Towsyfyan, Hossein (2017) Investigation of the Nonlinear Tribological Behaviour of Mechanical Seals for Online Condition Monitoring. Doctoral thesis, University of Huddersfield.

Mechanical seals have increasingly been used for sealing rotating shafts in centrifugal pumps, propeller shafts in ships and submarines, compressors, liquid propellant rocket motors in aerospace industry, pumps, turbines, mixers and many other rotating machines during last two decades. Abnormal operating conditions in the mechanical seals will degrade machine performance, increase operating cost and may cause unexpected sudden failures which are dangerous in both engineering and safety terms. Hence it is necessary to investigate the tribological behaviour of mechanical seals operating based on nonlinear coupling between fluid and surface dynamics, in order to develop more advanced diagnostic technologies to improve the reliability of such machines operating with mechanical seals.

Different condition monitoring techniques have been studied to evaluate the lubrication state and severity of contact between the mating faces in mechanical seals. However, some of them are not cost effective others are not practical in industrial applications. Acoustic emission (AE) has been proved to be a sensitive indicator of lubrication conditions and changes in the lubricant properties, however the application of technique for identification of lubrication regimes in mechanical seals has not been reported yet. Moreover, previous studies give relatively little information to acoustic emission condition monitoring of mechanical seals, nor has comprehensive fault detection been implemented for a particular case. In addition, a review on previous works reveals the lack of comprehensive mathematical models to explain the relationship between AE energy and tribological characteristics of the mating faces under healthy and faulty conditions.

In this research, the tribological behaviour of mechanical seals is investigated using acoustic emission measurements to pave a way for fault detection at early stage. Three common seal failures i.e. dry running, spring fault, and defective seal are studied in this thesis. The main objective is to extract AE features that can explain the tribological behaviour of mechanical seals under both healthy and faulty conditions. To achieve this, a purpose-built test rig was employed for collecting AE signals from the mechanical seals. Then, the collected data was processed using time domain, frequency domain and time frequency domain analysis methods which are of the most common techniques used for monitoring in AE applications.

Based on results the main frequency band that can present the tribological behaviour of mechanical seals was detected. Also it has been proved that AE features in time domain and frequency domain can be effectively applied to indicate the lubrication condition of mechanical seals as well as early fault detection.

Moreover, mathematical models were developed to establish a relationship between AE root mean square (RMS) value of AE signals and working parameters of seals (rotational speed, load and number of asperities in contact) under different lubrication regimes. A good agreement was achieved between measured and predicted signals that gives a good evidence of the effectiveness of proposed models. Especially in case of leakage that is one of the main situations indicating the seal failure, a significant difference was observed between the predicted signal for healthy case and the measured signal under faulty conditions.

Therefore, it can be deduced that the AE measurement system and signal processing developed in this work has a promising potential to be used to diagnose and monitor the mechanical seals online.

Finally, the conclusions and achievements are given based on the entirety of this research work, and online monitoring incorporating with AE features and mathematical models developed in this thesis are suggested as the main works for further research.

FINAL THESIS - Towsyfyan.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (7MB) | Preview


Downloads per month over past year

Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email