Loading...

Process Capability Analysis of Additive Manufacturing Process through Predictive Model of Dimensional and Geometric Errors based on Machine Learning

Abdolahi, Alireza | 2022

1160 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55178 (08)
  4. University: Sharif University of Technolog
  5. Department: Mechanical Engineering
  6. Advisor(s): Khodaygan, Saeed
  7. Abstract:
  8. Additive manufacturing (AM)has gained extensive industrial and research attention in recent years. Reducing manufacturing waste, lead-time and costs and the ability to build surfaces and parts with complex shapes, assemblies all at once or parts with internal features are some benefits of AM. However, complex error generation mechanisms underlying AM digital physical chains are likely to result in geometrical inaccuracies of the final product, thus posing significant challenges to design and tolerancing for AM. Therefore, predictive modeling of shape deviations is critical for AM. With increasing volumes and varieties of data, machine learning has gained extraordinary popularity due to its ability to explore complex patterns in observed data and make data-driven predictions or decisions on new data. process capability is a measurable property of a process to the specification, expressed as a process capability index that predicts how many parts will be produced out of specification. The output of a process is expected to meet customer requirements, specifications, or engineering tolerances.Engineers can conduct a process capability study to determine the extent to which the process can meet these xpectations. In this research, a machine learning method is proposed which enables the prediction of deviations for different shapes and process settings, then process capability of AM process is analyzed using relevant indices. The Process parameters are specified and a number of simulations are run for parts of different shapes and sizes to collect input deviation data. Through a case study, it’s demonstrated that the trained network manages to predict the three geometric deviations of 95.44% of shapes manufactured with varied size and process parameter settings with an accuracy of more than 87.00%, 88.34% and 85.85%.

  9. Keywords:
  10. Additive Manufacturing ; Machine Learning ; Process Capability ; Defect Prediction ; Probability Theory

 Digital Object List

 Bookmark

...see more