Система мониторинга продуктов горения угольных котлов по сигналам акустической эмиссии для предотвращения опасных аварийных ситуаций

Abstract

The system for monitoring combustion products in the coal-fired boilers is considered to prevent accidents during their operation. The method is based on the processing of the acoustic signal coming from the sensor, followed by the analysis of the obtained data. Acoustic signal from the moving dust cloud contains excess information about the flow dustiness. It can be used for determining the sources of an audio signal, the ways for emitting it, and the physical mechanisms of generation. Using the discrete fast Fourier transformation, two mathematical models for the analysis of dispersed dust were compiled — energy and amplitude-phase. The proposed spectral-timbre mathematical model for the time-series decomposition into a Fourier spectrum allows to obtain spectra consisting of the main and multiple timbre harmonics. Block diagram of the adaptive system for determining the concentration and disperse composition of the coal dust flow is presented (in the form of an algorithm). Concentration of the coal dust corresponding to one or another class of maximum permissible concentrations is determined by the pattern recognition method. Coefficients of the separating functions are calculated in accordance with the developed program of differential diagnostics. This problem is solved by the method of potential functions from the theory of pattern recognition. Training experiment was conducted on a closed circulation stand using the developed measuring complex. Obtained graphic map (acoustic diagram) visualizes the dust concentration zones depending on the frequencies and amplitudes of the Fourier spectral harmonics. This allows to clearly interpret the roles of the main and timbre harmonics of the spectrum. It also provides the possibility of a simple comparison of the concentrations of complex dust mixtures for the additive components of the spectrum of the analyzed acoustic signal.