On-line Condition Monitoring System Research Articles

An On-line Condition Monitoring and Diagnosis System for Feed Rolls in the Plate Mill

An on-line condition monitoring and diagnosis system for feed rolls was developed for providing early warnings of malfunctions so as to avoid great losses resulting from unexpected shutdown of the production line. This system measures the bearing vibration signals on-line and judges the feed roll condition automatically according to the diagnosis rules stored in the computer. These rules were established based on results of long-term vibration measurements and proved by many real cases to be very effective and reliable. The derivation of these diagnosis rules and the development of the on-line monitoring and diagnosis system based on these rules are depicted in this article.

Electrostatic monitoring of oil lubricated sliding point contacts for early detection of scuffing

Electrostatic sensors have been used to investigate scuffing events in oil lubricated sliding point contacts. A pin-on-disc and a reciprocating laboratory wear rig have been used for this investigation in a fixed ball-on-disc configuration. Electrostatic charging events have been detected up to 1200 s prior to severe scuffing and have been shown to coincide with microstructural changes in the contact and regions of increased wear. The experiments presented demonstrate that the electrostatic detection technology has the potential to be used as the basis of an on-line (real-time) machinery condition monitoring system.

An on-line condition monitoring system linked by fibre-optic digital network

This paper introduces the basis structure and major function of an on-line condition monitoring system linked by a fibre-optic digital network. Fibre-optic digital networks have obvious advantages to larger enterprises with various locations for condition monitoring data transmission, which not only is an information highway between the main station (diagnosis enter) and substation (monitoring locations), but also combines the monitoring functions into one system. This is a two-level condition monitoring system. The first level monitoring is realized by a sub-station and the second level monitoring is realized by the main station. The system has strong hardware functions and software functions, which became a new model of real-time condition monitoring and was used in many oil refineries.

On-line tool condition monitoring system with wavelet fuzzy neural network

In manufacturing systems such as flexible manufacturing systems (FMS), one of the most important issues is accurate detection of the tool conditions under given cutting conditions. An investigation is presented of a tool condition monitoring system (TCMS), which consists of a wavelet transform preprocessor for generating features from acoustic emission (AE) signals, followed by a high speed neural network with fuzzy inference for associating the preprocessor outputs with the appropriate decisions. A wavelet transform can decompose AE signals into different frequency bands in the time domain. The root mean square (RMS) values extracted from the decomposed signal for each frequency band were used as the monitoring feature. A fuzzy neural network (FNN) is proposed to describe the relationship between the tool conditions and the monitoring features; this requires less computation than a back propagation neural network (BPNN). The experimental results indicate the monitoring features have a low sensitivity to changes of the cutting conditions and FNN has a high monitoring success rate in a wide range of cutting conditions; TCMS with a wavelet fuzzy neural network is feasible.

Parametric Modeling of Exhaust Gas Emission From Natural Gas Fired Gas Turbines

Increased focus on air pollution from gas turbines in the Norwegian sector of the North Sea has resulted in taxes on CO2. Statements made by the Norwegian authorities imply regulations and/or taxes on NOx emissions in the near future. The existing CO2 tax of NOK 0.82/Sm3 (US Dollars 0.12/Sm3) and possible future tax on NOx are analyzed mainly with respect to operating and maintenance costs for the gas turbine. Depending on actual tax levels, the machine should be operated on full load/optimum thermal efficiency or part load to reduce specific exhaust emissions. Based on field measurements, exhaust emissions (CO2, CO, NOx, N20, UHC, etc.) are established with respect to load and gas turbine performance, including performance degradation. Different NOx emission correlations are analyzed based on test results, and a proposed prediction model presented. The impact of machinery performance degradation on emission levels is particularly analyzed. Good agreement is achieved between measured and predicted NOx emissions from the proposed correlation. To achieve continuous exhaust emission control, the proposed NOx model is implemented to the on-line condition monitoring system on the Sleipner A platform, rather than introducing sensitive emission sensors in the exhaust gas stack. The on-line condition monitoring system forms an important tool in detecting machinery condition/degradation and air pollution, and achieving optimum energy conservation.

COMOS-an online system for problem-orientated vibration monitoring

The monitoring of the time dependence of the vibrational behaviour of mechanical components is essential for malfunction detection. It contrast to passive primary components of PWRs like pressure vessels and steam generators, mechanical failures in rotating equipment may develop rapidly. An on-line Condition Monitoring System (COMOS) has been developed for main coolant pump shaft vibration monitoring as well as for general vibration monitoring of passive primary components. The monitoring procedure is based on the detection of deviations from reference signatures. For this purpose, a statistical discriminant is used which provides increased sensitivity for the detection of spectral changes. To qualify observed spectral changes, the frequency and amplitude of characteristic resonance are used as additional monitoring parameters. During the development phase of COMOS, comprehensive tests were carried out to test the suitability and sensitivity of various statistical discriminants or similarity measures to detect spectral changes. Although the discriminants are global measures of spectral change and as such unsensitive, an application to narrow frequency intervals enables some discriminants to detect spectral change sensitively, results in a significant feature extraction and data reduction and allows the integration of results obtained from a detailed correlation analysis. It was found that deterministic influences, such as operational influences, normal changes of vibrational behaviour during a fuel cycle and electrical disturbances might cause false alarms in an on-line system if they are not properly dealt with. To account for the redundancy of information in different signals and to incorporate the causal relationship between different primary components, a linkage of signatures was introduced. This new approach allows a significant incorporation of a knowledge base using normal computational methods and gives the system diagnostic potential. Furthermore, with this principle and the application of a time criterion, most transient deterministic influences could be coped with. A modular hardware and software concept is discussed. A high degree of automation and information extraction and a user-friendly representation of measurement results allow utility people to apply early failure detection methods without continuous expert assistance. Finally operational experience and system acceptance is highlighted.

Condition Monitoring Of Railway Single-throwEquipment

This paper presents the research carried out to improve the reliability of singlethrow equipment used in railway signalling, rolling stock and power systems. The motivation is to achieve improved operational availability and quality of service to passengers as well as the need for more cost-effective and efficient maintenance management. A number of railway equipment have so far been considered such as point machines, train-stops, DC Circuit Breakers (DCCB) and train doors. In this paper, the DCCB is presented as a case study for the development of an on-line condition monitoring system providing early-failure warning to maintainers.