Intelligent Optimization Of Mine Ventilation Parameter Monitoring And Intelligent Diagnosis Of Ventilation Network Response

As an important part of mine system,the reasonable and reliable operation of ventilation system is the cornerstone of sustainable coal mining.At present,under the new situation of promoting mine intelligence comprehensively,the quality upgrade of mine ventilation and safety system from automation to intelligence is the key path to realize intelligent mine reduction of personnel,improvement of efficiency and increase of safety.Aiming at the key issues in the construction of intelligent ventilation in mines,this thesis carries out theoretical and model research in the aspects of intelligent perception of ventilation conditions,intelligent matching of safety monitoring system and ventilation network,intelligent diagnosis of ventilation abnormal information and intelligent decision-making of emergency wind control mode,etc.The results will help realize accurate wind measurement in mines,accurate identification and optimization of ventilation network status.It has important research value and practical guiding significance to ensure the intelligent and reliable operation of mine ventilation system.Based on the analysis of mine ventilation measurement model and the study of ventilation network response characteristics,this thesis combines artificial intelligence(AI)technology with monitoring,warning and decision-making methods of mine ventilation network.Carry out research on intelligent calibration of wind speed in shaft and roadway,optimization of location selection of ventilation network monitoring system,intelligent diagnosis of ventilation network anomalies and intelligent regulation and control of emergency response.The main achievements are as follows:(1)In terms of intelligent perception of ventilation parameters,a multi-principle intelligent calibration model for wind measurement based on machine learning is established.The characteristics of local airflow disturbance and wind measurement errors caused by the intrusion of anemometer sensors into the flow field are analyzed.BP neural network algorithm is used to train and establish an intelligent correction model.The test results show that the average errors of ultrasonic and thermal anemometer measurements decrease from 2.42%and 3.04%to 1.19%and 0.46%,respectively.(2)In terms of ventilation monitoring system optimization site selection,an intelligent optimization method of ventilation monitoring network based on ventilation network resistance-air volume characteristics is proposed.The ventilation network solution method was used to study the resistance variation dynamic response characteristics of the wind network,and the branch sensitivity characteristics under variable wind resistance were analyzed.The air volume monitoring vector was integrated with the wind network ill-condition matrix,and the intelligent optimization model of the distributed monitoring network was proposed based on the objective function of the air volume prediction error,and the optimal configuration of ventilation monitoring was realized.(3)In terms of intelligent decision-making for ventilation,deep learning is used to develop an abnormal intelligent diagnosis and wind control model based on ventilation network response characteristics.Convolutional neural network(CNN)algorithm was introduced to establish the diagnosis and regulation model of ventilation anomalies with resistance and variation.Meanwhile,training optimization was carried out.Test results showed that the localization and identification accuracy of ventilation anomalies improved from 82.54%to 96.30%by the optimized model.The coefficient of determination(R~2)between the resistance variation constant and the valve opening diagnostic regression prediction results and the true value were 0.99 and 0.98,respectively.This thesis contains 79 figures,25 tables and 116 references.