Feature And Extraction Method Of Electrostatic Signals Induced From Gas-Solid Two-Phase Flow

Gas-solid two phase flow widely exists in nature and industrial production process.The accurate measurement of the parameters is helpful to improve the production efficiency,reduce energy consumption and environmental pollution.Gas-solid two phase flow has the characteristics of complex flow and uneven distribution of parameters,so its parameter measurement is difficult.In recent decades,the detection method based on electrostatic method has achieved rapid development in theory and application.At present,the inner flush-mounted electrostatic sensor,with simple structure,good wear resistance,good pressure resistance and high sensitivity,is the main type of electrostatic sensor used in industrial field.The measuring signal from the inner flush-mounted electrostatic sensor includes induced charge signal and transferred charge signal.Owing to the different sources,the induced and transferred charge signals can represent different characters of the gas-solid two-phase flow.The effective decoupling of the two kinds of signals is of great value for improving measurement accuracy and reliability.This paper focuses on the extraction of induced signal.1.COMSOL software is used to model and analyze the inner flush-mounted electrostatic sensor.And the spatial sensitivity distribution of the sensor is obtained.Then the model of induced charge signal is established.It is found that the induced charge signal spectrum is a single-peak curve with certain bandwidth from simulation.Meanwhile,the characteristics of frequency band are subjected to the velocity and position of the particle.When the particle moves faster and is closer to the electrode,the bandwidth range of the induced charge signal is wider and moves to higher frequency.The experimental results of the single particle experiment device and belt device are in agreement with the simulation results.2.The comparison experiments between the outer surface-mounted electrostatic sensor and inner flush-mounted electrostatic sensor was conducted on the gas-solid two-phase platform in Tianjin University.It is found that the induced charge signals of the measuring signals from outer surface-mounted electrostatic sensor have obvious band characteristics and the transfer charge signal occupies an important proportion in the whole signal.3.The induced charge signal mainly reflects the general information of the flowing particle in pipe.Because of the short distance between upstream and downstream electrodes,the spatial distribution of particles can keep similarly.Thus the similarity is high between the induced charge signals.In contrast,the transferred charge signals from the upstream and downstream electrodes have poor similarity because of the random collision between particles and electrodes.Based on these findings,this paper presents an adaptive signal decomposition method based on discrete wavelet transform(DWT).The correlation coefficient of the signal component obtained by DWT decomposition is the choosing criterion for the reconstruction of the induced charge signal.Thus,the measured signal can be decomposed adaptively into induced charge signal and transferred charge signal.4.The signals from upstream and downstream electrodes were used to calculate the cross-correlation velocity.After signal decomposition,the mean value of cross-correlation coefficient was increased from 0.48 using measured signals to 0.75 using induced charge signals,indicating a higher accuracy of the measurement.The relative standard deviation(RSD)of cross-correlation velocity was decreased from 3.35% to 2.69%,leading to higher stability of measurement.The root-mean-square(RMS)of decomposed transferred charge signals increases linearly with solid mass flow rate,which offers an effective method to predict the solid mass flow rate.