A wavelet-based de - noising method was introduced into the breakout prediction system in the continuous casting process to conduct the noise reduction. Temperature data collected by thermocouples was multi- scale decomposed. The threshold was applied to the wavelet coefficients, the temperature signals were reconstructed and the temperature date were de-noised. The trend of the temperature changing can be well shown by the de -noised data results.%将小波分析引入到连铸过程漏钢预报系统中,对热电偶所采集的温度数据进行降噪处理.通过对连铸现场采集的温度数据进行多尺度小波分解,并对小波分解系数作相应的阈值处理,最后重建温度信号,去除了温度数据中的噪声.结果表明,经过小波分析降噪后的温度数据能更好地反应测温点的温度变化趋势.
展开▼