Research on Prediction of Installation Location of Wind Speed Sensor Based on BP Neural Network

Abstract

In order to quickly and effectively determine the installation position of the wind speed sensor suitable for the average wind speed of the roadway section, so as to improve the wind measurement accuracy of the downhole wind speed sensor. Starting from the type of roadway section, roadway width, height, support form and other factors, a large number of numerical simulations are carried out by using numerical simulation software by controlling variables, and the changes in the position of the average wind speed circle in the roadway under different variables are obtained. Taking this data as a sample, a prediction model for the distance between the average wind speed circle and the roadway roof based on the BP neural network is constructed. MATLAB software is used for network training, and the predicted value is compared with the simulated value. At the same time, the predicted result is applied to the installation process of the actual wind speed sensor, and the monitoring data of the wind speed sensor is compared with the manual wind measurement data. The results show that the structure of the BP neural network is relatively simple, and it can converge at a relatively fast speed. The maximum error between the predicted value of the average wind speed circle and the top plate position and the simulated value is within 4%. The monitoring data is basically consistent with the manual wind measurement data, and the neural network model is feasible to solve the position change of the wind speed sensor from the roof under different conditions.