Page 46 - Practical Control Engineering a Guide for Engineers, Managers, and Practitioners

Introduction to Developing Control Algorithms 21.

Compartmentalize process into
DYNAMIC MODULES

Gather initial information and
develop requirements for each module

Trme domain analysis
Problem revelation
....,__....;;,__ problem solution
variance reduction

Problem revelation
problem solution Problem revelation
variance reduction problem solution
variance reduction

Problem revelation
problem solution
variance reduction

FIGURE 2-1 The diamond road map.

at or satisfactorily near that set point. The second benefit is the impact
of less variance in that module on the downstream end product. The
third benefit is that once the module is put under control, the set point
can be adjusted to optimize the end-of-line product.
Gathering information about each module, especially its perfor-
mance requirements, is often the most difficult step. What defines
"good performance" for each module? At the end of the manufactur-
ing process where the product emerges, good performance is rela-
tively easy to define. But as you move back into the process this can
become quite difficult. There may be no measurements available for
many of the "interior" or upstream modules in the process. How do
you know if it is performing properly? Could this module be a big
player in the observed poor performance at the end of the process? To
make sense of these studies one must have a reference point that
describes the satisfactory behavior of the module in quantitative
terms. In subsequent sections we will discuss in detail methods for
studying the performance of a module.