论文部分内容阅读
为了有效地将时滞信息引入到软测量建模过程中,同时实时跟踪过程动态,本文提出一种基于模糊曲线分析(FCA)估计过程时滞参数的新方法,用离线条件下得到的时滞参数集对软测量建模的数据进行重构;对于新的输入数据,基于一定时刻之前采集的历史变量值,采用时间差—高斯过程回归(TDGPR)模型对当前时刻主导变量值进行在线预测.通过对脱丁烷塔过程的仿真研究,验证了所提方法的有效性和精度.
In order to effectively introduce time-delay information into the process of soft-sensing modeling and to track the process dynamics in real time, a new method based on fuzzy curve analysis (FCA) to estimate the process delay parameters is proposed. With time-delay Parameter sets to reconstruct the data of the soft-sensor modeling; for the new input data, the TDGPR model is used to predict the current dominant value of the variable based on the historical variable values collected before a certain time. The simulation study of the debutanizer process validates the effectiveness and accuracy of the proposed method.