论文部分内容阅读
针对回转窑过程多时间尺度信号的耦合关系往往会掩盖过程变化规律,导致未经时间尺度分解的直接的因子分析方法难以得到过程固有的变化模式的问题,提出采用基于EMD的多尺度因子分析方法对锌钡白煅烧过程的数据进行分析。该方法利用EMD有较强的自适应性,分解结果往往具有较强的物理意义等特性,首先对锌钡白回转窑平稳运行时的生产数据进行EMD分解,根据分解结果提取不同的时间尺度数值,然后对不同尺度下的数据进行因子分析。分析结果显示,数据在周期3 h以上的大时间尺度和周期30 min以下的中小时间尺度上都分别表现出定常的特点,表明人工控制可分解为长周期控制和短周期控制两方面。实验结果表明该方法在处理回转窑过程多尺度数据处理方面的有效性。
The coupling relationship of multi-time scale signals in rotary kiln process often obscures the process variation, which leads to the problem that the direct factor analysis method without time-scale decomposition can not get the inherent change pattern of the process. A multi-scale factor analysis method based on EMD The lithopone calcination process data were analyzed. The EMD method has strong adaptability and the decomposition result has strong physical meaning. The EMD method is firstly used to decompose the production data of the lithized lithopone rotary kiln during steady operation, and different time scales are extracted according to the decomposition results , Then carry on the factor analysis to the data under different scales. The analysis results show that the data show the characteristics of regularity on the large time scale of more than 3 h and the small-scale time scale of 30 min respectively, which shows that the artificial control can be decomposed into long-period control and short-period control. The experimental results show the effectiveness of this method in the multi-scale data processing of rotary kiln process.