公开(公告)号：US07062417B2

公开(公告)日：2006-06-13

申请号：US09815274

申请日：2001-03-23

申请人： Uwe Kruger ,  Qian Chen ,  David J. Sandoz

发明人： Uwe Kruger ,  Qian Chen ,  David J. Sandoz

IPC分类号： G06F17/10 ,  G06F7/60 ,  G05B13/02 ,  H03F1/26

CPC分类号： G06F17/18

摘要： An extended partial least squares (EPLS) approach for the condition monitoring of industrial processes is described. This EPLS approach provides two statistical monitoring charts to detect abnormal process behaviour as well as contribution charts to diagnose this behaviour. A theoretical analysis of the EPLS monitoring charts is provided, together with two application studies to show that the EPLS approach is either more sensitive or provides easier interpretation than conventional PLS.Generalised scores are calculated by constructing an augmented matrix, of the form Z=[Y{dot over (:)}X], where X is the predictor matrix and Y is the response matrix, and constructing a score matrix Tn=T*n−E*n in which T*n and E*n are generally of the form: T n * = [ Y ⁢ ⁢ ⋮ ⁢ ⁢ X ] ⁡ [ B PLS ( n ) : ] 1 ⁢ R n E n * = [ E n ⁢ ⁢ ⋮ ⁢ ⁢ F n ] ⁡ [ B PLS ( n ) : ] 1 ⁢ R n the columns of the matrix T*n providing the generalised t-scores and the columns of the matrix E*n the generalised residual scores, where ℑ denotes an M×M identity matrix, BPLS(n) is the PLS regression matrix.

公开(公告)号：US06816810B2

公开(公告)日：2004-11-09

申请号：US09815275

申请日：2001-03-23

申请人： Manus P. Henry ,  David J. Sandoz ,  David W. Clarke

发明人： Manus P. Henry ,  David J. Sandoz ,  David W. Clarke

IPC分类号： G06F10114

CPC分类号： G08B31/00 ,  G01K15/00 ,  G08B29/20

摘要： Self-validating (SEVA) sensors implemented in a control process provide various metrics regarding sensed variables to a central control unit. Specifically, SEVA sensors provide measurements of the variables and validity information about the measurements, which may include fault information about the sensors themselves. A control unit utilizes the various SEVA metrics even when large numbers of SEVA sensors are used, a situation that is otherwise problematic due to difficulties in assimilating data from multiple SEVA sensors. Accordingly, the control unit distinguishes sensor faults from actual process changes, and responds as needed, even when large numbers of SEVA sensors are implemented together. Specifically, the monitoring and control unit assimilates signals from multiple SEVA sensors using a multi-variate statistical analysis, and compares results of this analysis with a model characterizing behavior of the process (where the model may take into account actuator position information) and/or historical statistical data.

摘要翻译： 在控制过程中实现的自验证（SEVA）传感器向中央控制单元提供关于感测变量的各种指标。 具体来说，SEVA传感器提供关于测量的变量和有效性信息的测量，其可以包括关于传感器本身的故障信息。 即使在使用大量SEVA传感器的情况下，控制单元还利用各种SEVA指标，否则由于难以吸收来自多个SEVA传感器的数据而存在问题。 因此，即使在一起实施大量的SEVA传感器的情况下，控制单元将传感器故障与实际过程变化区分开，并根据需要进行响应。 具体来说，监测和控制单元使用多变量统计分析来吸收来自多个SEVA传感器的信号，并将该分析的结果与该过程的模型特征行为（其中模型可以考虑执行器位置信息）和/或 历史统计资料。