Importance

Loading Importance Indicators...

Abstract

Sensors are spaced along an extended process from its beginning end to its finishing end. Each sensor is associated with a parallel inferential control loop (PICL) and generates an actual measurement signal for a portion of the process. Each PICL includes a process model emulating the sensed portion of the process with the process models being cascaded from the beginning end through and including the finishing end. Each PICL generates a loop control signal corresponding to its sensed portion of process with the loop control signals from the PICLs being summed to generate a total control signal which controls a control element at the beginning end of the process and also is received by the first or most up-stream process model. The process models produce expected measurement signals which are combined with the actual measurement signals to determine difference signals for each PICL which includes a reset model emulating the sensed portion of the process. Each reset model receives a reset input signal from a preceding PICL and generates a reset output signal to decouple control of preceding PICLs from the PICL receiving the reset input signal. One or more of the PICLs can be disabled provided its deviation signal and setpoint signal are nullified, its reset output signal is routed to be the reset input signal for the succeeding PICL and any setpoint signal for the loop is converted and transferred from the disabled loop to a preceding loop.

Loading the Abstract Image...

First Claim