File(s) under permanent embargo
Reason: Restricted by author. A copy can be supplied under Section 51(2) of the Australian Copyright Act 1968 by submitting a document delivery request through your library or by emailing firstname.lastname@example.org
A remote machine condition monitoring framework for intelligent fault diagnosis and prognosis
thesisposted on 14.02.2017, 02:14 by Yaqub, Muhammad Farrukh
Accurate remote equipment health monitoring systems can improve the efficiency of the agile processes e.g. agile manufacturing facilities, processing, mining and utilities etc. The impairments in industrial equipment are inevitable over the time. These impairments could be due to continuous meshing of the mechanical components, environment issues or nonidealistic operating conditions such as contaminated electrical supplies etc. Despite of deterioration in the industrial equipment, in order to ensure reliable process operation, a robust online condition monitoring system becomes need of the hour. An efficient online condition monitoring system can provide predictive repair maintenance schedule well ahead of the breakdown instances. In this research work, a framework for remote machine condition monitoring is presented to address the challenges in reliable data acquisition, inchoate fault diagnosis under varying industrial conditions and fault prognosis in terms of residual life prediction to minimise unplanned outages. Reliable and flexible data acquisition is necessary for continuous equipment monitoring in agile systems. In this thesis, wireless data acquisition, in the presence of co-existing networks, has been investigated for agile processes, e.g., shop floors of manufacturing facility where the underlying processes are prone to change both in orientation and type, due to new market trends. The problem of fault diagnosis essentially deals with categorizing the type of the fault arising in the machine, using the fault signatures captured from the wireless data acquisition unit. The fault diagnostic module comprises of signal processing, feature extraction and fault detection. The proposed framework optimally and adaptively tunes these constituents, particularly for inchoate, severity invariant machine fault diagnosis and multiple points-defect modelling. Fault prognosis demands continuous monitoring of the fault propagation to estimate the severity of the fault and hence, the residual life prediction of the equipment accordingly. The proposed residual life prediction scheme helps in achieving an optimized scheduling of maintenance to meet industrial demands, challenges and commitments. Proposed remote equipment health monitoring framework is flexible; and can be applied to industries e.g., mining, petroleum, utilities, transport, agriculture, space and building etc.