Loading...

Tool Life Prediction Using Bayesian Updating

Rahmanpanah, Hadi | 2018

510 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 52413 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Akbari, Javad
  7. Abstract:
  8. Traditional tool life models and wear prediction do not take into account the variation inherent in metal cutting processes. These lead to additional manufacturing costs. This research presents a stochastic method for both tool life and tool wear prediction. In the first part, Bayesian inference is applied to estimate the two model constants of Taylor tool life prediction model using a discrete grid method. Tool wear tests are performed using an uncoated M42 HSS tool and AISI 1018 carbon steel work material. Test results are used to update initial beliefs about the constants and the updated beliefs are then used to predict tool life using a probability density function (pdf). Finally, the research compares the tool life predictions using Bayesian inference to the least squares method, and the effect of prior selection and likelihood uncertainty on tool life predictions is illustrated. As a result, uncertainty in the prior distributions is reduced, and the life prediction on average shows better prediction. In the second part, the research provides a Bayesian wear prediction method. A dynamic system is estimated through Bayesian approach with a state model and ta measurement model. The data taken from a Mastuura milling machine MC-510V under different machining setting which sampled by five different sensors two acoustic emission, two vibrations and one current sensors. The research compares the Bayesians result with Artificial Neural Network ANN using Levenberg–Marquardt algorithm. The results demonstrate the Bayesian approach particularly with limited prior information is effective. In conclusion, the Bayesian method improves the both life and wear predictions
  9. Keywords:
  10. Tool Wear ; Bayes Probablity Theorm ; Uncertainty ; Neural Network ; Monte Carlo Method ; Tools Life

 Digital Object List

 Bookmark

No TOC