面向通风智能化的风速传感器结构化数据降噪方法对比

Comparison of structured data noise reduction methods for airflow speed sensor of intelligent ventilation

为实现风网实时解算、系统优化调节等关键技术,需要去除监测数据噪声,得到结构清晰、纯度较高的通风数据。利用FCM,Rloess和S-G等平滑降噪算法对300组实测风速数据进行分析处理。结果表明:FCM算法处理过程变量引起噪声较为优越,但需提前给定分类数目;窗宽参数选7时,Rloess算法去除由状态变量引起的风速异常数据最优;在窗宽选5、次数为2时,S-G算法降噪和保持数据特性最佳;结合使用FCM-Rloess或FCM-SG算法可有效处理过程变量和状态变量引起的风速异常数据。研究结果可为矿井通风的异常诊断、灾变识别等研究提供合理的基础数据。

In order to realize the key technologies such as the real-time resolution of air network and the optimal adjustment of system,it is necessary to remove the noise of monitoring data and obtain the ventilation data with clear structure and high purity.The smoothing noise reduction algorithms such as FCM,Rloess and S-G were used to analyze and process 300 sets of measured airflow speed data.The results showed that the FCM algorithm was superior to process the noise caused by the process variables,but the number of classifications needed to be given in advance.When the window width parameter was selected as 7,the Rloess algorithm was optimal to remove the abnormal airflow speed data caused by the state variables.When the window width was 5 and the number of times was 2,the S-G algorithm had the best effect of noise reduction and data characteristics retention.The combination of FCM-Rloess or FCM-SG algorithm could effectively process the abnormal airflow speed data caused by the process variables and state variables.The research provides reasonable basic data for the abnormal diagnosis and disaster identification of mine ventilation.