The project investigates the feasibility of using sensor-based detection and processing systems to provide a reliable means of monitoring rolling contact fatigue (RCF) wear failures of silicon nitride in hybrid bearings. To fulfil this investigation, a decision was made early in the project to perform a series of hybrid rolling wear tests using a twin disc machine modified for use on hybrid bearing elements. The initial part of the thesis reviews the current understanding of the general wear mechanisms and RCF with a specific focus to determine the appropriate methods for their detection in hybrid bearings. The study focusses on vibration, electrostatic and acoustic emission (AE) techniques and reviews their associated sensing technologies currently deployed with a view of adapting them for use in hybrids. To provide a basis for the adaptation, an understanding of the current sensor data enhancement and feature extraction methods is presented based on a literature review. The second part describes the test equipment, its modifications and instrumentation required to capture and process the vibration, electrostatic and AE signals generated in hybrid elements. These were identified in an initial feasibility test performed on a standard twin disc machine. After a detailed description of the resulting equipment, the thesis describes the calibration tests aimed to provide base data for the development of the signal processing methods. The development of the signal processing techniques is described in detail for each of the sensor types. Time synchronous averaging (TSA) technique is used to identify the location of the signal sources along the surfaces of the specimens and the signals are enhanced by additional filtering techniques. The next part of the thesis describes the main hybrid rolling wear tests; it details the selection of the run parameters and the samples seeded with surface cracks to cover a variety of situations, the method of execution of each test run, and the techniques to analyse the results. The research establishes that two RCF fault types are produced in the silicon nitride rolling element reflecting essentially different mechanisms in their distinct and separate development; i) cracks, progressing into depth and denoted in this study as C-/Ring crack Complex (CRC) and ii) Flaking, progressing primarily on the surface by spalls. Additionally and not reported in the literature, an advanced stage of the CRC fault type composed of multiple and extensive c-cracks is interpreted as the result of in duced sliding in these runs. In general, having reached an advanced stage, both CRC and Flaking faults produce significant wear in the steel counterface through abrasion, plastic deformation or 3-body abrasion in at least three possible ways, all of which are described in details.

A recent series of tapered roller bearing tests have been conducted at the University of Southampton to evaluate the effectiveness of using multiple sensing technologies to detect incipient faults. The test rig was instrumented with on-line sensors including vibration, temperature and electrostatic wear and oil-line debris sensors. Off-line techniques were also used such as debris analysis and bearing surface examination. The electrostatic sensors, in particular, have the potential to detect early decay of tribological contacts within rolling element bearings. These sensors have the unique ability to detect surface charge associated with surface phase transformations, material transfer, tribofilm breakdown and debris generation. Thus, they have the capability to detect contact decay before conventional techniques such as vibration and debris monitoring. However, precursor electrostatic events can not always be clearly seen using time and frequency based techniques. Therefore, an intelligent system that can process signals from multiple sensors is needed to enable early and automatic detection of novel events and provide reasoning to these detected anomalies. Operators could then seek corroborative trends between sensors and set robust alarms to ensure safe running. This has been the motivation of this study.

- / -

M. Tech. (Mechanical Engineering) Vaal University of Technology / Monitoring the condition of equipment in industry is very important to prevent unplanned breakdowns and to prolong their life. This is necessary, since it is not always economically viable to stop equipment at regular intervals to do maintenance. Failure on machines can lead to high repair costs and production losses. It is thus of paramount importance that early failure symptoms be identified by means of condition monitoring. This study in the field of condition monitoring is performed to determine if the mechanical shock values induced in defect bearings could be used to measure the condition of a bearing while operating under variable speed and variable load. Variable speed and variable load is becoming more popular in industry because variable speed drives applications ensure effective process control. Variable speed application, cause fault frequencies to fluctuate and therefore vibration applications for constant speed applications, which are speed-dependent, can no longer apply. Vibration-monitoring techniques that have applied for many years have now become obsolete in these variable speed applications. Methods such as Short Time Fourier Transformation (STFT), time scale like wavelet transform, and Order tracking has been applied in variable speed applications with some success. These methods analyses the vibration phases on the signal buy compensating for the speed changes. In this thesis, the Shock pulse method is selected as the analyses tool to measure the mechanical shock. Shock pulse monitoring does not focus on the vibration phases but measures in a small-time window when mechanical shocks are induced in the bearing material before the vibration phase. There is very little documented research in the field of mechanical shock pulse monitoring for conditions of variable speed and variable loads, and therefore this research focuses on recording these mechanical shock values by empirical tests. The tests were performed on a bearing with an induced defect on the outer race. The rolling element of the bearing strikes the defect and the mechanical shock value (dBsv) is measured. The mechanical shock is measured with the Shock pulse method in a small-time window before vibration occurs. In this time window, the dBsv is recorded over time to provide diagnostic information of the bearing during acceleration, deceleration and various loading conditions. These mechanical shocks are elastic waves that mirror the impact-contact-force's time function and the Shock pulse monitoring accelerometer, which is tuned to 32 kHz, will respond to the elastic wave fronts with transient amplitudes proportional to the square of the impact velocities. The mechanical shock values were analysed and reoccurring fault levels were identified on each empirical test. These recurring events from the empirical tests were used as primary data for analysis in this research. These tests were performed on a bearing with an induced failure and it was found that the dBsv measured over time could not be used to monitor the condition of the bearing under variable speed applications. This was because the dBsv changed as the speed increased. To overcome this problem Sohoel’s theory was applied and the initial mechanical shock value (dBi) was calculated for the bearing. The dbi value was subtracted from the dBsv and a value called the maximum mechanical shock value (dBm) was obtained. The dBm values stayed constant for the duration of the test and this allowed the condition of the bearing to be measured under variable speed and variable load conditions with some exception. The exception to the findings was that the dBm values stayed constant during acceleration phases, but during the deceleration phases the values were erratic and scattered. At speed below 200rpm the dBm values did not stay constant and therefore it was concluded that the dBm value recorded the best results only when thrust on the bearing was maximum. The other exception was under no-load conditions. The values were erratic and scattered, and therefore the results were not a true reflection of the bearing condition. The third exception was that the results on bearings with various loads remained constant during increased load changes unless the loading was erratic. During erratic load changes, the results were affected. The results also indicated that the larger the defect on the bearing raceway, the higher the dBm values were. Multipil defects on the bearing race ways were not part of this thesis and this gives an opertunity for futher research. The Shock pulse monitoring technique was 100% successful in monitoring the bearing condition only while the speed of the bearing was increasing. The results obtained in this work demonstrated that the condition of bearings can be monitored in applications of variable speed and variable load if the exception are eliminated and to obtain conclusive results the mechanical shock pulses should be measured over time and not be used as once-off value.

The focus of this thesis was the development of failure prognosis methods (prognostics) in rotating machinery with use of multiple sensors digital signal processing and machine learning techniques. The motivation stems from the void in literature concerning prognostics in meshing gearboxes. Moreover, there are several but inconclusive works regarding bearing prognosis. Few research groups have studied multi-hour fatigue gear experiments and this was one of the contributions of this thesis. Moreover, the study expanded beyond the sheer application of vibration monitoring with the addition of an Oil Debris Monitoring probe (ODM) as well as Acoustic emission (AE). The method of AE monitoring is, once again, proposed as a robust technique for failure prognosis being better correlated with gear pitting level compared to the classic vibration monitoring technique. Moreover, judging from ODM recordings the gear pitting comprises of two phases i) a linear phase, with an almost constant pitting rate and ii) a very short non linear phase where the pitting rate increases exponentially, an explicit indication of a critical failure. Multi-hour gear experiments that are close to real scale applications are very demanding in time as well as in invested capital. To bypass this shortfall a gear failure like simulation is proposed. The simulation framework is based on real life experiments and is applied to assess a number of data-driven Remaining Useful Life (RUL) estimation techniques namely i) Proportional Hazards Μodel (PHM), ii) ε- Support Vector Regression ε-SVR and iii) Exponential extrapolation based on bootstrap sampling. In the current thesis a feature extraction scheme for prognosis is proposed and assessed based on time domain, frequency domain statistical features and Wavelet Packet (WP) energy derived from AE and vibration recordings. ICA is proposed as a preferable fusion technique for gear failure prognostics. Application of ICA for feature fusion provided a clear improvement regarding the earlier presented bootstrap extrapolation technique. Bearings are also taken into account since they are closely connected to gearboxes. In the current thesis a wavelet denoising method is proposed for bearing vibration recordings aiming to the improvement of the diagnostic and prognostic potential of vibration. Finally the importance of data fusion is highlighted in the case of bearings. It is observed that a feature extraction scheme can generalize the application of prognostics, even in cases where RMS may yield no important degradation trend. / Η παρούσα εργασία εστιάζεται στην ανάπτυξη μεθοδολογιών πρόβλεψης τελικής αστοχίας σε περιστρεφόμενα συστήματα με χρήση πολλαπλών αισθητήρων και μεθόδων μηχανικής μάθησης και επεξεργασίας σήματος. Το κίνητρο προήλθε από το κενό που υπάρχει στη βιβλιογραφία όσον αφορά την προγνωστική σε κιβώτια ταχυτήτων. Η προγνωστική σε έδρανα έχει μεν μελετηθεί αλλά σε μικρό βαθμό και η παρούσα εργασία έρχεται να συμβάλλει και σε αυτό τον τομέα. Στα πλαίσια αυτής της εργασίας εκπονήθηκε ένας αριθμός πειραμάτων κόπωσης κιβωτίων ταχυτήτων. Η μελέτη επεκτάθηκε πέραν της παρακολούθησης κατάστασης με τη μέθοδο των κραδασμών και συγκεκριμένα μελετήθηκαν καταγραφές σωματιδίων σιδήρου στο λιπαντικό (ODM) καθώς και Ακουστική Εκπομπής (AE). Η μέθοδος ΑΕ ευρέθη πιο στενά συσχετισμένη με τη σταδιακή υποβάθμιση της ακεραιότητας του κιβωτίου ταχυτήτων σε σχέση με τις καταγραφές κραδασμών. Επίσης με βάση τις καταγραφές του αισθητήρα σωματιδίων σιδήρου διακρίθηκαν δύο στάδια  υποβάθμισης i) μια γραμμική περιοχή με σχεδόν σταθερό ρυθμό απελευθέρωσης υλικού από την επιφάνεια των δοντιών και ii) μια σύντομη αλλά έντονα μη γραμμική αύξηση στο ρυθμό αυτό πολύ κοντά στο τέλος της λειτουργίας του κιβωτίου. Tα πολύωρα πειράματα κόπωσης σε γρανάζια είναι πολύ απαιτητικά. Για να παρακαμφθεί αυτή η δυσκολία αναπτύχθηκε ένα φαινομενολογικό μοντέλο για αναπαραγωγή χρονοσειρών που ομοιάζουν σε καταγραφές γραναζιών σε κόπωση. Το μοντέλο αυτό στηρίχθηκε σε πραγματικά πειράματα κόπωσης. Έτσι έγινε δυνατό να εξεταστούν και να συγκριθούν ένας αριθμός μεθοδολογιών εκτίμησης εναπομένουσας ζωής και συγκεκριμένα i) Proportional Hazards Model (PHM), ii) ε- Support Vector Regression ε-SVR και iii) Exponential extrapolation βασισμένο σε μια διαδικασία bootstrap sampling. Στην παρούσα μελέτη προτείνεται ένα σύνολο παραμέτρων προερχόμενο από το πεδίο της συχνότητας, του χρόνου και κυματοπακέτων. Αυτό, συνδυαζόμενο με μια διαδικασία σύμπτυξης δεδομένων (ανάλυση σε πρωταρχικές και ανεξάρτητες συνιστώσες) αξιοποιείται για πρόγνωση σε γρανάζια σε κόπωση. Η τεχνική ανεξάρτητων συνιστωσών προτείνεται σαν προτιμότερη από τη σκοπιά της προγνωστικής καθώς βελτιώνει την εκτίμηση της εναπομένουσας ζωής. Η εργασία επεκτάθηκε και σε έδρανα κύλισης. Προτάθηκε μια διαδικασία wavelet denoising η οποία ενισχύει τόσο τη διαγνωστική όσο και την προγνωστική δυνατότητα του αισθητήρα κραδασμών. Τέλος, η σημασία της εξαγωγής παραμέτρων υπογραμμίζεται και στην περίπτωση της προγνωστικής σε έδρανα. Συνδυάζοντας πολλαπλές παραμέτρους και αισθητήρες κραδασμών μαζί με ένα μοντέλο ε-SVR παρέχεται ένα ολοκληρωμένο μοντέλο πιθανοτικής εκτίμησης εναπομένουσας ζωής σε έδρανα κύλισης ακόμα και σε περιπτώσεις που η τιμή RMS των κραδασμών δεν παρέχει πληροφορία.

Η παρούσα διδακτορική διατριβή αναφέρεται σε έδρανα ολίσθησης. Συγκεκριμένα, αναφέρεται στο σχεδιασμό και την κατασκευή ενός νέου, πρωτότυπου υβριδικού εδράνου ολίσθησης, το οποίο έχει τη δυνατότητα να λειτουργεί είτε ως υδροδυναμικό, είτε ως ενεργό ηλεκτρομαγνητικό, είτε ως υβριδικό, δηλαδή υδροδυναμικό και ενεργό ηλεκτρομαγνητικό ταυτόχρονα. Προκειμένου να πραγματοποιηθεί ο σχεδιασμός και η κατασκευή του συγκεκριμένου υβριδικού εδράνου ολίσθησης, έγιναν υπολογισμοί και προσομοίωση του υδροδυναμικού πεδίου για τα υδροδυναμικά έδρανα ολίσθησης και του ηλεκτρομαγνητικού πεδίου για τα ηλεκτρομαγνητικά έδρανα στο πρόγραμμα ANSYS. Η προσομοίωση των υβριδικών εδράνων ολίσθησης έγινε με επαναληπτική διαδικασία μεταξύ του υδροδυναμικού και του ηλεκτρομαγνητικού πεδίου προκειμένου να υπολογίζεται το κάθε επιθυμητό μέγεθος. Για το σχεδιασμό του υβριδικού εδράνου ολίσθησης χρησιμοποιήθηκε το σχεδιαστικό πρόγραμμα CATIA. Για την προσομοίωση του ελέγχου του υβριδικού εδράνου ολίσθησης, χρησιμοποιήθηκε το πρόγραμμα Matlab, το οποίο ενεργοποιεί το πρόγραμμα Ansys για τον υπολογισμό των χαρακτηριστικών του υδροδυναμικού και του ηλεκτρομαγνητικού πεδίου του εδράνου και το Simulink module του Matlab για τον υπολογισμό των χαρακτηριστικών του ελέγχου, ώστε να γίνονται όλοι οι απαραίτητοι υπολογισμοί αυτοματοποιημένα. Τέλος, έγιναν υπολογισμοί και προσομοίωση δυναμικής περιστρεφόμενων αξόνων, με στήριξη του ρότορα σε δύο έδρανα και δίσκο προσαρμοσμένο στον ρότορα, στο πρόγραμμα Matlab, με σύνδεση με το Ansys για τον υπολογισμό των συντελεστών ελαστικότητας και απόσβεσης και το Simulink module του Matlab για τον υπολογισμό των χαρακτηριστικών του ελέγχου, με σκοπό την ολοκληρωμένη ανάλυση του συστήματος. Συμπερασματικά, μέσω της συγκεκριμένης διδακτορικής διατριβής αποδεικνύεται πως ο σχεδιασμός, η λειτουργία και η κατασκευή του συγκεκριμένου πρωτότυπου υβριδικού εδράνου ολίσθησης είναι εφικτά, και από τα αποτελέσματα προκύπτει πως η λειτουργία ενός τέτοιου υβριδικού εδράνου ολίσθησης, έχει πολλά πλεονεκτήματα σε σχέση με άλλα έδρανα στήριξης περιστρεφόμενων αξόνων, είτε αυτά είναι απλά είτε υβριδικά και μπορεί να εφαρμοστεί είτε σε εργαστηριακό επίπεδο είτε σε βιομηχανικές εφαρμογές. / The present doctoral thesis refers to hydrodynamic journal bearings. Specifically, it refers to the design and construction of a new, innovative hybrid journal bearing, which has the ability to function either as hydrodynamic or active magnetic or hybrid, i.e. both hydrodynamic and active magnetic, at the same time. In order to be performed the design and construction of the specific hybrid journal bearing, calculations and simulation of the hydrodynamic field for the hydrodynamic journal bearings in the program ANSYS, have been made. The simulation of the hybrid journal bearings has been made with iterative process, between the hydrodynamic and active magnetic field, so that every desired magnitude can be calculated. For the design of the hybrid journal bearing, the designing program CATIA has been used. For the simulation of the control of the hybrid journal bearing, the program Maltab has been used, which activates the program Ansys, for the calculation of the features of the hydrodynamic and active magnetic field of the bearing and the Simulink module of Maltab, for the calculation of the features of control, so that all necessary calculations can be made automated. Finally, calculations and simulation of rotor dynamics, with support of the rotor in two bearings and disc adjusted to the rotor, in Maltab, in connection to Ansys, for the calculation of the elastic and damping coefficients and the Simulink module of Maltab, for the calculation of the features of control, have been made, having as a purpose the complete analysis of the system. In conclusion, through the specific doctoral thesis, it is proved that the design, function and construction of the specific, new hybrid journal bearing, can be achieved and the results show that the function of such a hybrid journal bearing, has many advantages compared to other bearings of support of rotors, either they are simple or hybrid and it can be applied either to laboratory level or industrial applications.