LAUSR

The ultrasonic anemometer is widely used in the field of wind speed measurement in the coal mine. To compensate for the error of wind speed caused by the shadow effect, we propose a method based on kernel extreme learning machine combined with particle swarm optimization (PSO-KELM). Firstly, we established and analyzed the simulation experiment by computational fluid dynamics (CFD). Secondly, we used the PSO to optimize the parameters of the KELM. The evaluation indexes of the PSO-KELM on the test set are better than other regression models. Finally, the accuracy of the PSO-KELM was verified by experimental data obtained by the ultrasonic anemometer in the wind tunnel. The results show that the compensation effect of the PSO-KELM is better than other regression models. The PSO-KELM model proposed in the paper can effectively compensate for the error of wind speed, and make the measurement results of the ultrasonic anemometer more accurate.