This patent describes a method and system for monitoring the performance of process models (a mathematical or physical model that indicates relationships between inputs and outputs). Rules are used by the monitor to determine when to trigger an alarm.

A continuous emission monitoring system for a manufacturing plant includes a control system which has associated therewith a virtual sensor network. The network is a predictive network that receives as inputs both control values to the plant and also sensor values. The network is then operable to map the inputs through a stored representation of the plant to output a predicted pollutant sensor level. This predicted pollutant sensor level is essentially the prediction of an actual pollutant sensor level that can be measured by a pollutant sensor. The network therefore is a substitute for the pollutant sensor; a virtual sensor.

This patent claims a method for using a predictive model of a engine emissions monitor to predict the noxious pollutants output by a engine, adjusting the predictive model to reduce error, and generating an alarm or modifying the operation of the engine to reduce emissions.

This patent claims a method for training a predictive network by comparing the predicted output with the target output data to generate an error, and adjusting the parameters of the model to minimise the error in accordance with a predetermined training algorithm.