The use of Bayesian Networks in Detecting the States of Ventilation Mills in Power Plants

Journal Title: Elektronika - Year 2014, Vol 18, Issue 1

Abstract

The main objective of this paper is to present a new method of predictive maintenance which can detect the states of coal grinding mills in thermal power plants using Bayesian networks. Several possible structures of Bayesian networks are proposed for solving this problem and one of them is implemented and tested on an actual system. This method uses acoustic signals and statistical signal pre-processing tools to compute the inputs of the Bayesian network. After that the network is trained and tested using signals measured in the vicinity of the mill in the period of 2 months. The goal of this algorithm is to increase the efficiency of the coal grinding process and reduce the maintenance cost by eliminating the unnecessary maintenance checks of the system.

Authors and Affiliations

Sanja Vujnović, Predrag Todorov, Željko Đurović, Aleksandra Marjanović

Keywords

Related Articles

Real Measurements and Evaluation of the Influence of Atmospheric Phenomena on FSO Combined with Modulation Formats

The influence of atmospheric environment is fundamental for Free-Space Optical link (FSO). The atmosphere can significantly degrade the communication quality of FSO up to so low received power/RSSI level that it can lead...

A Modified Approximate Internal Model-based Neural Control for the Typical Industrial Processes

A modification of the Approximate Internal Modelbased Neural Control (AIMNC), using Multi Layer Neural Network (MLNN) is introduced. A necessary condition that the system provides zero steady-state err...

New Single VDCC-based Explicit CurrentMode SRCO Employing All Grounded Passive Components

This paper proposes a new single resistance controlled sinusoidal oscillator (SRCO) which employs only one voltage differencing current conveyor (VDCC), two grounded resistors and two grounded capacitors. The presented c...

Improving the methodology of main power equipment choice for the gas turbine plants

Our paper considers the problem of economic substantiation of the choice of the main power equipment at the stage of functional studies of investment projects in conditions of uncertainty and incompleteness of initial da...

Tool-Based Curricula and Visual Learning

In the last twenty years nanotechnology has revolutionized the world of information theory, computers and other important disciplines, such as medicine, where it has contributed significantly in the creation of more soph...

Download PDF file
  • EP ID EP88635
  • DOI 10.7251/ELS1418016V
  • Views 92
  • Downloads 0

How To Cite

Sanja Vujnović, Predrag Todorov, Željko Đurović, Aleksandra Marjanović (2014). The use of Bayesian Networks in Detecting the States of Ventilation Mills in Power Plants. Elektronika, 18(1), 16-22. https://europub.co.uk./articles/-A-88635