将转炉产粗铜成分的以往化验数据作为二次变量,化验数据严重滞后的当前炉次粗铜成分作为主导变量,并以延迟时间T=2(炉次)将转炉粗铜成分时间序列分成两个子时间序列,采用重构相空间理论、最大Lyapunov指数方法和样本及时更新的办法,建立了一种仅对当前炉次的炼铜转炉粗铜成分进行软测量的混沌时间序列软测量模型.实验结果表明,该软测量模型能快速得出转炉当前炉次的粗铜成分数据,具有很高的精度.
A new soft-sensing model of current crude copper composition based on the chaos time series of foregone crude copper composition was established by using the phase-space reconstruction theory and the method of maximum Lyapunov exponent and by updating samples in time. In this soft-sensing model, foregoing assay data of crude copper composition were regarded as the secondary variables, the assay data of crude copper composition in current copper convertor were taken as the primary variables because of its great delay, and the chaos time series of foregone crude copper composition was divided into two sub-series via the time lag T = 2 furnace orders. Experimental results indicate that the crude copper composition in current copper convertor can be quickly and precisely determined by the proposed soft-sensing model.