ABSTRACT
There are three levels of monitoring and controlling of a
product in a processing plant. The simplest monitoring is
achieved by randomly pulling samples from a production
line and analyzing these samples offline in a quality control
laboratory. These results are then studied using statistical
processing control[1] to ensure the consistency of the plant
output. The next level is a continuous or semicontinuous
auditing function done by an automatic analyzer. A con-
tinuous monitoring of the finished products can alleviate
the product liability problem but cannot prevent the large-
scale massive production of faulty finished products.
Therefore, there is a need in manufacturing processes to
monitor and control the intermediate products, as well as
the finished products, and make process control decisions
based on the results of on-line measurements. This type of
on-line process control can be visualized as a single-loop-
feedback control system,[2] as shown in Fig. 1. The goal of
the feedback control is to design the controller transfer
function, Gc, to maintain the output, y, at a desired set
point value, r, in spite of the disturbance, d, that affect
the process.
