Study On Online Oil Monitoring System Based On Multiple Sensors

The healthy operation of the transmission system is a prerequisite for ensuring the safe and efficient production of coal mines.Oil monitoring is an effective technical means to carry out fault diagnosis of the transmission system of mining equipment,and online monitoring is one of its current research hotspots Aiming at the problems of current particle counters unable to classify wear particles and the redundancy of monitoring parameters in the multi-sensor oil analysis technology,etc,a four-coil particle sensor is designed,which can more effectively monitor the metal abrasive particles in the oil,and at the same time,it is combined with the moisture sensor and viscosity sensor to build a multi-sensor remote oil online monitoring system.The belt conveyor reduction gearbox is taken as the research object,moisture,viscosity and metal abrasive content are selected as the monitoring characteristic information of the remote oil online monitoring system.By analyzing the characteristics and advantages of different oil analysis technologies,the overall scheme of the remote oil online monitoring system based on multi-sensor is proposed,and then the overall architecture of the fault diagnosis system integrating multi-feature information is designed.Aiming at the problem that the particle counter cannot classify wear particles,Biot-Savart’s law and the signal defects caused by the coil layout of the traditional threecoil type wear particle sensor are combined to propose the design scheme of a new fourcoil wear particle sensor.The simulation software COMSOL is used to simulate and analyze the three-coil and four-coil wear particle sensor,verifing the superiority of the four-coil wear particle sensor in the coil structure.At the same time,the detection pipeline and circuit of the particle sensor are designed,then the performance of the sensor is discussed.The overall construction of the remote oil online monitoring platform is completed with the gearbox as the object.Aiming at the characteristics of the lubricating oil circuit of the monitored object,a new type of oil sampling module to simplify the complexity of sampling is designed.According to the selected and designed communication mode of the oil analysis sensor,RS485,CAN and Zig Bee are combined to complete the overall layout of the sensor data acquisition module.At the same time,Labview is used to compile the software modules of the upper computer to make the design of humancomputer interaction functions.The fault diagnosis technology based on multi-sensor is discussed,and the online oil fault diagnosis model based on BP neural network is established.The oil fault diagnosis model is trained based on the experimental data collected by the oil online monitoring platform.Aiming at the over-fitting problem of the oil fault diagnosis model,genetic algorithm is used to optimize the parameters of the oil fault diagnosis model,and setup a new GA-BP oil fault diagnosis model.According to the experimental results,the training error of the GA-BP oil fault diagnosis model is only 14% of the BP oil fault diagnosis model,and the training speed,generalization ability of GA-BP oil fault diagnosis model have been greatly improved.This paper contains 65 figures,12 tables,and 93 references.