【摘要】 根据采集数据噪声的特点,提出了一种 PI 数据库中实时数据的预处理方法。分析了 PI 数据库采集数据中干扰的主要存在形式,给出了一种异常数据剔除法,对异常跳变点进行了处理。仿真结果表明,所提出的方法既有效地去除了异常点,又保持了数据序列的连续性。更多还原

【Abstract】 Based on the characteristics of noise data,a date preprocessing method is put for- ward for real-time data in PI database.The main noise forms are analyzed and a method.for the rejection of outlier data is given to deal with the abnormal data.The simulationl results show that the proposed method effectively eliminates the abnormal data and maintains the continuity of the date sequence.更多还原