v基于电荷感应原理的矿井粉尘浓度测量方法研究

山东科技大学硕士学位论文摘要AbstractThe measurement of dust concentration at coal mine is very important for the safety incoal mine production and the life safety of the coal mine workers.In this paper,a new methodbased on electrostatic induction is researched.The basic principle is as follows:The coal minedust will be charged when the dust is produced and moved,and there are some relations withthe charge and the dust concentration;the dust concentration will be obtained by using asensing element which is sensitive to the static to measure the electric charges of the dust.The measurement method includes two parts:the inductive charge single acquisitionsystem and the digital signal processing method.The acquisition system contains a chargesensitive preamplifier and a sensing element which consists of a circular ring electrode,aninsulation-tube and a shield.The geometry and construction of the sensing element areoptimized by ANSYS.A dynamic equation is set up in order to study the dynamic response ofthe sensing element.The single acquisition system is experimented in the coal mine dustsimulate experiment system.The digital signal of the inductive charge with different dustdensity is collected.For the digital signal processing part,the time-frequency properties of thesignal are studied firstly.Based on the time-frequency properties,characteristics of the signalare extracted.Factor Analysis is applied to analyze the influence of the characteristics to thedust concentration.Three characteristics which have more influence than others are chosen forthe input of the BP neural network optimized by the Genetic algorithm.With this method,therelationship between the inductive charge single and the dust concentration is built.The calculation result and error analysis shows that this measurement method has highprecision up to the application standard.And it has great perspective and practicability withthe advantage of Maintenance Free and online monitoring.Keywords:coal-dust concentration,electrostatic induction,ANSYS,factor analysis,artificialneural network