초록 ▼

An apparatus and method for monitoring a process involve development and application of a statistically qualified neuro-analytic (SQNA) model to accurately and reliably identify process change. The development of the SQNA model is accomplished in two stages: deterministic model adaption and stochast

An apparatus and method for monitoring a process involve development and application of a statistically qualified neuro-analytic (SQNA) model to accurately and reliably identify process change. The development of the SQNA model is accomplished in two stages: deterministic model adaption and stochastic model modification of the deterministic model adaptation. Deterministic model adaption involves formulating an analytic model of the process representing known process characteristics, augmenting the analytic model with a neural network that captures unknown process characteristics, and training the resulting neuro-analytic model by adjusting the neural network weights according to a unique scaled equation error minimization technique. Stochastic model modification involves qualifying any remaining uncertainty in the trained neuro-analytic model by formulating a likelihood function, given an error propagation equation, for computing the probability that the neuro-analytic model generates measured process output. Preferably, the developed SQNA model is validated using known sequential probability ratio tests and applied to the process as an on-line monitoring system. Illustrative of the method and apparatus, the method is applied to a peristaltic pump system.

대표청구항 ▼

1. A method for monitoring operation of an apparatus having known and unknown operating characteristics to detect a change in the operation of the apparatus, comprising the steps of:generating an analytic model of the apparatus from the known operating variables of the apparatus; collecting input da

1. A method for monitoring operation of an apparatus having known and unknown operating characteristics to detect a change in the operation of the apparatus, comprising the steps of:generating an analytic model of the apparatus from the known operating variables of the apparatus; collecting input data and output data characteristic of the operation of the apparatus through a neural network; augmenting the analytic model with a neural network resulting in a combined neuro-analytic model, whereby the neuro-analytic model is capable of learning relationships between the collected input data and collected output data and representing the unknown operating characteristics of the apparatus in the neuro-analytic model; training the neuro-analytic model by minimizing any error between the collected output data and output data generated by the neuro-analytic model for a given set of input data; statistically qualifying any remaining error in the neuro-analytic model by writing an error propagation equation to describe any effect of the error on output data generated by the neuro-analytic model, and applying a statistical technique to maximize the likelihood that the output data generated by the error propagation equation is the same as the collected output data; and utilizing the statistically qualified neuro-analytic model to monitor the operation of the apparatus by observing output data generated by the statistically qualified neuro-analytic model to detect the change in the operation of the apparatus.