Page 21 - Practical Control Engineering a Guide for Engineers, Managers, and Practitioners

XX Preface

Just when you might start to feel comfortable in this new domain
we will leave Chapter Three and I will kick you into the frequency
domain. Chapter Four also adds two more process models to the
reader's toolkit-the pure dead-time process and the first-order with
dead-time process.
Chapter Five expands the first-order process into a third-order
process. This process will be studied in the time and frequency
domains. A new mathematical tool, matrices, will be introduced to
handle the higher dimensionality. Matrices will also provide a means
of looking at processes from the state-space approach which will be
applied to the third-order process.
Chapter Six is devoted to the next new process-the mass/
spring/ dash pot process that has underdamped behavior on its own.
This process is studied in the time, Laplace, frequency and state-space
domains. Proportional-integral control is shown to be lacking so an
extra term containing the derivative is added to the controller. The
chapter concludes with an alternative approach, using state feedback,
which produces a modified process that does not have underdamped
behavior and is easier to control.
Chapter Seven moves on to yet another new process-the
distributed process, epitomized by a tubular heat exchanger. To study
this process model, a new mathematical tool is introduced-partial
differential equations. As before, this new process model will be
studied in the time, Laplace, and frequency domains.
At this point we will have studied five different process models:
first-order, third-order, pure dead-time, first-order with dead-time,
underdamped, and distributed. This set of models covers quite a bit
of territory and will be sufficient for our purposes.
We need control algorithms because processes and process signals
are exposed to disturbances and noise. To properly analyze the
process we must learn how to characterize disturbances and noise.
So, Chapter Eight will open a whole new can of worms, stochastic
processes, that often is bypassed in introductory control engineering
texts but which, if ignored, can be your control engineer's downfall.
Chapters Eight and Nine deal with the discrete time domain, which
also has its associated transform-the Z-transform, which is introduced
in the latter chapter. As we move into these two new domains I will
introduce alternative mathematical structures for our set of process
models which usually require more sophisticated mathematics.
In Chapter Five, I started frequently referring to the state of the
process or system. Chapter Ten comes to grips with the estimation of
the state using the Kalman filter. A state-space based approach to
process control using the Kalman filter is presented and applied to
several example processes.
Although the simple proportional-integral-derivative control
algorithm is used in the development of concepts in Chapters Three
through Nine, the eleventh chapter revisits control algorithms using