Skip to main content

Advanced Valve Control System

Article Category

Article available in the folowing languages:

Monitoring and diagnosing machinery

Rotating machinery constitutes the heart of many industrial technologies. An important prerequisite for its proper use is the implementation of condition monitoring programmes. Due to the application difficulties and the increased subjectivity involved with such programmes, so far European industries rarely adopt them and if they do, they are unsuccessful. A European consortium of partners representing a wide variety of industries developed an advanced and very promising diagnostic system for condition monitoring of machinery that achieves increased equipment reliability.

Industrial Technologies

Benchmarking studies on the maintenance of industrial equipment revealed that the implementation of condition-machinery programmes is highly beneficial for industrial productivity. These programmes consist of the collection of data measurements from machinery and their analysis for the diagnosis of the condition of that machinery before serious operation faults develop. Such maintenance measures result not only in the minimisation of failures, but also in the prevention of their possible occurrence. Nevertheless, this good practice has not been extensively adopted yet and when it has, it has not been properly applied. This is mainly due to the fact that it is a highly complicated method that involves increased subjectivity in data analysis, mainly for vibration data. Unfortunately, any effort made for overcoming these problems has already failed. Motivated by their own needs, a wide spectrum of large European industries joined forces, and coordinated their experience and resources to develop a robust and reliable system for the diagnosis of industrial equipment, namely VISION. Initially, the consortium used simulation techniques to develop appropriate models of industrial plants that were further validated with real data sets. With the aid of specific artificial intelligence techniques, the simulation models were used to train a developed software system about the specificities of the plant. Finally, the models were decoupled from the adaptively trained system and it was tested on the recognition of different fault types from various machines. The end result is the diagnostic system that has achieved 98.4% accuracy on previously undetectable data. The software can be easily implemented on a Windows interface to analyse and interpret data from a plant and to locally advise maintenance and plant engineers in a wide range of industrial sectors. The system will be further developed to provide diagnostic services remotely via the Internet or Intranet remote servers. This intelligent, adaptive monitoring and diagnostic system exhibits increased marketability capabilities. Recent surveys revealed that 84% of the market share is available to adopt a reliable and simple to use diagnostic system to perform the complicated data analysis. Therefore, the VISION system not only promotes the industrial good practice of condition monitoring of machinery, but it is a useful tool for its successful implementation.

Discover other articles in the same domain of application