Research On Arrangement Of Wind Speed Sensor For Fault Diagnosis Of Ventilation System Based On Neighborhood Rough Set

The normal operation of the mine ventilation system is an important link to ensure the normal operation of the underground production,the health and safety of the staff,and the disaster prevention.The mine ventilation system should be in a stable state for a long time.The abnormal change of the air volume of the ventilation system caused by the change of the roadway caving or deformation,the door opening or damage,the fan performance degradation and so on is called the resistance type fault of the mine ventilation system,and the size of the change of the roadway air resistance caused by the fault is called the fault equivalent resistance.According to the monitoring value of the air volume of the roadway after the occurrence of the mine ventilation system's resistance type fault,the location of the resistance type fault and its equivalent resistance can be diagnosed by using the support vector machine SVM,genetic algorithm GA and other artificial intelligence methods.The core problem is to use the minimum number of wind speed sensors to achieve the accuracy of fault diagnosis to meet the actual needs of the site In this paper,an optimal layout model of resistance type fault diagnosis wind speed sensor of mine ventilation system based on neighborhood rough set is proposed,which can effectively monitor the wind speed by using the least number of wind speed sensors and obtain high accuracy of fault diagnosis.Due to the strict requirements of mine ventilation,not enough samples of "mine ventilation system resistance type fault-air volume" can be produced in practice.Based on the network solution method of mine ventilation simulation system(MVSS),combined with the network topology of specific mine,fault simulation samples are generated as the data base of machine learning.The problem of sensor placement is transformed into a decision-making problem.The fault simulation samples are compiled into neighborhood rough set(NRS)decision-making table.The air volume is regarded as the condition attribute,the fault branch number and the fault equivalent wind resistance are regarded as the decision-making attribute.The importance of the condition attribute to the decision-making attribute is used to determine the installation of the wind speed sensor whose goal is to monitor the fault location and whose goal is to monitor the fault amount In addition,the factors that affect the result of attribute reduction are analyzed.The fault diagnosis model is established by SVM,the air volume of the branch in the fault simulation sample is taken as the input of SVM,the fault branch number and the fault equivalent wind resistance are taken as the output of SVM,and the fault location diagnosis classification model and the fault equivalent wind resistance diagnosis regression model are constructed respectively.Taking the fault location and the fault equivalent wind resistance diagnosis accuracy as the evaluation criteria,verify the effect of using NRS method to determine the installation location optimization of wind speed sensor.The optimization of the position of the wind speed sensor in the mine ventilation system is verified.By using the algorithm of wind speed sensor location optimization proposed in this paper,35 branches need to be installed to diagnose the fault location,accounting for 24%of the total branches,39 branches need to be installed to diagnose the fault location,accounting for 27%of the total branches,and most branches need to install the wind speed sensor are wind resistance branches.In the field,the air volume data is obtained by opening the air door test,and then the fault location and fault volume are diagnosed.The accuracy of fault location diagnosis is 100%,and the absolute error range between the predicted value of fault volume and the real value is(0?0.1037),which proves the effectiveness of the method proposed in this dissertation.The dissertation has 35 pictures,38 tables,and 122 references.