Inferential process modeling, quality prediction and fault detection using multi-stage data segregation

摘要

A process modeling technique uses a single statistical model developed from historical data for a typical process and uses this model to perform quality prediction or fault detection for various different process states of a process. The modeling technique determines means (and possibly standard deviations) of process parameters for each of a set of product grades, throughputs, etc., compares on-line process parameter measurements to these means and uses these comparisons in a single process model to perform quality prediction or fault detection across the various states of the process. In this manner, a single process model can be used to perform quality prediction or fault detection while the process is operating in any of the defined process stages or states.

主权项

1. A computer implemented method of generating a process model for use in analyzing the operation of a process that is capable of operating in a number of different process states as defined by a state variable associated with the process, comprising:
collecting training data from the process during operation of the process, the training data including a value for each of a set of process parameters, a value for the state variable and a value of a result variable associated with each of a multiplicity of different process measurement times; dividing the training data into time slices of data, using a computer processing device, to produce a set of time sliced data for each time slice of data, wherein each set of time sliced data includes a value for each of the set of process parameters, a value for the state variable and a value for the result variable; storing the sets of time sliced data in a computer memory; determining, using a computer processing device, a set of process state means from the training data, the set of process state means including a state variable mean for each of the process states and one or more process parameter means for each of the process states; storing the set of process state means in a computer memory; developing, using a computer processing device, a set of time slice means for each of the time slices of data using the stored process state means, each of the sets of time slice means including a time slice mean for each of the process parameters; developing, using a computer processing device, a set of deviations from the mean for each time slice of data, the set of deviations from the mean for a particular time slice of data including, for each process parameter within the particular time slice of data, using the process parameter value of the particular time slice of data and the time slice mean for the process parameter for the particular time slice of data to develop the deviation from the mean for the process parameter for the particular time slice of data; and generating, using a computer processing device, a process model using the sets of deviations from the mean for the time slices of data and the result variable values for the time slices of data.