robot arm is one of its important components but with the highest failure rate. The reducer is a complex system including input shaft, output shaft, gears and bearings, etc. When the reducer starts to be damaged, performance of the robotic arm will be affected, and even worse the system shut down and the production efficiency might be induced, to name only a few. Therefore, how to extend the useful life of the reducer has become an important issue. In general, the clamping jaws (grippers) are installed on the 6th axis in charge of the loading and unloading, which will inevitably higher the reducer failure rate than that of the other 5 axes. Therefore, this research aims at the useful life optimization of the 6th axis reducer. The machine learning algorithms were adopted to establish methodologies to find the key factors. In addition, since the movement path of the robot arm determines the life of the reducer, multiple paths with the same starting and ending position will be generated through the forward and reverse processing, and then the RMSF (Root Mean Squares of Features) values of various paths are calculated. The optimal path with the optimum useful life of the reducer will be the one with the minimum RMSF value. This study has successfully shown that significant differences exist among the various movement paths based on the healthy and abnormal data from the cooperated reducer manufacturing company. This means the developed methodology could be used as an helpful index to extend the useful life of the reducer and also to serve as the basis in futuristic predictive maintenance system.