Characteristic analysis and preprocessing of mine gas monitoring data
-
摘要: 针对矿井瓦斯监测数据包含异常数据、存在数据缺失及数据含噪等特征,提出了瓦斯监测数据预处理方法。首先利用移动平均线处理法或自回归模型处理法进行异常数据替代,然后采用三次指数平滑法补齐缺失数据,最后通过小波软阈值法进行数据消噪处理。实例分析表明,该方法可在不改变瓦斯监测数据统计特征的基础上,消除异常数据的干扰,保证监测数据的完整性,使监测数据表现特征平滑、波动性较小。Abstract: In view of characteristics of abnormal data, missing data and noisy data of mine gas monitoring data, a preprocessing method of gas monitoring data was proposed. Abnormal data is replaced by use of moving average line processing method or auto regressive model processing method, missing data is filled by employing cubic exponential smoothing method and data denoising is processed though wavelet soft threshold method. The case analysis shows that the method can eliminate interference of abnormal data, ensure integrity of monitoring data and smooth characteristic curve of monitoring data without changing statistical characteristics of gas monitoring data.
-
Key words:
- gas monitoring /
- monitoring data /
- characteristic analysis /
- preprocessing
点击查看大图
计量
- 文章访问数: 85
- HTML全文浏览量: 9
- PDF下载量: 7
- 被引次数: 0