In a continuous chemical process the accuracy and reliability of proce
ss and analytical measurements are often a critical factor in control
system performance and ultimately in product uniformity. This paper fo
cuses on our experiences in the area of continuous process improvement
and product variability reduction. It explores some of the present se
nsor deficiencies and their effect on feedback control systems. It dis
cusses intermittent and cyclic disturbances and sensor errors, as well
as some intelligent methods for validation. Finally, it discusses som
e statistical analysis tools for assisting in the detection of ''speci
al causes'' and prerequisites to consider when applying advisory and s
upervisory control strategies.