Research And Application Of Accurate Measurement Of Mine Wind Speed Based On Data Fusion

With the development of mine disaster ventilation intelligent decision-making and emergency control,it is required to accurately measure the wind speed of the coal mine section.The commonly used single-point measurement of wind speed cannot accurately reflect the average wind speed of the section.The non-contact ultrasonic time difference sensor is used to measure the linear wind speed and convert it to the average wind speed of the section,which can improve the accuracy of the section wind speed measurement.As the flow field and environmental factors in the well will affect the sensor measurement results,the linear wind speed measurement results are inaccurate.Aiming at the problem of inaccurate measurement of mine wind speed,a flow compensation is proposed to correct the linear wind speed;and a two-level data fusion method combining Kalman filter(KF)and improved firework algorithm-neural network(IFWA-BPNN)to reduce The influence of environmental factors improves the accuracy of wind speed measurement results.First,the flow characteristics of the fluid under different conditions are analyzed through fluid simulation software.When the flow is fully developed in the pipeline,turbulent pulsations with different vortices will be generated,and the transducer and its holder are placed in the pipeline,which will affect the measurement route.The vorticity also increases with the flow field distribution above,resulting in inaccurate linear wind speed measurement results.The dynamic Reynolds number is used to calculate the velocity compensation coefficient in the turbulent state to correct the current linear wind speed.The corrected result is compared with the uncorrected result,showing a higher accuracy.After verifying the flow velocity compensation through experiments,the compensation accuracy of the flow velocity compensation method to the wind speed data reached 97.96%.Furthermore,from the analysis of the time difference method measurement principle,it is concluded that environmental factors will affect the time difference measurement,and a two-level data fusion model is proposed for this effect.The first stage uses Kalman filter(KF)data layer fusion algorithm to perform homogeneous data fusion on environmental parameter data,which can effectively reduce the error of environmental measurement values and improve the measurement accuracy of the sensor.Finally,a second-level decision-making layer multi-sensor data fusion method based on improved firework algorithm-neural network(IFWA-BPNN)is proposed.This method uses multi-sensor data fusion of wind speed information and environmental information,and improves the accuracy of wind speed measurement results through multi-source heterogeneous data complementation.The improved firework algorithm is aimed at the phenomenon of premature precocity in the firework algorithm(FWA).The adaptive merging strategy is adopted to enhance the search ability and search speed of FWA,as well as the optimization of the hybrid mutation operator based on Gauss and Levy.Gaussian mutation has better stability and can improve local search capabilities.Levy has a stronger perturbation mutation effect,and occasionally a larger value can help jump out of local optimization.Then integrate the host computer configuration and PLC intelligent control to build a ventilation experiment measurement and control platform,and apply IFWA-BPNN to heterogeneous data fusion.Based on the IFWA-BPNN fusion output,the average relative error of the wind speed value relative to the actual wind speed value is 1.51%,the average error rate dropped by 174.83%,and the accuracy of the wind speed value output by the IFWA-BPNN fusion model reached 98.49%.This paper includes 68 figures,20 tables and 92 references.