VIBRATION DIAGNOSTICS OF ROTATING MACHINES
Academic year and teacher
If you can't find the course description that you're looking for in the above list,
please see the following instructions >>
- Versione italiana
- Academic year
- 2020/2021
- Teacher
- GIANLUCA D'ELIA
- Credits
- 6
- Didactic period
- Primo Semestre
- SSD
- ING-IND/13
Training objectives
- The course provides the basic methodology for the condition monitoring and diagnostics of rotating machines. In particular, the pivotal aspects of the diagnostics of industrial components such as gears and bearings are highlighted in the course. Information concerning the status of the particular mechanical component are extracted for the vibration signal by means of advanced signal processing techniques working in several domains, such as: time, angle, time-frequency and cycliclostationary.
Prerequisites
- Courses of Mathematics, Physics, Vibration mechanics
Course programme
- Signal classification and analog to digital conversion. Continuous-time signal analysis, Fourier Transform, Autopower Spectrum, PSD, Short-Time Fourier Transform, amplitude/phase modulation and Hilbert transform. Discrete-time signal analysis, Discrete Fourier Transform, Leakage, windowing and spectrum estimators. (about 5 h)
Angular domain sampling, run-up analysis and Time Synchronous Average. Decomposition of the energy flow (cyclostationarity). (about 2.5 h).
Diagnostics of rotor unbalance and couplings. Diagnostics of localised and distributed faults in gears and bearings (about 5 h)
Introduction to machine learning for the classification of the diagnostic parameters, such as: Bayesian statistics, kNN, neural network. (about 5 h)
Laboratory (about 42.5 h): development of pseudo and Matlab codes for the diagnostics of rotating machines. Students will analysis simulated signals as well as real signals of sound and faulty mechanical transmissions. Didactic methods
- The course includes: theoretical lectures made with the aid of multimedia systems, training and laboratory activities concerning the diagnostics of rotating machines.
Learning assessment procedures
- The exam includes an oral test subdivided in three questions, two of them regarding the theoretical part of the courses and the last one concerning the practical section, i.e. the development of a pseudo or Matlab code of a particular signal processing technique. The exam is passed if the score of each question is at least sufficient. The final mark is the arithmetic mean of the marks obtained in the three questions.
Contest (not mandatory). During the course will be assigned to the students a real industrial case concerning the diagnostics of an ordinary gearbox. The contest will be divided in two sections: i) diagnostics of an ordinary gearbox, ii) implementation in Matlab environment of an advanced signal processing technique. The contest will have a strong influence on the final mark of the exam, however IS NOT MANDATORY. Reference texts
- Robert Bond Randall, Vibration-based Condition Monitoring: Industrial, Aerospace and Automotive Applications, Wiley
Andres Brandt, Noise and Vibration Analysis: Signal Analysis and Experimental Procedures, Wiley
Christopher M. Bishop, Pattern Recognition and Machine Learning, Springer
