Design of intelligent reliability prediction models for industrial automation machinery: A robotic injection molding automation case study

In this thesis study, a Robotic Injection Molding System (RIMS) is taken in the consideration dealing with existing and forecasted equipment and process reliability. Failure Mode and Effect Analysis (FMEA) is used for finding out RIMS most critical cases that effect overall system reliability according to historical 48 months real shop floor performance data. Obtained FMEA results are processed for training and structuring of best performance Artificial Neural Network (ANN) model to predict future performances in terms of equipment Mean Time Between Failure (MTBF). Additionally, 300 cycle (24 Hours production period) process data is used to structure another best performance ANN model to predict future Cycle Time (CT). Consolidated equipment and process reliability prediction results show that the pneumatic gripper of the manipulator is most critical equipment for both system and process reliabilities. In the end of the study preventive and improvement actions that are acquired from FMEA tables are suggested in accordance with forecasted MTBF and CT to eliminate overall potential risk factors to maintain best equipment and process life cycle.