Monitoring of vibrating machinery using artificial neural networks
- Tennessee Univ., Knoxville, TN (United States). Dept. of Nuclear Engineering
- Tennessee Univ., Knoxville, TN (United States). Dept. of Nuclear Engineering Oak Ridge National Lab., TN (United States)
The primary source of vibration in complex engineering systems is rotating machinery. Vibration signatures collected from these components render valuable information about the operational state of the system and may be used to perform diagnostics. For example, the low frequency domain contains information about unbalance, misalignment, instability in journal bearing and mechanical looseness; analysis of the medium frequency range can render information about faults in meshing gear teeth; while the high frequency domain will contain information about incipient faults in rolling-element bearings. Trend analysis may be performed by comparing the vibration spectrum for each machine with a reference spectrum and evaluating the vibration magnitude changes at different frequencies. This form of analysis for diagnostics is often performed by maintenance personnel monitoring and recording transducer signals and analyzing the signals to identify the operating condition of the machine. With the advent of portable fast Fourier transform (FFT) analyzers and laptop'' computers, it is possible to collect and analyze vibration data an site and detect incipient failures several weeks or months before repair is necessary. It is often possible to estimate the remaining life of certain systems once a fault has been detected. RMS velocity, acceleration, displacements, peak value, and crest factor readings can be collected from vibration sensors. To exploit all the information embedded in these signals, a robust and advanced analysis technique is required. Our goal is to design a diagnostic system using neural network technology, a system such as this would automate the interpretation of vibration data coming from plant-wide machinery and permit efficient on-line monitoring of these components.
- Research Organization:
- Tennessee Univ., Knoxville, TN (United States). Dept. of Nuclear Engineering
- Sponsoring Organization:
- USDOE; USDOE, Washington, DC (United States)
- DOE Contract Number:
- FG07-88ER12824; AC05-84OR21400
- OSTI ID:
- 6922588
- Report Number(s):
- CONF-9109447-1; ON: DE93003566
- Resource Relation:
- Conference: 2. government neural network applications workshop, Huntsville, AL (United States), 10-12 Sep 1991
- Country of Publication:
- United States
- Language:
- English
Similar Records
Rotating machinery surveillance system reduces plant downtime and radiation exposure
Condition Monitoring of Wind Turbine Planetary Gearboxes Under Different Operating Conditions